htsthread_wait() returned before threads it should join had started

process_chain was incremented by the child in hts_entry_point(), so a caller
that spawned threads and immediately called htsthread_wait() saw a zero count
and returned at once, free to tear down state the children were about to read.

Count at spawn instead, under the same mutex the waiter reads.

httrack.c freed the option block before waiting; wait first.

Closes #747

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Signed-off-by: Xavier Roche <xroche@gmail.com>

.github/workflows/windows-build.yml

@@ -226,8 +226,10 @@ jobs:
           # Every gate here exits 77, so an all-skipped suite would report green having
           # tested nothing: pin the skips, and floor the passes in case the glob empties.
           # footer-overflow skips on Windows (needs a path past MAX_PATH); crange pending #581;
-          # webdav-mime needs a reapable background listener, which MSYS cannot give it.
-          expected_skips=" 01_engine-footer-overflow.test 48_local-crange-memresume.test 71_local-crange-repaircache.test 79_local-proxytrack-webdav-mime.test"
+          # webdav-mime needs a reapable background listener, which MSYS cannot give it;
+          # badmtime needs a filesystem that stores an mtime past gmtime's range;
+          # single-file ends on a GUI half needing htsserver, which this job does not build.
+          expected_skips=" 01_engine-footer-overflow.test 48_local-crange-memresume.test 71_local-crange-repaircache.test 79_local-proxytrack-webdav-mime.test 88_local-proxytrack-badmtime.test 94_local-single-file.test"
           [ "$pass" -ge 90 ] || { echo "::error::only $pass tests passed ($skip skipped)"; exit 1; }
           [ "$skipped" = "$expected_skips" ] || { echo "::error::unexpected skips:$skipped"; exit 1; }
           [ "$fail" -eq 0 ] || { echo "::error::failing:$failed"; exit 1; }

AGENTS.md

@@ -26,6 +26,12 @@ the operational checklist: toolchain, invariants, and how to ship a change.
   check`, or `PATH="<bld>/src:$PATH"` for a manual run.
 - Give new `.test` scripts `set -e`: the older ones predate the rule, so several
   `local-crawl.sh` calls with no `set -e` report PASS on any non-last failure.
+- Never assert with `cmd | grep -q MARKER && fail`. Under `pipefail` the
+  pipeline is non-zero both when `cmd` fails and when `grep -q` matches early
+  and SIGPIPEs it, so the `&&` never fires and a probe that proved nothing reads
+  as "marker absent". Capture the reply, assert the status line it must carry
+  (an empty, truncated or redirected one is marker-free too), then match with a
+  here-string.
 
 ## Hard invariants
 - **Generated autotools files are NOT in git.** `configure`, every
@@ -48,6 +54,16 @@ the operational checklist: toolchain, invariants, and how to ship a change.
 - Bounds-check every copy. Overflow-safe form: put the untrusted value alone,
   `untrusted < limit - controlled` — never `controlled + untrusted < limit`,
   which can wrap and pass.
+- **Abort or clip is a decision, not a default.** The `*_safe_` helpers
+  (`strcpybuff`, `strlcpybuff`, `strcatbuff`) **abort** on overflow. Right for
+  our own data, wrong for anything read back from a cache, a header or the
+  wire, where it trades a memory smash for a crash on malformed input. Clip
+  with `dst[0] = '\0'; strlncatbuff(dst, src, size, size - 1)`.
+- **A warning class is not the unsafe set.** `-Wformat-truncation` fires only on
+  a *bounded* `snprintf` whose return is discarded, so an unbounded `sprintf`
+  into the same buffer never appears on it. Before scoping a hardening pass off
+  compiler output, grep the unguarded forms yourself (`\bsprintf\s*\(`,
+  `\bstrcpy\s*\(`, `\bstrcat\s*\(`).
 
 ## C conventions
 - **Use the `*t` allocator wrappers, never raw libc** (`htssafe.h`):
@@ -84,6 +100,17 @@ Before pushing, and when reviewing others, don't skim for bugs:
   layout/ABI, cache/wire format, or a security path? A static or unit check
   isn't enough; exercise the wrong behavior at runtime. Claude Code:
   `/review-recipe`.
+- **Poison a canary, never compare it against zero.** Checking that a
+  neighbouring field is still `'\0'` cannot see the stray NUL an off-by-one
+  terminator writes — the exact bug the canary is there for. Fill it with a
+  non-zero byte, and prove it by killing both the stray-`'X'` and the
+  stray-NUL mutant. Neither ASan nor `_FORTIFY_SOURCE` sees an overflow that
+  lands inside the same struct.
+- **Overshoot every destination, not one.** A bounds test that oversizes a
+  single field cannot tell a per-field bound from a one-size-fits-all one, nor
+  from a fix that bounds that field and leaves its neighbours raw. Exercise
+  each destination the path touches, spanning at least two capacities, and
+  check what the code actually emits before writing the expected values.
 
 ## Commits
 - **Sign-off is mandatory.** Every commit carries a `Signed-off-by` trailer:

configure.ac

@@ -100,7 +100,8 @@ AX_CHECK_COMPILE_FLAG([-fstack-protector-strong], [DEFAULT_CFLAGS="$DEFAULT_CFLA
 	[AX_CHECK_COMPILE_FLAG([-fstack-protector], [DEFAULT_CFLAGS="$DEFAULT_CFLAGS -fstack-protector"], [], [-Werror])], [-Werror])
 AX_CHECK_COMPILE_FLAG([-fstack-clash-protection], [DEFAULT_CFLAGS="$DEFAULT_CFLAGS -fstack-clash-protection"], [], [-Werror])
 AX_CHECK_COMPILE_FLAG([-fcf-protection], [DEFAULT_CFLAGS="$DEFAULT_CFLAGS -fcf-protection"], [], [-Werror])
-AX_CHECK_LINK_FLAG([-Wl,--discard-all], [DEFAULT_LDFLAGS="$DEFAULT_LDFLAGS -Wl,--discard-all"])
+# No --discard-all: it drops the local symbols naming every static function, so
+# a trace misattributes them to the nearest surviving global. Costs 0.6% size.
 AX_CHECK_LINK_FLAG([-Wl,--no-undefined], [DEFAULT_LDFLAGS="$DEFAULT_LDFLAGS -Wl,--no-undefined"])
 AX_CHECK_LINK_FLAG([-Wl,-z,relro,-z,now], [DEFAULT_LDFLAGS="$DEFAULT_LDFLAGS -Wl,-z,relro,-z,now"])
 AX_CHECK_LINK_FLAG([-Wl,-z,noexecstack], [DEFAULT_LDFLAGS="$DEFAULT_LDFLAGS -Wl,-z,noexecstack"])
@@ -127,8 +128,8 @@ AX_CHECK_LINK_FLAG([-pie], [LDFLAGS_PIE="-pie"])
 AC_SUBST([CFLAGS_PIE])
 AC_SUBST([LDFLAGS_PIE])
 
-## -rdynamic must be a link flag; DEFAULT_CFLAGS (AM_CPPFLAGS) never reaches the linker.
-AX_CHECK_LINK_FLAG([-rdynamic], [DEFAULT_LDFLAGS="$DEFAULT_LDFLAGS -rdynamic"])
+# Ties a crash trace from a stripped build back to its separate debug symbols.
+AX_CHECK_LINK_FLAG([-Wl,--build-id], [DEFAULT_LDFLAGS="$DEFAULT_LDFLAGS -Wl,--build-id"])
 
 ### Check for -fvisibility=hidden support
 gl_VISIBILITY

fuzz/Makefile.am

@@ -2,7 +2,7 @@
 if FUZZERS
 noinst_PROGRAMS = fuzz-charset fuzz-meta fuzz-idna fuzz-entities \
 	fuzz-unescape fuzz-filters fuzz-url fuzz-header fuzz-cachendx \
-	fuzz-htsparse
+	fuzz-htsparse fuzz-singlefile
 endif
 
 AM_CPPFLAGS = \
@@ -27,6 +27,7 @@ fuzz_url_SOURCES = fuzz-url.c fuzz.h
 fuzz_header_SOURCES = fuzz-header.c fuzz.h
 fuzz_cachendx_SOURCES = fuzz-cachendx.c fuzz.h
 fuzz_htsparse_SOURCES = fuzz-htsparse.c fuzz.h
+fuzz_singlefile_SOURCES = fuzz-singlefile.c fuzz.h
 
 # List corpus files explicitly: automake does not expand EXTRA_DIST globs.
 EXTRA_DIST = README.md run-fuzzers.sh \
@@ -47,4 +48,8 @@ EXTRA_DIST = README.md run-fuzzers.sh \
 	corpus/cachendx/regress-overadvance.bin \
 	corpus/cachendx/regress-truncated-entry.bin \
 	corpus/htsparse/basic.html corpus/htsparse/script-inscript.html \
-	corpus/htsparse/meta-usemap.html corpus/htsparse/malformed.html
+	corpus/htsparse/meta-usemap.html corpus/htsparse/malformed.html \
+	corpus/singlefile/img-src.html corpus/singlefile/link-rel.html \
+	corpus/singlefile/style-block.html corpus/singlefile/style-attr.html \
+	corpus/singlefile/srcset.html corpus/singlefile/rawtext.html \
+	corpus/singlefile/malformed.html corpus/singlefile/many-attrs.html

fuzz/corpus/singlefile/img-src.html

@@ -0,0 +1,3 @@
+<img src="a.png">
+<img src=big.png alt=over-cap>
+<img src="../escape.png"><img src="/abs.png"><img src="data:,x">

fuzz/corpus/singlefile/link-rel.html

@@ -0,0 +1,4 @@
+<link rel="stylesheet" href="s.css">
+<link rel=icon href=a.png>
+<link rel="next" href="p2.html">
+<link rel="preload" href="j.js">

fuzz/corpus/singlefile/malformed.html

@@ -0,0 +1,5 @@
+<img src="unterminated.png
+<div style="background:url(a.png">
+<style>@import url(
+<!-- unterminated comment
+<a href=

fuzz/corpus/singlefile/many-attrs.html

@@ -0,0 +1 @@
+<img src="a.png" a0="v" a1="v" a2="v" a3="v" a4="v" a5="v" a6="v" a7="v" a8="v" a9="v" a10="v" a11="v" a12="v" a13="v" a14="v" a15="v" a16="v" a17="v" a18="v" a19="v" a20="v" a21="v" a22="v" a23="v" a24="v" a25="v" a26="v" a27="v" a28="v" a29="v" a30="v" a31="v" a32="v" a33="v" a34="v" a35="v" a36="v" a37="v" a38="v" a39="v" a40="v" a41="v" a42="v" a43="v" a44="v" a45="v" a46="v" a47="v" a48="v" a49="v" a50="v" a51="v" a52="v" a53="v" a54="v" a55="v" a56="v" a57="v" a58="v" a59="v" a60="v" a61="v" a62="v" a63="v" a64="v" a65="v" a66="v" a67="v" a68="v" a69="v">

fuzz/corpus/singlefile/rawtext.html

@@ -0,0 +1,4 @@
+<script>var s="</scripting>"; if(a</b) x=1;</script>
+<script src="j.js"></script>
+<textarea></textareas></textarea>
+<title></titles></title>

fuzz/corpus/singlefile/srcset.html

@@ -0,0 +1,2 @@
+<img srcset="a.png 1x, big.png 2x, a.png 100w">
+<source srcset="a.png,, a.png 2x," src="a.png">

fuzz/corpus/singlefile/style-attr.html

@@ -0,0 +1,2 @@
+<div style="background:url(a.png);list-style:url('a.png')"></div>
+<p style='background:url("a.png")'>x</p>

fuzz/corpus/singlefile/style-block.html

@@ -0,0 +1,5 @@
+<style>@import "s.css";
+@import url(sub/b.css);
+div{background:url(a.png)}
+/* url(a.png) */ p:after{content:"url(a.png)"}
+</style>

fuzz/fuzz-singlefile.c

@@ -0,0 +1,131 @@
+/* ------------------------------------------------------------ */
+/*
+HTTrack Website Copier, Offline Browser for Windows and Unix
+Copyright (C) 2026 Xavier Roche and other contributors
+
+SPDX-License-Identifier: GPL-3.0-or-later
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation, either version 3 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program. If not, see <http://www.gnu.org/licenses/>.
+
+Ethical use: we kindly ask that you NOT use this software to harvest email
+addresses or to collect any other private information about people. Doing so
+would dishonor our work and waste the many hours we have spent on it.
+
+Please visit our Website: http://www.httrack.com
+*/
+
+/* Fuzz the --single-file rewriter (htssinglefile.c): hostile HTML walked
+   through the tag, CSS url()/@import and srcset parsers, then re-serialized.
+   The resolver is aimed at a private temp tree, so the inlining half (MIME
+   guess, base64, nested stylesheet) is reached and nothing else on disk is. */
+#include "fuzz.h"
+
+#include "httrack-library.h"
+#include "htssinglefile.h"
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/stat.h>
+
+/* Between a.png and big.png, so one input reaches both the inline path and the
+   over-cap fallback. */
+#define FUZZ_SF_CAP 64
+
+static char sf_root[512];
+static char sf_page[600];
+
+/* The asset tree, in removal order: the subdirectory comes after its file. */
+static const char *const sf_files[] = {"a.png",     "big.png", "j.js", "s.css",
+                                       "sub/b.css", "sub",     NULL};
+
+static void sf_cleanup(void) {
+  char path[700];
+  int i;
+
+  for (i = 0; sf_files[i] != NULL; i++) {
+    snprintf(path, sizeof(path), "%s/%s", sf_root, sf_files[i]);
+    (void) remove(path);
+  }
+  (void) remove(sf_root);
+}
+
+/* A missing asset would silently reduce the target to its parser half. */
+static void sf_write(const char *name, const char *data, size_t len) {
+  char path[700];
+  FILE *fp;
+
+  snprintf(path, sizeof(path), "%s/%s", sf_root, name);
+  fp = fopen(path, "wb");
+  if (fp == NULL || fwrite(data, 1, len, fp) != len)
+    abort();
+  fclose(fp);
+}
+
+static void sf_text(const char *name, const char *data) {
+  sf_write(name, data, strlen(data));
+}
+
+static void sf_init(void) {
+  static const char png[] = "\x89PNG\r\n\x1a\n";
+  static const char big[4096] = "\x89PNG";
+  const char *tmp = getenv("TMPDIR");
+  char path[700];
+
+  hts_init();
+  snprintf(sf_root, sizeof(sf_root), "%s/httrack-fuzz-sf-XXXXXX",
+           tmp != NULL && tmp[0] != '\0' ? tmp : "/tmp");
+  if (mkdtemp(sf_root) == NULL)
+    abort();
+  atexit(sf_cleanup);
+  snprintf(sf_page, sizeof(sf_page), "%s/page.html", sf_root);
+  snprintf(path, sizeof(path), "%s/sub", sf_root);
+  if (mkdir(path, 0700) != 0)
+    abort();
+  sf_write("a.png", png, sizeof(png) - 1);
+  sf_write("big.png", big, sizeof(big));
+  sf_text("j.js", "var x=1;\n");
+  /* @import plus a url(), so an inlined stylesheet recurses and its own
+     relative reference is rebased. */
+  sf_text("s.css", "@import url(sub/b.css);\ndiv{background:url(a.png)}\n");
+  sf_text("sub/b.css", "p{background:url(../a.png)}\n");
+}
+
+int LLVMFuzzerTestOneInput(const uint8_t *data, size_t size) {
+  static int inited = 0;
+
+  String out = STRING_EMPTY;
+  httrackp *opt;
+  /* Exact-length, unterminated: the rewriter is span-based, so ASan bounds a
+     read past html_len instead of it landing on a terminator. */
+  char *html = malloct(size != 0 ? size : 1);
+
+  if (!inited) {
+    sf_init();
+    inited = 1;
+  }
+  memcpy(html, data, size);
+
+  opt = hts_create_opt();
+  opt->log = opt->errlog = NULL;
+  opt->single_file_max_size = FUZZ_SF_CAP;
+
+  StringClear(out);
+  (void) singlefile_rewrite_html(opt, sf_root, sf_page, html, size,
+                                 SINGLEFILE_MAX_PAGE_SIZE, &out);
+
+  StringFree(out);
+  freet(html);
+  hts_free_opt(opt);
+  return 0;
+}

html/changes.html

@@ -0,0 +1,263 @@
+<html xmlns="http://www.w3.org/1999/xhtml" lang="en">
+
+<head>
+	<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1" />
+	<meta name="description" content="HTTrack is an easy-to-use website mirror utility. It allows you to download a World Wide website from the Internet to a local directory,building recursively all structures, getting html, images, and other files from the server to your computer. Links are rebuiltrelatively so that you can freely browse to the local site (works with any browser). You can mirror several sites together so that you can jump from one toanother. You can, also, update an existing mirror site, or resume an interrupted download. The robot is fully configurable, with an integrated help" />
+	<meta name="keywords" content="httrack, HTTRACK, HTTrack, winhttrack, WINHTTRACK, WinHTTrack, offline browser, web mirror utility, aspirateur web, surf offline, web capture, www mirror utility, browse offline, local  site builder, website mirroring, aspirateur www, internet grabber, capture de site web, internet tool, hors connexion, unix, dos, windows 95, windows 98, solaris, ibm580, AIX 4.0, HTS, HTGet, web aspirator, web aspirateur, libre, GPL, GNU, free software" />
+	<title>HTTrack Website Copier - Change report format specification</title>
+
+	<style type="text/css">
+	<!--
+
+body {
+	margin: 0;  padding: 0;  margin-bottom: 15px;  margin-top: 8px;
+	background: #77b;
+}
+body, td {
+	font: 14px "Trebuchet MS", Verdana, Arial, Helvetica, sans-serif;
+	}
+
+#subTitle {
+	background: #000;  color: #fff;  padding: 4px;  font-weight: bold; 
+	}
+
+#siteNavigation a, #siteNavigation .current {
+	font-weight: bold;  color: #448;
+	}
+#siteNavigation a:link    { text-decoration: none; }
+#siteNavigation a:visited { text-decoration: none; }
+
+#siteNavigation .current { background-color: #ccd; }
+
+#siteNavigation a:hover   { text-decoration: none;  background-color: #fff;  color: #000; }
+#siteNavigation a:active  { text-decoration: none;  background-color: #ccc; }
+
+
+a:link    { text-decoration: underline;  color: #00f; }
+a:visited { text-decoration: underline;  color: #000; }
+a:hover   { text-decoration: underline;  color: #c00; }
+a:active  { text-decoration: underline; }
+
+#pageContent {
+	clear: both;
+	border-bottom: 6px solid #000;
+	padding: 10px;  padding-top: 20px;
+	line-height: 1.65em;
+	background-image: url(images/bg_rings.gif);
+	background-repeat: no-repeat;
+	background-position: top right;
+	}
+
+#pageContent, #siteNavigation {
+	background-color: #ccd;
+	}
+
+
+.imgLeft  { float: left;   margin-right: 10px;  margin-bottom: 10px; }
+.imgRight { float: right;  margin-left: 10px;   margin-bottom: 10px; }
+
+hr { height: 1px;  color: #000;  background-color: #000;  margin-bottom: 15px; }
+
+h1 { margin: 0;  font-weight: bold;  font-size: 2em; }
+h2 { margin: 0;  font-weight: bold;  font-size: 1.6em; }
+h3 { margin: 0;  font-weight: bold;  font-size: 1.3em; }
+h4 { margin: 0;  font-weight: bold;  font-size: 1.18em; }
+
+.blak { background-color: #000; }
+.hide { display: none; }
+.tableWidth { min-width: 400px; }
+
+.tblRegular       { border-collapse: collapse; }
+.tblRegular td    { padding: 6px;  background-image: url(fade.gif);  border: 2px solid #99c; }
+.tblHeaderColor, .tblHeaderColor td { background: #99c; }
+.tblNoBorder td   { border: 0; }
+
+
+// -->
+</style>
+
+</head>
+
+<table width="76%" border="0" align="center" cellspacing="0" cellpadding="0" class="tableWidth">
+	<tr>
+	<td><img src="images/header_title_4.gif" width="400" height="34" alt="HTTrack Website Copier" title="" border="0" id="title" /></td>
+	</tr>
+</table>
+<table width="76%" border="0" align="center" cellspacing="0" cellpadding="3" class="tableWidth">
+	<tr>
+	<td id="subTitle">Open Source offline browser</td>
+	</tr>
+</table>
+<table width="76%" border="0" align="center" cellspacing="0" cellpadding="0" class="tableWidth">
+<tr class="blak">
+<td>
+	<table width="100%" border="0" align="center" cellspacing="1" cellpadding="0">
+	<tr>
+	<td colspan="6"> 
+		<table width="100%" border="0" align="center" cellspacing="0" cellpadding="10">
+		<tr> 
+		<td id="pageContent"> 
+<!-- ==================== End prologue ==================== -->
+
+<h2 align="center"><em>Change report format specification</em></h2>
+
+<br />
+
+Run with <tt>--changes</tt> (<tt>-%d</tt>), HTTrack writes <tt>hts-changes.json</tt>
+in the project directory, next to <tt>hts-log.txt</tt>, describing what the crawl
+left new, changed, unchanged and gone compared to the previous mirror. The file is
+rewritten from scratch at the end of every run, and the log carries a one-line
+summary of the same counts.
+
+<br /><br />
+
+<h3>What "changed" means</h3>
+
+A resource is changed when its bytes differ, not when the server merely re-sent
+it. HTTrack compares the payload it just received against the copy the previous
+run left behind: for pages it parses, the previous payload comes from the cache
+(the file on disk carries the mirror footer and its crawl date, so its bytes
+differ on every run); for everything else, the mirrored file is the payload
+verbatim and is compared directly.
+
+<br /><br />
+
+Where no digest can be taken on either side, because the cache is disabled or
+the previous copy is gone, the report falls back to the transfer signal, and a
+server that answers 200 rather than 304 reads as changed. Keeping the cache on
+(the default) is what makes the report precise.
+
+<br /><br />
+
+<h3>With the cache off</h3>
+
+<tt>--cache=0</tt> costs the report more than the digest of a parsed page. The
+mirror's file index (<tt>hts-cache/new.lst</tt>) is what records which files a
+run produced, so without it there is no previous mirror to subtract from: nothing
+is reported <tt>gone</tt>, and whether the run is a first crawl cannot be decided
+at all, which <tt>first_crawl</tt> states as <tt>null</tt> rather than guess. What
+is on disk is still compared byte for byte, so the other three lists stay
+meaningful, except for the pages HTTrack parses: those have no cached payload to
+compare against and fall back to the transfer signal.
+
+<br /><br />
+
+<h3>Fields</h3>
+
+<ul>
+<li><tt>schema</tt>: format version, currently <tt>1</tt>. It is bumped only
+on an incompatible change; new fields may appear without one.</li>
+<li><tt>generator</tt>: the HTTrack build that wrote the file.</li>
+<li><tt>date</tt>: when the report was written, UTC, <tt>YYYY-MM-DDThh:mm:ssZ</tt>.</li>
+<li><tt>first_crawl</tt>: true when no index of a previous mirror
+(<tt>hts-cache/old.lst</tt>) was found, so there was nothing to compare against and
+everything is listed as new. Null when the run kept no index at all and the
+question cannot be answered (see above).</li>
+<li><tt>partial</tt>: true when the report ran out of memory and lists only
+part of the mirror.</li>
+<li><tt>purged</tt>: true when <tt>--purge-old</tt> was in effect, so the
+files under <tt>gone</tt> were also deleted from disk.</li>
+<li><tt>counts</tt>: the size of each of the four lists.</li>
+<li><tt>new</tt>, <tt>changed</tt>, <tt>unchanged</tt>, <tt>gone</tt>: the
+lists themselves. Every mirrored file appears in exactly one of them.</li>
+</ul>
+
+Each entry is an object:
+
+<ul>
+<li><tt>url</tt>: the absolute URL the file came from. Empty under
+<tt>gone</tt>: deletions are computed from the mirror's file index, which records
+paths, not URLs.</li>
+<li><tt>file</tt>: the path relative to the mirror root, with forward
+slashes. This is the entry's identity: a URL and a redirect that resolve to the
+same local file are one entry, not two.</li>
+<li><tt>size</tt>: the mirrored file's size in bytes, absent when the file
+is not on disk.</li>
+<li><tt>previous_size</tt>: under <tt>changed</tt> only, the size of the
+copy the previous run left.</li>
+</ul>
+
+<br />
+
+<h3>Encoding</h3>
+
+The file is JSON, UTF-8. URLs and local paths reach HTTrack as raw bytes and are
+not guaranteed to be valid UTF-8; any byte sequence that is not becomes
+U+FFFD (<tt>\ufffd</tt>), so the file always parses. Compare on <tt>file</tt>
+rather than on <tt>url</tt> when a mirror is known to carry legacy-charset URLs.
+
+<br /><br />
+
+<h3>Example</h3>
+
+<pre>
+{
+  "schema": 1,
+  "generator": "HTTrack Website Copier/3.49-14",
+  "date": "2026-07-26T15:29:03Z",
+  "first_crawl": false,
+  "partial": false,
+  "purged": true,
+  "counts": { "new": 1, "changed": 1, "unchanged": 1, "gone": 1 },
+  "new": [
+    { "url": "http://example.com/d.html", "file": "example.com/d.html", "size": 280 }
+  ],
+  "changed": [
+    { "url": "http://example.com/a.html", "file": "example.com/a.html", "size": 281, "previous_size": 273 }
+  ],
+  "unchanged": [
+    { "url": "http://example.com/b.html", "file": "example.com/b.html", "size": 277 }
+  ],
+  "gone": [
+    { "url": "", "file": "example.com/c.html" }
+  ]
+}
+</pre>
+
+<br /><br />
+
+<h3>Notes</h3>
+
+<ul>
+<li>A file listed under <tt>gone</tt> is only deleted when <tt>--purge-old</tt> is
+on. Left in place it drops out of the mirror's index, so it is reported once and
+not again.</li>
+<li>A resource whose local file name changed since the previous mirror (a new
+MIME type, say) is reported as <tt>new</tt> under its new name; the old name is
+reported as <tt>gone</tt> only if the file is still on disk. The two entries are
+not paired.</li>
+<li>A resource this run tried and failed to transfer also drops out of the
+mirror's index, but its previous copy is untouched, so it is reported
+<tt>unchanged</tt>. Under <tt>--purge-old</tt> that copy is deleted anyway, and
+the report says <tt>gone</tt> to match.</li>
+<li>A run that transfers no data at all is rolled back: HTTrack restores the
+previous cache generation and leaves the previous report in place, so a lost
+connection does not overwrite a good report with an empty one.</li>
+<li>Content diffs, and keeping the previous copy of a changed page, are out of
+scope: both change what a mirror directory contains.</li>
+</ul>
+
+<br /><br />
+
+<!-- ==================== Start epilogue ==================== -->
+		</td>
+		</tr>
+		</table>
+	</td>
+	</tr>
+	</table>
+</td>
+</tr>
+</table>
+
+<table width="76%" border="0" align="center" valign="bottom" cellspacing="0" cellpadding="0">
+	<tr>
+	<td id="footer"><small>&copy; 1998-2026 Xavier Roche & other contributors - Web Design: Leto Kauler.</small></td>
+	</tr>
+</table>
+
+</body>
+
+</html>
+
+

html/cmdguide.html

@@ -300,6 +300,7 @@ ones it kept so the local copy browses offline. These options tune both halves.<
   <tr><td><tt>--preserve (-%p), --disable-passwords (-%x)</tt></td><td>Leave HTML untouched (no rewriting), and strip passwords out of saved links.</td></tr>
   <tr><td><tt>--extended-parsing (-%P), --parse-java (-j)</tt></td><td>Aggressive link discovery, and how much script content is parsed for links.</td></tr>
   <tr><td><tt>--mime-html (-%M)</tt></td><td>Save the whole mirror as a single MIME-encapsulated <tt>.mht</tt> archive (<tt>index.mht</tt>).</td></tr>
+  <tr><td><tt>--single-file (-%Z), --single-file-max-size N</tt></td><td>Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded as <tt>data:</tt> URIs. Assets over the cap (10&nbsp;MB by default) keep their link, as do audio, video, and the links from one page to another. A sibling of <tt>-%M</tt>, not a replacement: see the recipe below for which to pick.</td></tr>
   <tr><td><tt>--index (-I), --build-top-index (-%i), --search-index (-%I)</tt></td><td>Build a per-mirror index, a top index across projects, and a searchable keyword index.</td></tr>
 </table>
 
@@ -482,6 +483,34 @@ add <tt>--warc-cdx</tt> for a sorted CDXJ index, or <tt>--wacz</tt> to bundle th
 archive, index and pages into one WACZ for replay tools such as
 replayweb.page.</small></p>
 
+<h4>See what a re-crawl changed</h4>
+<p><tt>httrack https://example.com/ --update --changes --path mydir</tt><br>
+<small>Writes <tt>hts-changes.json</tt> in the project folder, listing every
+mirrored file as new, changed, unchanged or gone, plus a one-line summary in the
+log. &quot;Changed&quot; means the bytes really differ: a server that answers 200
+with the same content it served last time lands in <tt>unchanged</tt>. Deletions
+are reported whether or not <tt>--purge-old</tt> is deleting them. The format is
+documented in <a href="changes.html">the change report specification</a>.</small></p>
+
+<h4>Pages you can hand to someone as one file</h4>
+<p><tt>httrack https://example.com/ --single-file --path mydir</tt><br>
+<small>Rewrites each saved page after the crawl so its stylesheets, scripts,
+images and fonts are embedded as <tt>data:</tt> URIs. The mirror stays a normal
+browsable tree, with links between pages relative and the assets still on disk,
+but any single <tt>.html</tt> file can now be mailed or archived on its own.
+Raise or lower the 10&nbsp;MB per-asset limit with
+<tt>--single-file-max-size N</tt>; anything over it, plus audio and video, keeps
+its link. One caveat if you raise it: an inlined stylesheet becomes a
+<tt>data:</tt> URL, whose path is opaque, so an asset it referenced that stayed
+over the cap no longer resolves from inside it. Raising the cap past that asset
+embeds it too and the question goes away.<br>
+Reach for this when the file has to open for someone you cannot make assumptions
+about: it is plain HTML and needs no add-on. Reach for <tt>--mime-html</tt>
+instead when a Chromium-family browser is a given and the mirror is large: MIME
+carries text parts without the base64 tax, keeps each resource's original URL,
+and stores a shared asset once rather than re-embedding it in every page that
+uses it.</small></p>
+
 <h4>HTTrack as a fetch tool</h4>
 <p><tt>httrack --get https://host/file.bin --path tmp</tt><br>
 <small><tt>--get</tt> fetches one file with cache, index, depth, cookies and robots
@@ -490,7 +519,8 @@ all disabled.</small></p>
 <p><br>For the complete, always-current option list, see
 <a href="httrack.man.html">the manual page</a>. For the filter language, see
 <a href="filters.html">filters</a>; for the cache and updates, see
-<a href="cache.html">cache</a>.</p>
+<a href="cache.html">cache</a>; for the change report, see
+<a href="changes.html">changes</a>.</p>
 
 <!-- ==================== Start epilogue ==================== -->
 		</td>

html/dev.html

@@ -129,6 +129,9 @@ The library can be used to write graphical GUIs for httrack, or to run mirrors f
 <li><a href="cache.html">Cache format</a></li><br>
 HTTrack stores original HTML data and references to downloaded files in a cache, located in the hts-cache directory.
 This page describes the HTTrack cache format.
+<li><a href="changes.html">Change report format</a></li><br>
+With --changes, HTTrack writes hts-changes.json describing what the crawl left new, changed, unchanged and gone
+compared to the previous mirror. This page describes that file.
 </ul>
 
 

html/httrack.man.html

@@ -90,9 +90,10 @@ offline browser : copy websites to a local directory</p>
 ] [ <b>-NN, --structure[=N]</b> ] [ <b>-%N,
 --delayed-type-check</b> ] [ <b>-%D,
 --cached-delayed-type-check</b> ] [ <b>-%M, --mime-html</b>
-] [ <b>-LN, --long-names[=N]</b> ] [ <b>-KN,
---keep-links[=N]</b> ] [ <b>-x, --replace-external</b> ] [
-<b>-%x, --disable-passwords</b> ] [ <b>-%q,
+] [ <b>-%Z, --single-file</b> ] [ <b>-LN,
+--long-names[=N]</b> ] [ <b>-KN, --keep-links[=N]</b> ] [
+<b>-x, --replace-external</b> ] [ <b>-%x,
+--disable-passwords</b> ] [ <b>-%q,
 --include-query-string</b> ] [ <b>-%g, --strip-query</b> ] [
 <b>-o, --generate-errors</b> ] [ <b>-X, --purge-old[=N]</b>
 ] [ <b>-%p, --preserve</b> ] [ <b>-%T, --utf8-conversion</b>
@@ -108,16 +109,17 @@ offline browser : copy websites to a local directory</p>
 --footer</b> ] [ <b>-%l, --language</b> ] [ <b>-%a,
 --accept</b> ] [ <b>-%X, --headers</b> ] [ <b>-C,
 --cache[=N]</b> ] [ <b>-k, --store-all-in-cache</b> ] [
-<b>-%r, --warc</b> ] [ <b>-%n, --do-not-recatch</b> ] [
-<b>-%v, --display</b> ] [ <b>-Q, --do-not-log</b> ] [ <b>-q,
---quiet</b> ] [ <b>-z, --extra-log</b> ] [ <b>-Z,
---debug-log</b> ] [ <b>-v, --verbose</b> ] [ <b>-f,
---file-log</b> ] [ <b>-f2, --single-log</b> ] [ <b>-I,
---index</b> ] [ <b>-%i, --build-top-index</b> ] [ <b>-%I,
---search-index</b> ] [ <b>-pN, --priority[=N]</b> ] [ <b>-S,
---stay-on-same-dir</b> ] [ <b>-D, --can-go-down</b> ] [
-<b>-U, --can-go-up</b> ] [ <b>-B, --can-go-up-and-down</b> ]
-[ <b>-a, --stay-on-same-address</b> ] [ <b>-d,
+<b>-%r, --warc</b> ] [ <b>-%d, --changes</b> ] [ <b>-%n,
+--do-not-recatch</b> ] [ <b>-%v, --display</b> ] [ <b>-Q,
+--do-not-log</b> ] [ <b>-q, --quiet</b> ] [ <b>-z,
+--extra-log</b> ] [ <b>-Z, --debug-log</b> ] [ <b>-v,
+--verbose</b> ] [ <b>-f, --file-log</b> ] [ <b>-f2,
+--single-log</b> ] [ <b>-I, --index</b> ] [ <b>-%i,
+--build-top-index</b> ] [ <b>-%I, --search-index</b> ] [
+<b>-pN, --priority[=N]</b> ] [ <b>-S, --stay-on-same-dir</b>
+] [ <b>-D, --can-go-down</b> ] [ <b>-U, --can-go-up</b> ] [
+<b>-B, --can-go-up-and-down</b> ] [ <b>-a,
+--stay-on-same-address</b> ] [ <b>-d,
 --stay-on-same-domain</b> ] [ <b>-l, --stay-on-same-tld</b>
 ] [ <b>-e, --go-everywhere</b> ] [ <b>-%H,
 --debug-headers</b> ] [ <b>-%!,
@@ -648,6 +650,24 @@ don&rsquo;t wait) (--cached-delayed-type-check)</p></td></tr>
 <td width="4%">
 
 
+<p>-%Z</p></td>
+<td width="5%"></td>
+<td width="82%">
+
+
+<p>after the mirror, rewrite each saved page with its
+stylesheets, scripts, images and fonts inlined as data:
+URIs, so any page opens by double-click anywhere (links
+between pages stay relative; audio and video stay links);
+--single-file-max-size N caps each asset (default 10485760
+bytes). %M is the better container where a Chromium-family
+browser is a given: one archive, no base64 tax on text, a
+shared asset stored once (--single-file)</p></td></tr>
+<tr valign="top" align="left">
+<td width="9%"></td>
+<td width="4%">
+
+
 <p>-%t</p></td>
 <td width="5%"></td>
 <td width="82%">
@@ -1143,6 +1163,19 @@ past N bytes, --warc-cdx also writes a sorted CDXJ index,
 <td width="4%">
 
 
+<p>-%d</p></td>
+<td width="5%"></td>
+<td width="82%">
+
+
+<p>write hts-changes.json listing what this crawl left new,
+changed, unchanged and gone compared to the previous mirror
+(--changes)</p> </td></tr>
+<tr valign="top" align="left">
+<td width="9%"></td>
+<td width="4%">
+
+
 <p>-%n</p></td>
 <td width="5%"></td>
 <td width="82%">

html/server/finished.html

@@ -97,9 +97,8 @@ ${do:end-if}
 </pre>
 
 ${LANG_G8} : 
-${do:output-mode:html-urlescaped}
-<a href="file://${path}/${projname}/" target="_new">
-${do:output-mode:}
+${/* an http: page cannot navigate to file:, so the mirror is reached through the server */}
+<a href="/website/index.html" target="_new">
 ${path}/${projname}
 </a></li>
 <ul>

html/server/option10.html

@@ -98,6 +98,9 @@ ${do:end-if}
 <input type="hidden" name="redirect" value="">
 <input type="hidden" name="closeme" value="">
 
+<!-- clear if not checked -->
+<input type="hidden" name="ftpprox" value="">
+
 ${LANG_PROXYTYPE}
 <select name="proxytype"
 			title='${html:LANG_PROXYTYPETIP}' onMouseOver="info('${html:LANG_PROXYTYPETIP}'); return true" onMouseOut="info('&nbsp;'); return true"

html/server/option2.html

@@ -98,6 +98,14 @@ ${do:end-if}
 <input type="hidden" name="redirect" value="">
 <input type="hidden" name="closeme" value="">
 
+<!-- clear if not checked -->
+<input type="hidden" name="errpage" value="">
+<input type="hidden" name="external" value="">
+<input type="hidden" name="hidepwd" value="">
+<input type="hidden" name="hidequery" value="">
+<input type="hidden" name="nopurge" value="">
+<input type="hidden" name="singlefile" value="">
+
 ${LANG_I33}
 <br>
 <select name="build"
@@ -136,6 +144,13 @@ ${listid:build:LISTDEF_3}
 <tr><td><input type="checkbox" name="nopurge"   ${checked:nopurge}  
 			title='${html:LANG_I1a}' onMouseOver="info('${html:LANG_I1a}'); return true" onMouseOut="info('&nbsp;'); return true"
 > ${LANG_I57}</td></tr>
+<tr><td><input type="checkbox" name="singlefile" ${checked:singlefile}
+			title='${html:LANG_SINGLEFILETIP}' onMouseOver="info('${html:LANG_SINGLEFILETIP}'); return true" onMouseOut="info('&nbsp;'); return true"
+> ${LANG_SINGLEFILE}</td></tr>
+<tr><td>${LANG_SINGLEFILEMAX}
+<input name="singlefilemax" value="${singlefilemax}" size="12"
+			title='${html:LANG_SINGLEFILEMAXTIP}' onMouseOver="info('${html:LANG_SINGLEFILEMAXTIP}'); return true" onMouseOut="info('&nbsp;'); return true"
+></td></tr>
 </table>
 
 <tr><td>

html/server/option2b.html

@@ -103,7 +103,7 @@ ${LANG_Q3}
 <tr><td>
 	<table width="100%">
 	<tr><td align="left">
-	<input type="submit" value="${LANG_OK]" 
+	<input type="submit" value="${LANG_OK}"
 ${do:output-mode:html-urlescaped}
 		onClick="if (confirm(str_replace(str_replace('${LANG_DIAL7}', '%20', ' '), '%0a', ' '))) { form.closeme.value=1; form.submit(); } return false;"
 ${do:output-mode:}

html/server/option3.html

@@ -98,6 +98,9 @@ ${do:end-if}
 <input type="hidden" name="redirect" value="">
 <input type="hidden" name="closeme" value="">
 
+<!-- clear if not checked -->
+<input type="hidden" name="windebug" value="">
+
 ${LANG_I40c}
 <br>
 

html/server/option4.html

@@ -98,6 +98,11 @@ ${do:end-if}
 <input type="hidden" name="redirect" value="">
 <input type="hidden" name="closeme" value="">
 
+<!-- clear if not checked -->
+<input type="hidden" name="ka" value="">
+<input type="hidden" name="remt" value="">
+<input type="hidden" name="rems" value="">
+
 <table border="0" width="100%" cellspacing="0">
 
 <tr><td>

html/server/option8.html

@@ -98,6 +98,17 @@ ${do:end-if}
 <input type="hidden" name="redirect" value="">
 <input type="hidden" name="closeme" value="">
 
+<!-- clear if not checked -->
+<input type="hidden" name="cookies" value="">
+<input type="hidden" name="parsejava" value="">
+<input type="hidden" name="updhack" value="">
+<input type="hidden" name="urlhack" value="">
+<input type="hidden" name="keepwww" value="">
+<input type="hidden" name="keepslashes" value="">
+<input type="hidden" name="keepqueryorder" value="">
+<input type="hidden" name="toler" value="">
+<input type="hidden" name="http10" value="">
+
 <input type="checkbox" name="cookies" ${checked:cookies}
 			title='${html:LANG_I1b}' onMouseOver="info('${html:LANG_I1b}'); return true" onMouseOut="info('&nbsp;'); return true"
 > ${LANG_I58}

html/server/option9.html

@@ -98,6 +98,14 @@ ${do:end-if}
 <input type="hidden" name="redirect" value="">
 <input type="hidden" name="closeme" value="">
 
+<!-- clear if not checked -->
+<input type="hidden" name="warc" value="">
+<input type="hidden" name="changes" value="">
+<input type="hidden" name="norecatch" value="">
+<input type="hidden" name="logf" value="">
+<input type="hidden" name="index" value="">
+<input type="hidden" name="index2" value="">
+
 <input type="checkbox" name="cache2" ${checked:cache2}
 			title='${html:LANG_I1e}' onMouseOver="info('${html:LANG_I1e}'); return true" onMouseOut="info('&nbsp;'); return true"
 > ${LANG_I61}
@@ -114,6 +122,11 @@ ${LANG_WARCFILE}
 >
 <br><br>
 
+<input type="checkbox" name="changes" ${checked:changes}
+			title='${html:LANG_CHANGESTIP}' onMouseOver="info('${html:LANG_CHANGESTIP}'); return true" onMouseOut="info('&nbsp;'); return true"
+> ${LANG_CHANGES}
+<br><br>
+
 <input type="checkbox" name="norecatch" ${checked:norecatch}
 			title='${html:LANG_I5b}' onMouseOver="info('${html:LANG_I5b}'); return true" onMouseOut="info('&nbsp;'); return true"
 > ${LANG_I34b}

html/server/step2.html

@@ -143,6 +143,9 @@ ${do:copy:StripQuery:stripquery}
 ${do:copy:StoreAllInCache:cache2}
 ${do:copy:Warc:warc}
 ${do:copy:WarcFile:warcfile}
+${do:copy:Changes:changes}
+${do:copy:SingleFile:singlefile}
+${do:copy:SingleFileMaxSize:singlefilemax}
 ${do:copy:LogType:logtype}
 ${do:copy:UseHTTPProxyForFTP:ftpprox}
 ${do:copy:ProxyType:proxytype}

html/server/step4.html

@@ -114,16 +114,16 @@ ${do:end-if}
 
 ${/* Real commands and ini file generated below */}
 
-<!-- engine commandline -->
+<!-- engine commandline; ztest so a cleared default-on option still emits its disabling flag -->
 ${do:output-mode:html}
 <textarea name="command" cols="50" rows="4" style="visibility:hidden">
 httrack \
 	--quiet \
 	--build-top-index \
 	${test:todo:--mirror:--mirror:--mirror-wizard:--get:--mirrorlinks:--testlinks:--continue:--update}
-	${urls}
-	${test:filelist:-%L "}${filelist}${test:filelist:"}
-	--path "${html:path}/${html:projname}" 
+	${unquoted:urls}
+	${test:filelist:-%L "}${arg:filelist}${test:filelist:"}
+	--path "${arg:path}/${arg:projname}" 
 \
 	${test:parseall:--near}
 	${test:link:--test}
@@ -131,7 +131,7 @@ httrack \
 	${test:htmlfirst::--priority=7}
 \
 ${do:if-not-empty:BuildString}
-	--structure "${BuildString}"
+	--structure "${arg:BuildString}"
 ${do:end-if}
 	${test:build:-N0:-N0:-N1:-N2:-N3:-N4:-N5:-N100:-N101:-N102:-N103:-N104:-N105:-N99:-N199:}
 \
@@ -150,51 +150,55 @@ ${do:end-if}
 	${test:travel3::--keep-links=0:--keep-links:--keep-links=3:--keep-links=4}
 	${test:windebug:--debug-headers}
 \
-	${test:connexion:--sockets=}${connexion}
+	${test:connexion:--sockets=}${unquoted:connexion}
 	${test:ka:--keep-alive}
-	${test:timeout:--timeout=}${timeout}
+	${test:timeout:--timeout=}${unquoted:timeout}
 	${test:remt:--host-control=1}
-	${test:retry:--retries=}${retry}
-	${test:rate:--min-rate=}${rate}
+	${test:retry:--retries=}${unquoted:retry}
+	${test:rate:--min-rate=}${unquoted:rate}
 	${test:rems:--host-control=2}
 \
-	${test:depth:--depth=}${depth}
-	${test:depth2:--ext-depth=}${depth2}
-	${test:maxhtml:--max-files=,}${maxhtml}
-	${test:othermax:--max-files=}${othermax}
-	${test:sizemax:--max-files=}${sizemax}
-	${test:pausebytes:--max-pause=}${pausebytes}
-	${test:maxtime:--max-time=}${maxtime}
-	${test:maxrate:--max-rate=}${maxrate}
-	${test:maxconn:--connection-per-second=}${maxconn}
-	${test:maxlinks:--advanced-maxlinks=}${maxlinks}
+	${test:depth:--depth=}${unquoted:depth}
+	${test:depth2:--ext-depth=}${unquoted:depth2}
+${/* -m<n> resets the html limit, so the bare form must precede the -m,<n> one */}
+	${test:othermax:--max-files=}${unquoted:othermax}
+	${test:maxhtml:--max-files=,}${unquoted:maxhtml}
+	${test:sizemax:--max-size=}${unquoted:sizemax}
+	${test:pausebytes:--max-pause=}${unquoted:pausebytes}
+	${test:maxtime:--max-time=}${unquoted:maxtime}
+	${test:maxrate:--max-rate=}${unquoted:maxrate}
+	${test:maxconn:--connection-per-second=}${unquoted:maxconn}
+	${test:maxlinks:--advanced-maxlinks=}${unquoted:maxlinks}
 \
-	--user-agent "${html:user}"
-	--footer "${html:footer}"
+	--user-agent "${arg:user}"
+	--footer "${arg:footer}"
 \
-	${url2}
+	${unquoted:url2}
 \
-	${test:cookies:--cookies=0:}
-	${test:parsejava:--parse-java=0:}
+	${ztest:cookies:--cookies=0:}
+	${ztest:parsejava:--parse-java=0:}
 	${test:updhack:--updatehack}
 	${test:urlhack:--urlhack=0:--urlhack}
 	${test:keepwww:--keep-www-prefix}
 	${test:keepslashes:--keep-double-slashes}
 	${test:keepqueryorder:--keep-query-order}
-	${test:cookiesfile:--cookies-file "}${html:cookiesfile}${test:cookiesfile:"}
-	${test:pausefiles:--pause "}${pausefiles}${test:pausefiles:"}
-	${test:stripquery:--strip-query "}${html:stripquery}${test:stripquery:"}
+	${test:cookiesfile:--cookies-file "}${arg:cookiesfile}${test:cookiesfile:"}
+	${test:pausefiles:--pause "}${arg:pausefiles}${test:pausefiles:"}
+	${test:stripquery:--strip-query "}${arg:stripquery}${test:stripquery:"}
 	${test:toler:--tolerant}
 	${test:http10:--http-10}
 	${test:cache2:--store-all-in-cache}
 	${test:warc:--warc}
-	${test:warcfile:--warc-file "}${html:warcfile}${test:warcfile:"}
+	${test:warcfile:--warc-file "}${arg:warcfile}${test:warcfile:"}
+	${test:changes:--changes}
+	${test:singlefile:--single-file}
+	${test:singlefilemax:--single-file-max-size=}${unquoted:singlefilemax}
 	${test:norecatch:--do-not-recatch}
 	${test:logf:--single-log}
 	${test:logtype:::--extra-log:--debug-log}
 	${test:index:--index=0:}
 	${test:index2:--search-index=0:--search-index}
-	${test:prox:--proxy "}${do:if-not-empty:prox}${test:proxytype::socks5:connect}${test:proxytype:\3A//}${do:end-if}${do:output-mode:html}${prox}${test:prox:\3A}${portprox}${test:prox:"}
+	${test:prox:--proxy "}${do:if-not-empty:prox}${test:proxytype::socks5:connect}${test:proxytype:\3A//}${do:end-if}${do:output-mode:html}${arg:prox}${test:prox:\3A}${arg:portprox}${test:prox:"}
 	${test:ftpprox:--httpproxy-ftp=0:--httpproxy-ftp}
 </textarea>
 
@@ -213,7 +217,7 @@ ParseAll=${ztest:parseall:0:1}
 HTMLFirst=${ztest:htmlfirst:0:1}
 Cache=${ztest:cache:0:1}
 NoRecatch=${ztest:norecatch:0:1}
-Dos=${dos
+Dos=${dos}
 Index=${ztest:index:0:1}
 WordIndex=${ztest:index2:0:1}
 Log=${ztest:logf:0:1:2}
@@ -241,6 +245,9 @@ StripQuery=${stripquery}
 StoreAllInCache=${ztest:cache2:0:1}
 Warc=${ztest:warc:0:1}
 WarcFile=${warcfile}
+Changes=${ztest:changes:0:1}
+SingleFile=${ztest:singlefile:0:1}
+SingleFileMaxSize=${singlefilemax}
 LogType=${logtype}
 UseHTTPProxyForFTP=${ztest:ftpprox:0:1}
 ProxyType=${proxytype}

lang.def

@@ -1042,3 +1042,15 @@ LANG_WARCFILE
 WARC archive name:
 LANG_WARCFILETIP
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
+LANG_CHANGES
+Report what changed since the previous mirror
+LANG_CHANGESTIP
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+LANG_SINGLEFILE
+Inline assets as data: URIs (self-contained pages)
+LANG_SINGLEFILETIP
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+LANG_SINGLEFILEMAX
+Largest inlined asset (bytes):
+LANG_SINGLEFILEMAXTIP
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.

lang/Bulgarian.txt

@@ -964,3 +964,15 @@ WARC archive name:
 Èìå íà WARC àðõèâà:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Íåçàäúëæèòåëíî áàçîâî èìå çà WARC àðõèâà; îñòàâåòå ïðàçíî çà àâòîìàòè÷íî èìåíóâàíå â èçõîäíàòà äèðåêòîðèÿ.
+Report what changed since the previous mirror
+Îò÷åò çà ïðîìåíèòå ñïðÿìî ïðåäèøíîòî îãëåäàëî
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Çàïèñâàíå è íà hts-changes.json ñúñ ñïèñúê íà íîâèòå, ïðîìåíåíèòå, íåïðîìåíåíèòå è èç÷åçíàëèòå ôàéëîâå ñïðÿìî ïðåäèøíîòî îãëåäàëî.
+Inline assets as data: URIs (self-contained pages)
+Âãðàæäàíå íà ðåñóðñèòå êàòî data: URI (ñàìîñòîÿòåëíè ñòðàíèöè)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Ñëåä çàâúðøâàíå íà îãëåäàëîòî âñÿêà çàïàçåíà ñòðàíèöà ñå ïðåçàïèñâà ñ âãðàäåíè ñòèëîâå, ñêðèïòîâå, èçîáðàæåíèÿ è øðèôòîâå, òàêà ÷å ñòðàíèöàòà äà ìîæå äà ñå îòâîðè è ñàìîñòîÿòåëíî.
+Largest inlined asset (bytes):
+Íàé-ãîëÿì âãðàäåí ðåñóðñ (áàéòîâå):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Ðåñóðñ íàä òîçè ðàçìåð çàïàçâà îáèêíîâåíà âðúçêà; îñòàâåòå ïðàçíî çà ñòîéíîñòòà ïî ïîäðàçáèðàíå îò 10485760 áàéòà.

lang/Castellano.txt

@@ -964,3 +964,15 @@ WARC archive name:
 Nombre del archivo WARC:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Nombre base opcional para el archivo WARC; déjelo en blanco para nombrarlo automáticamente en el directorio de salida.
+Report what changed since the previous mirror
+Informar de los cambios desde la copia anterior
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Escribir también hts-changes.json con la lista de lo que esta captura deja como nuevo, modificado, sin cambios o desaparecido respecto a la copia anterior.
+Inline assets as data: URIs (self-contained pages)
+Incrustar los recursos como URI data: (páginas autónomas)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Una vez terminada la copia, reescribir cada página guardada con sus hojas de estilo, scripts, imágenes y fuentes incrustadas, de modo que una página también pueda abrirse por sí sola.
+Largest inlined asset (bytes):
+Tamaño máximo del recurso incrustado (bytes):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Un recurso mayor que este tamaño conserva un enlace normal; déjelo vacío para el valor predeterminado de 10485760 bytes.

lang/Cesky.txt

@@ -964,3 +964,15 @@ WARC archive name:
 Název archivu WARC:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Volitelný základní název archivu WARC; ponechte prázdné pro automatické pojmenování ve výstupním adresáøi.
+Report what changed since the previous mirror
+Hlásit, co se od pøedchozího zrcadlení zmìnilo
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Zapsat také hts-changes.json se seznamem toho, co je oproti pøedchozímu zrcadlení nové, zmìnìné, nezmìnìné nebo chybìjící.
+Inline assets as data: URIs (self-contained pages)
+Vložit zdroje jako URI data: (samostatné stránky)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Po dokonèení zrcadlení pøepsat každou uloženou stránku s vloženými styly, skripty, obrázky a písmy, aby ji bylo možné otevøít i samostatnì.
+Largest inlined asset (bytes):
+Nejvìtší vložený zdroj (bajty):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Zdroj vìtší než tato velikost si ponechá bìžný odkaz; ponechte prázdné pro výchozí hodnotu 10485760 bajtù.

lang/Chinese-BIG5.txt

@@ -964,3 +964,15 @@ WARC archive name:
 WARC 封存檔名稱：
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 WARC 封存檔的選用基本名稱；留空則於輸出目錄中自動命名。
+Report what changed since the previous mirror
+回報自上次鏡射以來的變更
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+同時寫入 hts-changes.json，列出這次擷取相對於上次鏡射的新增、變更、未變更與消失的項目。
+Inline assets as data: URIs (self-contained pages)
+將資源內嵌為 data: URI（獨立網頁）
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+鏡射完成後，重寫每個已儲存的網頁，將樣式表、指令碼、圖片與字型內嵌其中，讓網頁也能單獨開啟。
+Largest inlined asset (bytes):
+內嵌資源大小上限（位元組）：
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+超過此大小的資源會保留一般連結；留空則使用預設的 10485760 位元組。

lang/Chinese-Simplified.txt

@@ -964,3 +964,15 @@ WARC archive name:
 WARC 归档名称：
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 WARC 归档的可选基本名称；留空则在输出目录中自动命名。
+Report what changed since the previous mirror
+报告自上次镜像以来的变更
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+同时写入 hts-changes.json，列出本次抓取相对于上次镜像的新增、更改、未更改和消失的项目。
+Inline assets as data: URIs (self-contained pages)
+将资源内嵌为 data: URI（独立网页）
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+镜像完成后，重写每个已保存的网页，将样式表、脚本、图片和字体内嵌其中，使网页也能单独打开。
+Largest inlined asset (bytes):
+内嵌资源大小上限（字节）：
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+超过此大小的资源会保留普通链接；留空则使用默认的 10485760 字节。

lang/Croatian.txt

@@ -966,3 +966,15 @@ WARC archive name:
 Naziv WARC arhive:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Neobavezni osnovni naziv WARC arhive; ostavite prazno za automatsko imenovanje u izlaznom direktoriju.
+Report what changed since the previous mirror
+Prijavi ¹to se promijenilo od prethodnog zrcaljenja
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Zapi¹i i hts-changes.json s popisom onoga ¹to je u odnosu na prethodno zrcaljenje novo, promijenjeno, nepromijenjeno ili nestalo.
+Inline assets as data: URIs (self-contained pages)
+Ugradi resurse kao data: URI-je (samostalne stranice)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Nakon dovr¹etka zrcaljenja ponovno zapi¹i svaku spremljenu stranicu s ugraðenim stilskim listovima, skriptama, slikama i fontovima, tako da se stranica mo¾e otvoriti i zasebno.
+Largest inlined asset (bytes):
+Najveæi ugraðeni resurs (bajtovi):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Resurs veæi od ove velièine zadr¾ava obiènu poveznicu; ostavite prazno za zadanih 10485760 bajtova.

lang/Dansk.txt

@@ -1012,3 +1012,15 @@ WARC archive name:
 Navn på WARC-arkiv:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Valgfrit basisnavn til WARC-arkivet; lad feltet stå tomt for automatisk navngivning i outputmappen.
+Report what changed since the previous mirror
+Rapportér hvad der er ændret siden den forrige spejling
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Skriv også hts-changes.json med en liste over, hvad denne gennemgang efterlader som nyt, ændret, uændret eller forsvundet i forhold til den forrige spejling.
+Inline assets as data: URIs (self-contained pages)
+Indlejr ressourcer som data:-URI'er (selvstændige sider)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Når spejlingen er færdig, omskrives hver gemt side med dens typografiark, scripts, billeder og skrifttyper indlejret, så en side også kan åbnes alene.
+Largest inlined asset (bytes):
+Største indlejrede ressource (byte):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+En ressource over denne størrelse beholder et almindeligt link; lad feltet stå tomt for standardværdien på 10485760 byte.

lang/Deutsch.txt

@@ -964,3 +964,15 @@ WARC archive name:
 Name des WARC-Archivs:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Optionaler Basisname für das WARC-Archiv; leer lassen, um es automatisch im Ausgabeverzeichnis zu benennen.
+Report what changed since the previous mirror
+Melden, was sich seit der vorherigen Spiegelung geändert hat
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Zusätzlich hts-changes.json schreiben, das auflistet, was dieser Durchlauf gegenüber der vorherigen Spiegelung als neu, geändert, unverändert oder verschwunden hinterlässt.
+Inline assets as data: URIs (self-contained pages)
+Ressourcen als data:-URIs einbetten (eigenständige Seiten)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Nach Abschluss der Spiegelung jede gespeicherte Seite mit eingebetteten Stylesheets, Skripten, Bildern und Schriftarten neu schreiben, sodass eine Seite auch für sich allein geöffnet werden kann.
+Largest inlined asset (bytes):
+Größte eingebettete Ressource (Bytes):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Eine Ressource über dieser Größe behält einen gewöhnlichen Link; leer lassen für den Standardwert von 10485760 Bytes.

lang/Eesti.txt

@@ -964,3 +964,15 @@ WARC archive name:
 WARC-arhiivi nimi:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 WARC-arhiivi valikuline põhinimi; jäta tühjaks, et see väljundkataloogis automaatselt nimetada.
+Report what changed since the previous mirror
+Teata, mis on eelmisest peegeldusest saadik muutunud
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Kirjuta ka hts-changes.json, mis loetleb, mis on võrreldes eelmise peegeldusega uus, muutunud, muutumatu või kadunud.
+Inline assets as data: URIs (self-contained pages)
+Manusta ressursid data: URI-dena (iseseisvad lehed)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Pärast peegelduse lõppu kirjuta iga salvestatud leht ümber, manustades selle laaditabelid, skriptid, pildid ja fondid, nii et lehte saab avada ka eraldi.
+Largest inlined asset (bytes):
+Suurim manustatud ressurss (baiti):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Sellest suurem ressurss säilitab tavalise lingi; jäta tühjaks vaikeväärtuse 10485760 baiti jaoks.

lang/English.txt

@@ -1012,3 +1012,15 @@ WARC archive name:
 WARC archive name:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
+Report what changed since the previous mirror
+Report what changed since the previous mirror
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Inline assets as data: URIs (self-contained pages)
+Inline assets as data: URIs (self-contained pages)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Largest inlined asset (bytes):
+Largest inlined asset (bytes):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.

lang/Finnish.txt

@@ -966,3 +966,15 @@ WARC archive name:
 WARC-arkiston nimi:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Valinnainen WARC-arkiston perusnimi; jätä tyhjäksi, jotta se nimetään automaattisesti tulostehakemistoon.
+Report what changed since the previous mirror
+Raportoi, mikä on muuttunut edellisen peilauksen jälkeen
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Kirjoita myös hts-changes.json, joka luettelee, mikä on edelliseen peilaukseen verrattuna uutta, muuttunutta, muuttumatonta tai kadonnutta.
+Inline assets as data: URIs (self-contained pages)
+Upota resurssit data:-URI-osoitteina (itsenäiset sivut)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Kun peilaus on valmis, kirjoita jokainen tallennettu sivu uudelleen niin, että sen tyylitiedostot, komentosarjat, kuvat ja fontit upotetaan, jolloin sivun voi avata myös yksinään.
+Largest inlined asset (bytes):
+Suurin upotettu resurssi (tavua):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Tätä suurempi resurssi säilyttää tavallisen linkin; jätä tyhjäksi, jolloin käytetään oletusarvoa 10485760 tavua.

lang/Francais.txt

@@ -1012,3 +1012,15 @@ WARC archive name:
 Nom de l'archive WARC :
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Nom de base optionnel pour l'archive WARC ; laissez vide pour le générer automatiquement dans le répertoire de sortie.
+Report what changed since the previous mirror
+Signaler ce qui a changé depuis le miroir précédent
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Écrire aussi hts-changes.json, qui liste ce que ce crawl laisse nouveau, modifié, inchangé ou disparu par rapport au miroir précédent.
+Inline assets as data: URIs (self-contained pages)
+Intégrer les ressources en URI data: (pages autonomes)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Une fois le miroir terminé, réécrire chaque page enregistrée avec ses feuilles de style, scripts, images et polices intégrés, afin qu'une page puisse aussi être ouverte seule.
+Largest inlined asset (bytes):
+Taille maximale d'une ressource intégrée (octets) :
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Au-delà de cette taille, la ressource reste un lien ordinaire ; laissez vide pour la valeur par défaut de 10485760 octets.

lang/Greek.txt

@@ -966,3 +966,15 @@ WARC archive name:
 Όνομα αρχείου WARC:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Προαιρετικό βασικό όνομα για το αρχείο WARC. Αφήστε το κενό για αυτόματη ονομασία στον κατάλογο εξόδου.
+Report what changed since the previous mirror
+Αναφορά των αλλαγών από το προηγούμενο αντίγραφο
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Εγγραφή και του hts-changes.json, που παραθέτει τι είναι νέο, τι άλλαξε, τι έμεινε ίδιο και τι χάθηκε σε σχέση με το προηγούμενο αντίγραφο.
+Inline assets as data: URIs (self-contained pages)
+Ενσωμάτωση των πόρων ως URI data: (αυτόνομες σελίδες)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Μόλις ολοκληρωθεί το αντίγραφο, κάθε αποθηκευμένη σελίδα ξαναγράφεται με ενσωματωμένα τα φύλλα στυλ, τα σενάρια, τις εικόνες και τις γραμματοσειρές της, ώστε η σελίδα να μπορεί να ανοίξει και μόνη της.
+Largest inlined asset (bytes):
+Μέγιστος ενσωματωμένος πόρος (byte):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Πόρος μεγαλύτερος από αυτό το μέγεθος διατηρεί κανονικό σύνδεσμο. Αφήστε το κενό για την προεπιλογή των 10485760 byte.

lang/Italiano.txt

@@ -964,3 +964,15 @@ WARC archive name:
 Nome dell'archivio WARC:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Nome di base facoltativo per l'archivio WARC; lascia vuoto per assegnarlo automaticamente nella directory di output.
+Report what changed since the previous mirror
+Segnala che cosa è cambiato dalla copia precedente
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Scrive anche hts-changes.json, che elenca ciò che questa scansione lascia come nuovo, modificato, invariato o scomparso rispetto alla copia precedente.
+Inline assets as data: URIs (self-contained pages)
+Incorpora le risorse come URI data: (pagine autonome)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Al termine della copia, riscrive ogni pagina salvata incorporando fogli di stile, script, immagini e caratteri, in modo che una pagina possa essere aperta anche da sola.
+Largest inlined asset (bytes):
+Dimensione massima della risorsa incorporata (byte):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Una risorsa oltre questa dimensione mantiene un collegamento normale; lasciare vuoto per il valore predefinito di 10485760 byte.

lang/Japanese.txt

@@ -964,3 +964,15 @@ WARC archive name:
 WARC アーカイブ名:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 WARC アーカイブの任意のベース名。空欄にすると出力ディレクトリ内で自動的に名前が付けられます。
+Report what changed since the previous mirror
+前回のミラーからの変更点を報告する
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+hts-changes.json も書き出し、前回のミラーと比べて新規、変更、変更なし、消滅となったものを一覧にします。
+Inline assets as data: URIs (self-contained pages)
+リソースを data: URI として埋め込む (単独で開けるページ)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+ミラーの完了後、保存された各ページを、スタイルシート、スクリプト、画像、フォントを埋め込んだ形で書き直します。ページ単体でも開けるようになります。
+Largest inlined asset (bytes):
+埋め込む最大サイズ (バイト):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+このサイズを超えるリソースは通常のリンクのままになります。空欄にすると既定値の 10485760 バイトになります。

lang/Macedonian.txt

@@ -964,3 +964,15 @@ WARC archive name:
 ¸ÜÕ ÝÐ WARC ÐàåØÒÐâÐ:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 ¸×ÑÞàÝÞ ÞáÝÞÒÝÞ ØÜÕ ×Ð WARC ÐàåØÒÐâÐ; ÞáâÐÒÕâÕ ßàÐ×ÝÞ ×Ð ÐÒâÞÜÐâáÚÞ ØÜÕÝãÒÐúÕ ÒÞ Ø×ÛÕ×ÝØÞâ ÔØàÕÚâÞàØãÜ.
+Report what changed since the previous mirror
+¸×ÒÕáâØ èâÞ Õ ßàÞÜÕÝÕâÞ ÞÔ ßàÕâåÞÔÝÞâÞ ÞÓÛÕÔÐÛÞ
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+·ÐßØèØ Ø hts-changes.json áÞ áßØáÞÚ ÝÐ âÞÐ èâÞ Õ ÝÞÒÞ, ßàÞÜÕÝÕâÞ, ÝÕßàÞÜÕÝÕâÞ ØÛØ ØáçÕ×ÝÐâÞ ÒÞ ÞÔÝÞá ÝÐ ßàÕâåÞÔÝÞâÞ ÞÓÛÕÔÐÛÞ.
+Inline assets as data: URIs (self-contained pages)
+²ÓàÐÔØ ÓØ àÕáãàáØâÕ ÚÐÚÞ data: URI (áÐÜÞáâÞøÝØ áâàÐÝØæØ)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+¿Þ ×ÐÒàèãÒÐúÕ ÝÐ ÞÓÛÕÔÐÛÞâÞ, áÕÚÞøÐ ×ÐçãÒÐÝÐ áâàÐÝØæÐ áÕ ßàÕßØèãÒÐ áÞ ÒÓàÐÔÕÝØ áâØÛáÚØ ÛØáâÞÒØ, áÚàØßâØ, áÛØÚØ Ø äÞÝâÞÒØ, ×Ð ÔÐ ÜÞÖÕ áâàÐÝØæÐâÐ ÔÐ áÕ ÞâÒÞàØ Ø áÐÜÞáâÞøÝÞ.
+Largest inlined asset (bytes):
+½ÐøÓÞÛÕÜ ÒÓàÐÔÕÝ àÕáãàá (ÑÐøâØ):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+ÀÕáãàá ßÞÓÞÛÕÜ ÞÔ ÞÒÐÐ ÓÞÛÕÜØÝÐ ×ÐÔàÖãÒÐ ÞÑØçÝÐ ÒàáÚÐ; ÞáâÐÒÕâÕ ßàÐ×ÝÞ ×Ð áâÐÝÔÐàÔÝØâÕ 10485760 ÑÐøâØ.

lang/Magyar.txt

@@ -964,3 +964,15 @@ WARC archive name:
 WARC archívum neve:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 A WARC archívum opcionális alapneve; hagyja üresen az automatikus elnevezéshez a kimeneti könyvtárban.
+Report what changed since the previous mirror
+Jelentés arról, mi változott az elõzõ tükrözés óta
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+A hts-changes.json fájl írása is, amely felsorolja, mi új, mi változott, mi maradt változatlan és mi tûnt el az elõzõ tükrözéshez képest.
+Inline assets as data: URIs (self-contained pages)
+Erõforrások beágyazása data: URI-ként (önálló oldalak)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+A tükrözés befejezése után minden mentett oldal újraírása a stíluslapok, parancsfájlok, képek és betûtípusok beágyazásával, így az oldal önmagában is megnyitható.
+Largest inlined asset (bytes):
+Legnagyobb beágyazott erõforrás (bájt):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Az ennél nagyobb erõforrás közönséges hivatkozás marad; hagyja üresen a 10485760 bájtos alapértelmezéshez.

lang/Nederlands.txt

@@ -964,3 +964,15 @@ WARC archive name:
 Naam van WARC-archief:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Optionele basisnaam voor het WARC-archief; laat leeg om het automatisch een naam te geven in de uitvoermap.
+Report what changed since the previous mirror
+Rapporteren wat er sinds de vorige spiegeling is gewijzigd
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Ook hts-changes.json schrijven met een overzicht van wat deze doorloop nieuw, gewijzigd, ongewijzigd of verdwenen laat ten opzichte van de vorige spiegeling.
+Inline assets as data: URIs (self-contained pages)
+Bronnen insluiten als data:-URI's (op zichzelf staande pagina's)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Zodra de spiegeling klaar is, elke opgeslagen pagina herschrijven met ingesloten stijlbladen, scripts, afbeeldingen en lettertypen, zodat een pagina ook los geopend kan worden.
+Largest inlined asset (bytes):
+Grootste ingesloten bron (bytes):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Een bron boven deze grootte houdt een gewone koppeling; laat leeg voor de standaardwaarde van 10485760 bytes.

lang/Norsk.txt

@@ -964,3 +964,15 @@ WARC archive name:
 Navn på WARC-arkiv:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Valgfritt basisnavn for WARC-arkivet; la feltet stå tomt for automatisk navngivning i utdatamappen.
+Report what changed since the previous mirror
+Rapporter hva som er endret siden forrige speiling
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Skriv også hts-changes.json som lister opp hva denne gjennomgangen etterlater som nytt, endret, uendret eller forsvunnet sammenlignet med forrige speiling.
+Inline assets as data: URIs (self-contained pages)
+Bygg inn ressurser som data:-URI-er (selvstendige sider)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Når speilingen er ferdig, skrives hver lagrede side om med stilark, skript, bilder og skrifter innebygd, slik at en side også kan åpnes alene.
+Largest inlined asset (bytes):
+Største innebygde ressurs (byte):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+En ressurs over denne størrelsen beholder en vanlig lenke; la feltet stå tomt for standardverdien på 10485760 byte.

lang/Polski.txt

@@ -964,3 +964,15 @@ WARC archive name:
 Nazwa archiwum WARC:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Opcjonalna nazwa bazowa archiwum WARC; pozostaw puste, aby nazwaæ je automatycznie w katalogu wyj¶ciowym.
+Report what changed since the previous mirror
+Zg³o¶, co zmieni³o siê od poprzedniej kopii
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Zapisz równie¿ plik hts-changes.json z list± tego, co w porównaniu z poprzedni± kopi± jest nowe, zmienione, niezmienione lub usuniête.
+Inline assets as data: URIs (self-contained pages)
+Osad¼ zasoby jako identyfikatory data: URI (samodzielne strony)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Po zakoñczeniu tworzenia kopii przepisz ka¿d± zapisan± stronê z osadzonymi arkuszami stylów, skryptami, obrazami i czcionkami, aby stronê mo¿na by³o otworzyæ równie¿ samodzielnie.
+Largest inlined asset (bytes):
+Najwiêkszy osadzony zasób (bajty):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Zasób wiêkszy ni¿ ten rozmiar zachowuje zwyk³y odno¶nik; pozostaw puste, aby u¿yæ domy¶lnych 10485760 bajtów.

lang/Portugues-Brasil.txt

@@ -1012,3 +1012,15 @@ WARC archive name:
 Nome do arquivo WARC:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Nome base opcional para o arquivo WARC; deixe em branco para nomeá-lo automaticamente no diretório de saída.
+Report what changed since the previous mirror
+Relatar o que mudou desde o espelhamento anterior
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Gravar também hts-changes.json listando o que esta captura deixa como novo, alterado, inalterado ou removido em relação ao espelhamento anterior.
+Inline assets as data: URIs (self-contained pages)
+Incorporar os recursos como URIs data: (páginas autônomas)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Ao terminar o espelhamento, reescrever cada página salva com suas folhas de estilo, scripts, imagens e fontes incorporadas, para que a página também possa ser aberta sozinha.
+Largest inlined asset (bytes):
+Maior recurso incorporado (bytes):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Um recurso acima desse tamanho mantém um link comum; deixe em branco para o padrão de 10485760 bytes.

lang/Portugues.txt

@@ -964,3 +964,15 @@ WARC archive name:
 Nome do arquivo WARC:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Nome base opcional para o arquivo WARC; deixe em branco para o nomear automaticamente no diretório de saída.
+Report what changed since the previous mirror
+Comunicar o que mudou desde o espelho anterior
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Escrever também hts-changes.json, que lista o que esta recolha deixa como novo, alterado, inalterado ou desaparecido em relação ao espelho anterior.
+Inline assets as data: URIs (self-contained pages)
+Incorporar os recursos como URI data: (páginas autónomas)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Depois de terminado o espelho, reescrever cada página guardada com as suas folhas de estilo, scripts, imagens e tipos de letra incorporados, para que a página também possa ser aberta sozinha.
+Largest inlined asset (bytes):
+Maior recurso incorporado (bytes):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Um recurso acima deste tamanho mantém uma ligação normal; deixe em branco para o valor predefinido de 10485760 bytes.

lang/Romanian.txt

@@ -964,3 +964,15 @@ WARC archive name:
 Numele arhivei WARC:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Nume de baza optional pentru arhiva WARC; lasati gol pentru a-l denumi automat in directorul de iesire.
+Report what changed since the previous mirror
+Raporteaza ce s-a schimbat fata de copia anterioara
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Scrie si hts-changes.json, care listeaza ce este nou, modificat, nemodificat sau disparut fata de copia anterioara.
+Inline assets as data: URIs (self-contained pages)
+Încorporeaza resursele ca URI data: (pagini de sine statatoare)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Dupa terminarea copiei, fiecare pagina salvata este rescrisa cu foile de stil, scripturile, imaginile si fonturile încorporate, astfel încât pagina sa poata fi deschisa si singura.
+Largest inlined asset (bytes):
+Cea mai mare resursa încorporata (octeti):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+O resursa mai mare decât aceasta dimensiune pastreaza o legatura obisnuita; lasati gol pentru valoarea implicita de 10485760 de octeti.

lang/Russian.txt

@@ -964,3 +964,15 @@ WARC archive name:
 Èìÿ WARC-àðõèâà:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Íåîáÿçàòåëüíîå áàçîâîå èìÿ WARC-àðõèâà; îñòàâüòå ïóñòûì äëÿ àâòîìàòè÷åñêîãî èìåíîâàíèÿ â âûõîäíîì êàòàëîãå.
+Report what changed since the previous mirror
+Ñîîáùàòü, ÷òî èçìåíèëîñü ñ ïðåäûäóùåãî çåðêàëà
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Òàêæå çàïèñûâàòü hts-changes.json ñî ñïèñêîì òîãî, ÷òî ïî ñðàâíåíèþ ñ ïðåäûäóùèì çåðêàëîì ñòàëî íîâûì, èçìåí¸ííûì, íåèçìåí¸ííûì èëè èñ÷åçëî.
+Inline assets as data: URIs (self-contained pages)
+Âñòðàèâàòü ðåñóðñû êàê data: URI (ñàìîñòîÿòåëüíûå ñòðàíèöû)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Ïîñëå çàâåðøåíèÿ çåðêàëèðîâàíèÿ êàæäàÿ ñîõðàí¸ííàÿ ñòðàíèöà ïåðåçàïèñûâàåòñÿ ñî âñòðîåííûìè òàáëèöàìè ñòèëåé, ñöåíàðèÿìè, èçîáðàæåíèÿìè è øðèôòàìè, ÷òîáû ñòðàíèöó ìîæíî áûëî îòêðûòü îòäåëüíî.
+Largest inlined asset (bytes):
+Íàèáîëüøèé âñòðàèâàåìûé ðåñóðñ (áàéòû):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Ðåñóðñ áîëüøå ýòîãî ðàçìåðà ñîõðàíÿåò îáû÷íóþ ññûëêó; îñòàâüòå ïóñòûì äëÿ çíà÷åíèÿ ïî óìîë÷àíèþ 10485760 áàéò.

lang/Slovak.txt

@@ -964,3 +964,15 @@ WARC archive name:
 Názov archívu WARC:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Voliteµný základný názov archívu WARC; ponechajte prázdne pre automatické pomenovanie vo výstupnom adresári.
+Report what changed since the previous mirror
+Oznámi», èo sa od predchádzajúceho zrkadlenia zmenilo
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Zapísa» aj hts-changes.json so zoznamom toho, èo je oproti predchádzajúcemu zrkadleniu nové, zmenené, nezmenené alebo chýbajúce.
+Inline assets as data: URIs (self-contained pages)
+Vlo¾i» zdroje ako URI data: (samostatné stránky)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Po dokonèení zrkadlenia prepísa» ka¾dú ulo¾enú stránku s vlo¾enými ¹týlmi, skriptmi, obrázkami a písmami, aby sa stránka dala otvori» aj samostatne.
+Largest inlined asset (bytes):
+Najväè¹í vlo¾ený zdroj (bajty):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Zdroj väè¹í ne¾ táto veµkos» si ponechá be¾ný odkaz; ponechajte prázdne pre predvolených 10485760 bajtov.

lang/Slovenian.txt

@@ -964,3 +964,15 @@ WARC archive name:
 Ime arhiva WARC:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Neobvezno osnovno ime arhiva WARC; pustite prazno za samodejno poimenovanje v izhodni mapi.
+Report what changed since the previous mirror
+Porocaj, kaj se je spremenilo od prejsnjega zrcaljenja
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Zapisi tudi hts-changes.json s seznamom tega, kar je v primerjavi s prejsnjim zrcaljenjem novo, spremenjeno, nespremenjeno ali izginilo.
+Inline assets as data: URIs (self-contained pages)
+Vgradi vire kot data: URI (samostojne strani)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Ko je zrcaljenje koncano, se vsaka shranjena stran prepise z vgrajenimi slogovnimi listi, skripti, slikami in pisavami, tako da jo je mogoce odpreti tudi samostojno.
+Largest inlined asset (bytes):
+Najvecji vgrajeni vir (bajti):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Vir, vecji od te velikosti, ohrani obicajno povezavo; pustite prazno za privzetih 10485760 bajtov.

lang/Svenska.txt

@@ -964,3 +964,15 @@ WARC archive name:
 WARC-arkivets namn:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Valfritt basnamn för WARC-arkivet; lämna tomt för att namnge det automatiskt i utdatakatalogen.
+Report what changed since the previous mirror
+Rapportera vad som ändrats sedan föregående spegling
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Skriv även hts-changes.json som listar vad den här genomgången lämnar som nytt, ändrat, oförändrat eller försvunnet jämfört med föregående spegling.
+Inline assets as data: URIs (self-contained pages)
+Bädda in resurser som data:-URI:er (fristående sidor)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+När speglingen är klar skrivs varje sparad sida om med sina formatmallar, skript, bilder och teckensnitt inbäddade, så att en sida också kan öppnas för sig.
+Largest inlined asset (bytes):
+Största inbäddade resurs (byte):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+En resurs över den här storleken behåller en vanlig länk; lämna tomt för standardvärdet 10485760 byte.

lang/Turkish.txt

@@ -964,3 +964,15 @@ WARC archive name:
 WARC arþivi adý:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 WARC arþivi için isteðe baðlý temel ad; çýktý dizininde otomatik adlandýrma için boþ býrakýn.
+Report what changed since the previous mirror
+Önceki yansýmadan bu yana deðiþenleri bildir
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Ayrýca hts-changes.json yazarak bu taramanýn önceki yansýmaya göre neyi yeni, deðiþmiþ, deðiþmemiþ veya kaybolmuþ býraktýðýný listele.
+Inline assets as data: URIs (self-contained pages)
+Kaynaklarý data: URI olarak göm (kendi kendine yeten sayfalar)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Yansýlama bittiðinde, kaydedilen her sayfa stil sayfalarý, betikleri, görselleri ve yazý tipleri gömülü olarak yeniden yazýlýr; böylece sayfa tek baþýna da açýlabilir.
+Largest inlined asset (bytes):
+En büyük gömülü kaynak (bayt):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Bu boyutun üzerindeki bir kaynak sýradan baðlantýsýný korur; 10485760 baytlýk varsayýlan için boþ býrakýn.

lang/Ukrainian.txt

@@ -964,3 +964,15 @@ WARC archive name:
 ²ì'ÿ WARC-àðõ³âó:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 Íåîáîâ'ÿçêîâà áàçîâà íàçâà WARC-àðõ³âó; çàëèøòå ïîðîæí³ì äëÿ àâòîìàòè÷íîãî íàéìåíóâàííÿ ó âèõ³äíîìó êàòàëîç³.
+Report what changed since the previous mirror
+Ïîâ³äîìëÿòè, ùî çì³íèëîñÿ ç ïîïåðåäíüîãî äçåðêàëà
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+Òàêîæ çàïèñóâàòè hts-changes.json ç³ ñïèñêîì òîãî, ùî ïîð³âíÿíî ç ïîïåðåäí³ì äçåðêàëîì º íîâèì, çì³íåíèì, íåçì³íåíèì àáî çíèêëèì.
+Inline assets as data: URIs (self-contained pages)
+Âáóäîâóâàòè ðåñóðñè ÿê data: URI (ñàìîñò³éí³ ñòîð³íêè)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+ϳñëÿ çàâåðøåííÿ äçåðêàëþâàííÿ êîæíà çáåðåæåíà ñòîð³íêà ïåðåçàïèñóºòüñÿ ç âáóäîâàíèìè òàáëèöÿìè ñòèë³â, ñöåíàð³ÿìè, çîáðàæåííÿìè òà øðèôòàìè, ùîá ñòîð³íêó ìîæíà áóëî â³äêðèòè îêðåìî.
+Largest inlined asset (bytes):
+Íàéá³ëüøèé âáóäîâàíèé ðåñóðñ (áàéòè):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Ðåñóðñ, á³ëüøèé çà öåé ðîçì³ð, çáåð³ãຠçâè÷àéíå ïîñèëàííÿ; çàëèøòå ïîðîæí³ì äëÿ òèïîâîãî çíà÷åííÿ 10485760 áàéò³â.

lang/Uzbek.txt

@@ -964,3 +964,15 @@ WARC archive name:
 WARC arxivi nomi:
 Optional base name for the WARC archive; leave blank to auto-name it under the output directory.
 WARC arxivi uchun ixtiyoriy asosiy nom; chiqish katalogida avtomatik nomlash uchun bo'sh qoldiring.
+Report what changed since the previous mirror
+Oldingi nusxadan beri nima o’zgarganini xabar qilish
+Also write hts-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror.
+hts-changes.json ham yoziladi: unda ushbu yig’ish oldingi nusxaga nisbatan nimani yangi, o’zgargan, o’zgarmagan yoki yo’qolgan holda qoldirgani ro’yxati bo’ladi.
+Inline assets as data: URIs (self-contained pages)
+Resurslarni data: URI sifatida joylash (mustaqil sahifalar)
+Once the mirror is finished, rewrite every saved page with its stylesheets, scripts, images and fonts embedded, so a page can also be opened on its own.
+Nusxalash tugagach, har bir saqlangan sahifa uslublar jadvallari, skriptlar, rasmlar va shriftlar joylangan holda qayta yoziladi, shunda sahifani alohida ham ochish mumkin.
+Largest inlined asset (bytes):
+Eng katta joylangan resurs (bayt):
+An asset above this size keeps an ordinary link; leave blank for the 10485760-byte default.
+Bu o’lchamdan katta resurs oddiy havolani saqlab qoladi; standart 10485760 bayt uchun bo’sh qoldiring.

libtest/callbacks-example-filenameiisbug.c

@@ -71,7 +71,9 @@ static int mysavename(t_hts_callbackarg * carg, httrackp * opt,
       for(j = 0; iisBogus[i][j] == a[j] && iisBogus[i][j] != '\0'; j++) ;
       if (iisBogus[i][j] == '\0'
           && (a[j] == '\0' || a[j] == '/' || a[j] == '\\')) {
-        strncpy(a, iisBogusReplace[i], strlen(iisBogusReplace[i]));
+        /* j bytes matched, so j fit: copying j cannot overrun whatever the
+           table holds, and the tail must survive untouched */
+        memcpy(a, iisBogusReplace[i], (size_t) j);
         break;
       }
     }

man/httrack.1

@@ -3,7 +3,7 @@
 .\"
 .\" This file is generated by man/makeman.sh; do not edit by hand.
 .\" SPDX-License-Identifier: GPL-3.0-or-later
-.TH httrack 1 "23 July 2026" "httrack website copier"
+.TH httrack 1 "26 July 2026" "httrack website copier"
 .SH NAME
 httrack \- offline browser : copy websites to a local directory
 .SH SYNOPSIS
@@ -40,6 +40,7 @@ httrack \- offline browser : copy websites to a local directory
 [ \fB\-%N, \-\-delayed\-type\-check\fR ]
 [ \fB\-%D, \-\-cached\-delayed\-type\-check\fR ]
 [ \fB\-%M, \-\-mime\-html\fR ]
+[ \fB\-%Z, \-\-single\-file\fR ]
 [ \fB\-LN, \-\-long\-names[=N]\fR ]
 [ \fB\-KN, \-\-keep\-links[=N]\fR ]
 [ \fB\-x, \-\-replace\-external\fR ]
@@ -75,6 +76,7 @@ httrack \- offline browser : copy websites to a local directory
 [ \fB\-C, \-\-cache[=N]\fR ]
 [ \fB\-k, \-\-store\-all\-in\-cache\fR ]
 [ \fB\-%r, \-\-warc\fR ]
+[ \fB\-%d, \-\-changes\fR ]
 [ \fB\-%n, \-\-do\-not\-recatch\fR ]
 [ \fB\-%v, \-\-display\fR ]
 [ \fB\-Q, \-\-do\-not\-log\fR ]
@@ -198,6 +200,8 @@ delayed type check, don't make any link test but wait for files download to star
 cached delayed type check, don't wait for remote type during updates, to speedup them (%D0 wait, * %D1 don't wait) (\-\-cached\-delayed\-type\-check)
 .IP \-%M
 generate a RFC MIME\-encapsulated full\-archive (.mht) (\-\-mime\-html)
+.IP \-%Z
+after the mirror, rewrite each saved page with its stylesheets, scripts, images and fonts inlined as data: URIs, so any page opens by double\-click anywhere (links between pages stay relative; audio and video stay links); \-\-single\-file\-max\-size N caps each asset (default 10485760 bytes). %M is the better container where a Chromium\-family browser is a given: one archive, no base64 tax on text, a shared asset stored once (\-\-single\-file)
 .IP \-%t
 keep the original file extension, don't rewrite it from the MIME type (%t0 rewrite)
 .IP \-LN
@@ -279,6 +283,8 @@ create/use a cache for updates and retries (C0 no cache,C1 cache is prioritary,*
 store all files in cache (not useful if files on disk) (\-\-store\-all\-in\-cache)
 .IP \-%r
 write an ISO\-28500 WARC/1.1 archive; \-\-warc\-file NAME sets the output name, \-\-warc\-max\-size N rotates segments past N bytes, \-\-warc\-cdx also writes a sorted CDXJ index, \-\-wacz packages it all as a WACZ file (\-\-warc)
+.IP \-%d
+write hts\-changes.json listing what this crawl left new, changed, unchanged and gone compared to the previous mirror (\-\-changes)
 .IP \-%n
 do not re\-download locally erased files (\-\-do\-not\-recatch)
 .IP \-%v

src/Makefile.am

@@ -32,7 +32,9 @@ AM_LDFLAGS = \
 
 bin_PROGRAMS = proxytrack httrack htsserver
 
-httrack_LDADD = $(THREADS_LIBS) libhttrack.la
+httrack_SOURCES = httrack.c htsbacktrace.c htsbacktrace.h
+# $(DL_LIBS): dladdr() in the crash handler, still in libdl on pre-2.34 glibc.
+httrack_LDADD = $(THREADS_LIBS) $(DL_LIBS) libhttrack.la
 htsserver_LDADD = $(THREADS_LIBS) $(SOCKET_LIBS) libhttrack.la
 proxytrack_LDADD = $(THREADS_LIBS) $(SOCKET_LIBS)
 
@@ -46,7 +48,8 @@ htsserver_LDFLAGS = $(AM_LDFLAGS) $(LDFLAGS_PIE)
 
 lib_LTLIBRARIES = libhttrack.la
 
-htsserver_SOURCES = htsserver.c htsserver.h htsweb.c htsweb.h \
+htsserver_SOURCES = htsserver.c htsserver.h htsweb.c htsweb.h htsstats.h \
+	htscmdline.c htscmdline.h \
 	htsurlport.c htsurlport.h
 proxytrack_SOURCES = proxy/main.c \
 	proxy/proxytrack.c proxy/store.c \
@@ -60,21 +63,21 @@ whttrackrun_SCRIPTS = webhttrack
 libhttrack_la_SOURCES =  htscore.c htsparse.c htsback.c htscache.c \
 	htscache_selftest.c htsdns_selftest.c htsselftest.c \
 	htscatchurl.c htsfilters.c htsftp.c htshash.c coucal/coucal.c \
-	htshelp.c htslib.c htsurlport.c htscoremain.c \
+	htscmdline.c htshelp.c htslib.c htsurlport.c htscoremain.c \
 	htsname.c htsrobots.c htstools.c htswizard.c \
 	htsalias.c htsthread.c htsindex.c htsbauth.c \
-	htsmd5.c htscodec.c htswarc.c htsproxy.c htszlib.c htswrap.c htsconcat.c \
+	htsmd5.c htscodec.c htswarc.c htschanges.c htssinglefile.c htsproxy.c htszlib.c htswrap.c htsconcat.c \
 	htsmodules.c htscharset.c punycode.c htsencoding.c htssniff.c \
 	md5.c \
 	minizip/ioapi.c minizip/mztools.c minizip/unzip.c minizip/zip.c \
 	hts-indextmpl.h htsalias.h htsback.h htsbase.h htssafe.h \
 	htsbasenet.h htsbauth.h htscache.h htscache_selftest.h htsdns_selftest.h htsselftest.h htscatchurl.h  \
-	htsconfig.h htscore.h htsparse.h htscoremain.h htsdefines.h  \
+	htscmdline.h htsconfig.h htscore.h htsparse.h htscoremain.h htsdefines.h  \
 	htsfilters.h htsftp.h htsglobal.h htshash.h coucal/coucal.h \
 	htshelp.h htsindex.h htslib.h htsurlport.h htsmd5.h \
 	htsmodules.h htsname.h htsnet.h htssniff.h \
 	htsopt.h htsrobots.h htsthread.h  \
-	htstools.h htswizard.h htswrap.h htscodec.h htswarc.h htsproxy.h htszlib.h  \
+	htstools.h htswizard.h htswrap.h htscodec.h htswarc.h htschanges.h htssinglefile.h htsproxy.h htszlib.h  \
 	htsstrings.h htsarrays.h httrack-library.h \
 	htscharset.h punycode.h htsencoding.h \
 	htsentities.h htsentities.sh htsbasiccharsets.sh htscodepages.h \
@@ -84,7 +87,7 @@ libhttrack_la_LIBADD = $(THREADS_LIBS) $(ZLIB_LIBS) $(BROTLI_LIBS) $(ZSTD_LIBS)
 libhttrack_la_CFLAGS = $(AM_CFLAGS) -DLIBHTTRACK_EXPORTS -DZLIB_CONST
 libhttrack_la_LDFLAGS = $(AM_LDFLAGS) -version-info $(VERSION_INFO)
 
-EXTRA_DIST = httrack.h webhttrack \
+EXTRA_DIST = httrack.h htsstats.h webhttrack \
 		version.rc \
 		libhttrack.rc \
 		httrack.rc \

src/htsalias.c

@@ -114,6 +114,8 @@ const char *hts_optalias[][4] = {
    "strip [host/pattern=]key1,key2,... from URLs"},
   {"cookies-file", "-%K", "param1",
    "load extra cookies from a Netscape cookies.txt"},
+  {"changes", "-%d", "single",
+   "write hts-changes.json: what this crawl changed vs. the previous mirror"},
   {"warc", "-%r", "single", "write an ISO-28500 WARC/1.1 archive of the crawl"},
   {"warc-file", "-%rf", "param1", "write a WARC archive to the given base name"},
   {"warc-max-size", "-%rs", "param1",
@@ -123,6 +125,10 @@ const char *hts_optalias[][4] = {
   {"warc-cdxj", "-%rc", "single", ""},
   {"wacz", "-%rz", "single",
    "package the WARC archive, CDXJ index and pages as a WACZ file"},
+  {"single-file", "-%Z", "single",
+   "after the mirror, inline each page's assets as data: URIs"},
+  {"single-file-max-size", "-%Zs", "param1",
+   "per-asset cap for --single-file, in bytes (implies it; default 10485760)"},
   {"why", "-%Y", "param1",
    "explain which filter rule accepts or rejects a URL, then exit"},
   {"pause", "-%G", "param1",

src/htsback.c

@@ -38,6 +38,7 @@ Please visit our Website: http://www.httrack.com
 #include "htsnet.h"
 #include "htscore.h"
 #include "htswarc.h"
+#include "htschanges.h"
 #include "htsthread.h"
 #include <time.h>
 /* END specific definitions */
@@ -125,6 +126,22 @@ void back_free(struct_back ** sback) {
    above a normal handshake. The last candidate still gets the full timeout. */
 #define HTS_CONNECT_FALLBACK_TIMEOUT 10
 
+void back_read_ftp_result(FILE *fp, htsblk *r) {
+  size_t j = 0;
+
+  if (fscanf(fp, "%d ", &r->statuscode) != 1)
+    r->statuscode = STATUSCODE_INVALID;
+  // an external helper writes this file: stop at capacity, not at EOF
+  while (j + 1 < sizeof(r->msg)) {
+    const int c = fgetc(fp);
+
+    if (c == EOF)
+      break;
+    r->msg[j++] = (char) c;
+  }
+  r->msg[j] = '\0';
+}
+
 int back_connect_fallback_due(int addr_index, int addr_count, int elapsed,
                               int timeout) {
   int deadline;
@@ -542,8 +559,14 @@ static int create_back_tmpfile(httrackp *opt, lien_back *const back,
   // do not use tempnam() but a regular filename
   back->tmpfile_buffer[0] = '\0';
   if (back->url_sav[0] != '\0') {
-    snprintf(back->tmpfile_buffer, sizeof(back->tmpfile_buffer), "%s.%s",
-             back->url_sav, ext);
+    /* same capacity as url_sav, so truncation drops the extension and aliases
+       the temp name onto the live file that back_finalize_backup() UNLINKs */
+    if (!sprintfbuff(back->tmpfile_buffer, "%s.%s", back->url_sav, ext)) {
+      hts_log_print(opt, LOG_WARNING, "temporary filename too long for %s",
+                    back->url_sav);
+      back->tmpfile_buffer[0] = '\0';
+      return -1;
+    }
     back->tmpfile = back->tmpfile_buffer;
     if (structcheck(back->tmpfile) != 0) {
       hts_log_print(opt, LOG_WARNING, "can not create directory to %s",
@@ -551,8 +574,15 @@ static int create_back_tmpfile(httrackp *opt, lien_back *const back,
       return -1;
     }
   } else {
-    snprintf(back->tmpfile_buffer, sizeof(back->tmpfile_buffer), "%s/tmp%d.%s",
-             StringBuff(opt->path_html_utf8), opt->state.tmpnameid++, ext);
+    /* truncation here would collide distinct tmpnameid's onto one name */
+    if (!sprintfbuff(back->tmpfile_buffer, "%s/tmp%d.%s",
+                     StringBuff(opt->path_html_utf8), opt->state.tmpnameid++,
+                     ext)) {
+      hts_log_print(opt, LOG_WARNING, "temporary filename too long in %s",
+                    StringBuff(opt->path_html_utf8));
+      back->tmpfile_buffer[0] = '\0';
+      return -1;
+    }
     back->tmpfile = back->tmpfile_buffer;
   }
   /* OK */
@@ -699,6 +729,13 @@ int back_finalize(httrackp * opt, cache_back * cache, struct_back * sback,
                 if ((size = hts_codec_unpack(codec, back[p].tmpfile,
                                              unpacked)) >= 0) {
                   back[p].r.size = back[p].r.totalsize = size;
+                  if (back[p].r.is_write) {
+                    /* Sample the previous copy now: the rename below replaces
+                       it, and file_notify() only fires once it is gone. */
+                    hts_changes_notify(
+                        opt, back[p].url_adr, back[p].url_fil, back[p].url_sav,
+                        HTS_TRUE, back[p].r.notmodified ? HTS_TRUE : HTS_FALSE);
+                  }
                   if (!back[p].r.is_write) {
                     // fichier -> mémoire ; le fichier est écrit plus tard
                     deleteaddr(&back[p].r);
@@ -2947,7 +2984,7 @@ void back_wait(struct_back * sback, httrackp * opt, cache_back * cache,
               back[i].r.msg[0] = '\0';
               strncatbuff(back[i].r.msg, tmp, sizeof(back[i].r.msg) - 2);
               if (!strnotempty(back[i].r.msg)) {
-                sprintf(back[i].r.msg, "SSL/TLS error %d", err_code);
+                htsblk_failf(&back[i].r, "SSL/TLS error %d", err_code);
               }
               deletehttp(&back[i].r);
               back[i].r.soc = INVALID_SOCKET;
@@ -3020,16 +3057,7 @@ void back_wait(struct_back * sback, httrackp * opt, cache_back * cache,
             FOPEN(fconcat(OPT_GET_BUFF(opt), back[i].location_buffer, ".ok"),
                   "rb");
           if (fp) {
-            int j = 0;
-
-            fscanf(fp, "%d ", &(back[i].r.statuscode));
-            while(!feof(fp)) {
-              int c = fgetc(fp);
-
-              if (c != EOF)
-                back[i].r.msg[j++] = c;
-            }
-            back[i].r.msg[j++] = '\0';
+            back_read_ftp_result(fp, &back[i].r);
             fclose(fp);
             UNLINK(fconcat(OPT_GET_BUFF(opt), back[i].location_buffer, ".ok"));
             strcpybuff(fconcat
@@ -3330,10 +3358,11 @@ void back_wait(struct_back * sback, httrackp * opt, cache_back * cache,
                   deleteaddr(&back[i].r);
                   if (back[i].r.size < back[i].r.totalsize)
                     back[i].r.statuscode = STATUSCODE_CONNERROR;        // recatch
-                  sprintf(back[i].r.msg,
-                          "Incorrect length (" LLintP " Bytes, " LLintP
-                          " expected)", (LLint) back[i].r.size,
-                          (LLint) back[i].r.totalsize);
+                  htsblk_failf(&back[i].r,
+                               "Incorrect length (" LLintP " Bytes, " LLintP
+                               " expected)",
+                               (LLint) back[i].r.size,
+                               (LLint) back[i].r.totalsize);
                 } else {
                   // Un warning suffira..
                   hts_log_print(opt, LOG_WARNING,

src/htsback.h

@@ -74,6 +74,10 @@ void back_free(struct_back ** sback);
 // backing
 #define BACK_ADD_TEST "(dummy)"
 #define BACK_ADD_TEST2 "(dummy2)"
+/* Parse an external FTP helper's "<statuscode> <message>" result file into r,
+   clipping the message to r->msg. */
+void back_read_ftp_result(FILE *fp, htsblk *r);
+
 int back_index(httrackp * opt, struct_back * sback, const char *adr, const char *fil,
                const char *sav);
 int back_available(const struct_back * sback);

src/htsbacktrace.c

@@ -0,0 +1,238 @@
+/* ------------------------------------------------------------ */
+/*
+HTTrack Website Copier, Offline Browser for Windows and Unix
+Copyright (C) 1998 Xavier Roche and other contributors
+
+SPDX-License-Identifier: GPL-3.0-or-later
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation, either version 3 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program. If not, see <http://www.gnu.org/licenses/>.
+
+Ethical use: we kindly ask that you NOT use this software to harvest email
+addresses or to collect any other private information about people. Doing so
+would dishonor our work and waste the many hours we have spent on it.
+
+Please visit our Website: http://www.httrack.com
+*/
+
+/* ------------------------------------------------------------ */
+/* File: crash backtrace printer                                 */
+/* Author: Xavier Roche                                          */
+/* ------------------------------------------------------------ */
+
+/* Before every header: glibc gates dladdr() on it. */
+#if defined(__linux) && !defined(_GNU_SOURCE)
+#define _GNU_SOURCE
+#endif
+
+#include "htsbacktrace.h"
+
+#include "htsglobal.h"
+
+#include <stdlib.h>
+#include <string.h>
+#ifdef _WIN32
+#include <io.h> /* write */
+#endif
+#ifdef HAVE_UNISTD_H
+#include <unistd.h>
+#endif
+#if (defined(__linux) && defined(HAVE_EXECINFO_H))
+#include <dlfcn.h>
+#include <errno.h>
+#include <execinfo.h>
+#include <signal.h>
+#include <time.h>
+#include <sys/wait.h>
+#define USES_BACKTRACE
+#endif
+
+#ifdef USES_BACKTRACE
+#define BT_MAX_FRAMES 64     /* frames we try to name */
+#define BT_MAX_MODULES 8     /* distinct modules, one child each */
+#define BT_HEX_SIZE 19       /* "0x" + 16 nibbles + NUL */
+#define BT_PATH_SIZE 1024    /* module path; longer is skipped */
+#define BT_WAIT_TICKS 300    /* 10ms ticks, shared: cap a slow child */
+#define BT_NO_SYMBOLIZER 127 /* child exit: execvp() found none */
+
+static hts_boolean symbolize_crash = HTS_TRUE;
+
+/* "0x"-prefixed hex: the handler must stay stdio-free. */
+static void print_hex(char *buffer, uintptr_t value) {
+  static const char digits[] = "0123456789abcdef";
+  size_t i = 2, a, b;
+
+  buffer[0] = '0';
+  buffer[1] = 'x';
+  do {
+    buffer[i++] = digits[value & 0xf];
+    value >>= 4;
+  } while (value != 0);
+  buffer[i] = '\0';
+  for (a = 2, b = i - 1; a < b; a++, b--) {
+    const char c = buffer[a];
+
+    buffer[a] = buffer[b];
+    buffer[b] = c;
+  }
+}
+
+/* HTS_FALSE if src does not fit: a truncated module path would point the
+   symbolizer at the wrong file. */
+static hts_boolean copy_bounded(char *dest, size_t size, const char *src) {
+  size_t i;
+
+  for (i = 0; i < size - 1 && src[i] != '\0'; i++) {
+    dest[i] = src[i];
+  }
+  dest[i] = '\0';
+  return src[i] == '\0' ? HTS_TRUE : HTS_FALSE;
+}
+
+/* Run the symbolizer on argv, output on fd, within *budget ticks. HTS_FALSE
+   only if none could be run at all; otherwise silent, the raw trace stands. */
+static hts_boolean spawn_symbolizer(char **argv, int fd, int *budget) {
+  const pid_t pid = fork();
+  int status = 0;
+
+  if (pid == -1)
+    return HTS_FALSE;
+  if (pid == 0) {
+    static char llvm_prog[] = "llvm-symbolizer";
+    static char llvm_opts[] = "-p";
+
+    dup2(fd, 1); /* both symbolizers write on stdout */
+    execvp(argv[0], argv);
+    argv[0] = llvm_prog; /* an LLVM-only install ships no addr2line */
+    argv[1] = llvm_opts;
+    execvp(argv[0], argv);
+    _exit(BT_NO_SYMBOLIZER);
+  }
+  for (; *budget > 0; (*budget)--) {
+    const struct timespec tick = {0, 10 * 1000 * 1000};
+    const pid_t reaped = waitpid(pid, &status, WNOHANG);
+
+    if (reaped == pid)
+      return WIFEXITED(status) && WEXITSTATUS(status) == BT_NO_SYMBOLIZER
+                 ? HTS_FALSE
+                 : HTS_TRUE;
+    if (reaped == -1 && errno != EINTR)
+      return HTS_TRUE;
+    nanosleep(&tick, NULL);
+  }
+  kill(pid, SIGKILL);
+  waitpid(pid, NULL, 0);
+  return HTS_TRUE;
+}
+
+/* Name the frames backtrace_symbols_fd() leaves as module+offset:
+   -fvisibility=hidden keeps them out of .dynsym, but DWARF has them. dladdr()
+   is not formally async-signal-safe; accepted, this path is already fatal. */
+static void symbolize_backtrace(void *const *stack, int size, int fd) {
+  static char prog[] = "addr2line";
+  static char opts[] = "-Cfipa";
+  static char dashe[] = "-e";
+  char hex[BT_MAX_FRAMES][BT_HEX_SIZE];
+  const void *base[BT_MAX_FRAMES];
+  const char *name[BT_MAX_FRAMES];
+  hts_boolean grouped[BT_MAX_FRAMES];
+  char module[BT_PATH_SIZE];
+  char *argv[4 + BT_MAX_FRAMES + 1];
+  int budget = BT_WAIT_TICKS;
+  int i, spawned;
+
+  if (size > BT_MAX_FRAMES)
+    size = BT_MAX_FRAMES;
+
+  for (i = 0; i < size; i++) {
+    Dl_info info;
+
+    grouped[i] = HTS_TRUE; /* skipped unless dladdr() places the frame */
+    if (dladdr(stack[i], &info) == 0 || info.dli_fname == NULL ||
+        info.dli_fname[0] == '\0')
+      continue;
+    base[i] = info.dli_fbase;
+    name[i] = info.dli_fname;
+    print_hex(hex[i], (uintptr_t) ((const char *) stack[i] -
+                                   (const char *) info.dli_fbase));
+    grouped[i] = HTS_FALSE;
+  }
+
+  /* One child per module: addr2line takes a single -e. Each frame is claimed
+     once, so argc cannot exceed argv[]. */
+  for (spawned = 0; spawned < BT_MAX_MODULES; spawned++) {
+    int first, j, argc = 0;
+
+    for (first = 0; first < size && grouped[first]; first++)
+      ;
+    if (first >= size)
+      break;
+    argv[argc++] = prog;
+    argv[argc++] = opts;
+    argv[argc++] = dashe;
+    argv[argc++] = module;
+    for (j = first; j < size; j++) {
+      if (grouped[j] || base[j] != base[first])
+        continue;
+      grouped[j] = HTS_TRUE;
+      argv[argc++] = hex[j];
+    }
+    argv[argc] = NULL;
+    /* access(): skip pseudo-modules like linux-vdso, which have no file and
+       would draw nothing but an addr2line complaint. */
+    if (copy_bounded(module, sizeof(module), name[first]) &&
+        access(module, R_OK) == 0) {
+      const size_t len = strlen(module);
+
+      /* addr2line -a prints offsets only: say which module they are in. */
+      (void) (write(fd, module, len) == (ssize_t) len);
+      (void) (write(fd, ":\n", 2) == 2);
+      if (!spawn_symbolizer(argv, fd, &budget))
+        break; /* no symbolizer: stop at one header */
+    }
+  }
+}
+#endif
+
+void hts_backtrace_init(void) {
+#ifdef USES_BACKTRACE
+  symbolize_crash =
+      getenv("HTTRACK_NO_SYMBOLIZE") == NULL ? HTS_TRUE : HTS_FALSE;
+#endif
+}
+
+void hts_print_backtrace(int fd) {
+#ifdef USES_BACKTRACE
+  void *stack[256];
+  const int size = backtrace(stack, sizeof(stack) / sizeof(stack[0]));
+
+  /* A fault inside the handler lands back here: symbolizing twice interleaves
+     two traces on fd and spends a second budget. */
+  static volatile sig_atomic_t entered = 0;
+
+  if (size != 0) {
+    backtrace_symbols_fd(stack, size, fd);
+    if (symbolize_crash && entered == 0) {
+      entered = 1;
+      symbolize_backtrace(stack, size, fd);
+      entered = 0;
+    }
+  }
+#else
+  const char msg[] = "No stack trace available on this OS :(\n";
+
+  if (write(fd, msg, sizeof(msg) - 1) != sizeof(msg) - 1) {
+    /* sorry GCC */
+  }
+#endif
+}

src/htsbacktrace.h

@@ -0,0 +1,45 @@
+/* ------------------------------------------------------------ */
+/*
+HTTrack Website Copier, Offline Browser for Windows and Unix
+Copyright (C) 1998 Xavier Roche and other contributors
+
+SPDX-License-Identifier: GPL-3.0-or-later
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation, either version 3 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program. If not, see <http://www.gnu.org/licenses/>.
+
+Ethical use: we kindly ask that you NOT use this software to harvest email
+addresses or to collect any other private information about people. Doing so
+would dishonor our work and waste the many hours we have spent on it.
+
+Please visit our Website: http://www.httrack.com
+*/
+
+/* ------------------------------------------------------------ */
+/* File: crash backtrace printer                                 */
+/* Author: Xavier Roche                                          */
+/* ------------------------------------------------------------ */
+
+#ifndef HTSBACKTRACE_DEFH
+#define HTSBACKTRACE_DEFH
+
+/* Sample HTTRACK_NO_SYMBOLIZE before any crash: getenv() is not signal-safe.
+   Call once, from the process that installs the fatal-signal handlers. */
+void hts_backtrace_init(void);
+
+/* Write the calling thread's stack to fd, callable from a fatal signal handler:
+   raw frames first, then whatever an external symbolizer can name. Allocates
+   nothing; prints a one-line notice where the OS has no backtrace(). */
+void hts_print_backtrace(int fd);
+
+#endif

src/htscache.c

@@ -142,10 +142,12 @@ struct cache_back_zip_entry {
   int compressionMethod;
 };
 
+/* A corrupt cache can carry a field wider than ours; clipping it keeps the
+   entry, where aborting would take the crawl down. */
 #define ZIP_READFIELD_STRING(line, value, refline, refvalue, refvalue_size)    \
   do {                                                                         \
     if (line[0] != '\0' && strfield2(line, refline)) {                         \
-      strlcpybuff(refvalue, value, refvalue_size);                             \
+      (void) strclipbuff(refvalue, refvalue_size, value);                      \
       line[0] = '\0';                                                          \
     }                                                                          \
   } while (0)
@@ -1295,12 +1297,17 @@ char *readfile2(const char *fil, LLint * size) {
 }
 
 /* Note: utf-8 */
-char *readfile_utf8(const char *fil) {
+char *readfile_utf8(const char *fil) { return readfile2_utf8(fil, NULL); }
+
+/* Note: utf-8 */
+char *readfile2_utf8(const char *fil, LLint *size) {
   char *adr = NULL;
   char catbuff[CATBUFF_SIZE];
   const LLint len = fsize_utf8(fil);
   const size_t buflen = len >= 0 ? llint_to_size_t(len) : (size_t) -1;
 
+  if (size != NULL)
+    *size = len;
   if (buflen != (size_t) -1) { // exists, and is addressable (see readfile2)
     FILE *const fp = FOPEN(fconv(catbuff, sizeof(catbuff), fil), "rb");
 

src/htscache_selftest.c

@@ -1195,6 +1195,12 @@ int cache_legacy_refused_selftest(httrackp *opt, const char *dir) {
 
 /* --- read-side corruption injection --------------------------------------- */
 
+/* 100 'A's: a placeholder header line long enough to be overwritten by a
+   forged, over-long X-StatusMessage of the same byte length. */
+#define CORRUPT_LONG_ETAG                                                      \
+  "AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA"                         \
+  "AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA"
+
 /* canary read back intact after each corruption; victim gets the byte surgery
  */
 #define CORRUPT_ADR "corrupt.example.com"
@@ -1219,6 +1225,24 @@ static void corrupt_build(httrackp *opt) {
   selftest_close(&cache);
 }
 
+/* Like corrupt_build, but the victim carries a 100-char Etag placeholder. */
+static void corrupt_build_longetag(httrackp *opt) {
+  cache_back cache;
+
+  memset(corrupt_body_a, 'a', sizeof(corrupt_body_a) - 1);
+  memset(corrupt_body_b, 'b', sizeof(corrupt_body_b) - 1);
+  remove(reconcile_st_path(opt, "hts-cache/new.zip"));
+  remove(reconcile_st_path(opt, "hts-cache/old.zip"));
+  selftest_open_for_write(&cache, opt);
+  store_entry(opt, &cache, CORRUPT_ADR, "/canary.html", "canary.html", 200,
+              "OK", "text/html", "utf-8", "", "", "", "", corrupt_body_a,
+              strlen(corrupt_body_a));
+  store_entry(opt, &cache, CORRUPT_ADR, "/victim.html", "victim.html", 200,
+              "OK", "text/html", "utf-8", "", CORRUPT_LONG_ETAG, "", "",
+              corrupt_body_b, strlen(corrupt_body_b));
+  selftest_close(&cache);
+}
+
 /* Like corrupt_build, but the victim carries a 20-char Etag whose header line
    is later overwritten with a forged oversized X-Size (same byte length). */
 static void corrupt_build_etag(httrackp *opt) {
@@ -1403,6 +1427,48 @@ static int corrupt_expect_disk_header(httrackp *opt, LLint wantsize,
   return fail;
 }
 
+/* An over-long field from a foreign cache must clip, not abort: the entry
+   still reads, clipped to capacity, and the canary survives. */
+static int corrupt_expect_victim_clipped(httrackp *opt, size_t wantmsg,
+                                         size_t wantlastmod, const char *what) {
+  cache_back cache;
+  htsblk v, c;
+  char BIGSTK lv[HTS_URLMAXSIZE * 2];
+  char BIGSTK lc[HTS_URLMAXSIZE * 2];
+  int fail = 0;
+
+  selftest_open_for_read(&cache, opt);
+  lv[0] = lc[0] = '\0';
+  v = cache_readex(opt, &cache, CORRUPT_ADR, "/victim.html", "", lv, NULL, 1);
+  if (v.statuscode != 200) {
+    fprintf(stderr, "%s: %s: status %d, expected 200\n", selftest_tag, what,
+            v.statuscode);
+    fail++;
+  }
+  if (wantmsg != (size_t) -1 && strlen(v.msg) != wantmsg) {
+    fprintf(stderr, "%s: %s: msg len %u, expected %u\n", selftest_tag, what,
+            (unsigned) strlen(v.msg), (unsigned) wantmsg);
+    fail++;
+  }
+  if (wantlastmod != (size_t) -1 && strlen(v.lastmodified) != wantlastmod) {
+    fprintf(stderr, "%s: %s: lastmodified len %u, expected %u\n", selftest_tag,
+            what, (unsigned) strlen(v.lastmodified), (unsigned) wantlastmod);
+    fail++;
+  }
+  c = cache_readex(opt, &cache, CORRUPT_ADR, "/canary.html", "", lc, NULL, 1);
+  if (c.statuscode != 200) {
+    fprintf(stderr, "%s: %s: canary tainted (status %d)\n", selftest_tag, what,
+            c.statuscode);
+    fail++;
+  }
+  if (v.adr != NULL)
+    freet(v.adr);
+  if (c.adr != NULL)
+    freet(c.adr);
+  selftest_close(&cache);
+  return fail;
+}
+
 /* One zip corruption case: build, patch, then check victim+canary in-session.
  */
 static int corrupt_case_zip(httrackp *opt, const char *pat, const char *rep,
@@ -1439,6 +1505,31 @@ int cache_corruption_selftest(httrackp *opt, const char *dir) {
   failures += corrupt_expect_victim(opt, "Cache Read Error : Read Data",
                                     "garbled deflate stream");
 
+  /* A corrupt cache can hold a field wider than ours. Clipping keeps the
+     entry; aborting would take the crawl down. Overwrite the placeholder Etag
+     line in place, same byte length, so the zip offsets stay intact. */
+  corrupt_build_longetag(opt);
+  corrupt_patch(opt, "Etag: " CORRUPT_LONG_ETAG, 106,
+                "X-StatusMessage: " /* 17 + 89 = 106 */
+                "AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA"
+                "AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA",
+                1, 1);
+  failures +=
+      corrupt_expect_victim_clipped(opt, sizeof(((htsblk *) 0)->msg) - 1,
+                                    (size_t) -1, "over-long X-StatusMessage");
+
+  /* lastmodified[64] is narrower than msg[80]: one hardcoded clip length
+     cannot satisfy both. */
+  corrupt_build_longetag(opt);
+  corrupt_patch(opt, "Etag: " CORRUPT_LONG_ETAG, 106,
+                "Last-Modified: " /* 15 + 91 = 106 */
+                "AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA"
+                "AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA",
+                1, 1);
+  failures += corrupt_expect_victim_clipped(
+      opt, (size_t) -1, sizeof(((htsblk *) 0)->lastmodified) - 1,
+      "over-long Last-Modified");
+
   /* An X-Size above INT_MAX is positive as int64 (slips a bare sign check) but
      truncates negative in the (int) cast the malloc uses: a wraparound alloc.
      cache_add asserts size fits an int, so such a value only reaches the reader

src/htschanges.c

@@ -0,0 +1,679 @@
+/* ------------------------------------------------------------ */
+/*
+HTTrack Website Copier, Offline Browser for Windows and Unix
+Copyright (C) 2026 Xavier Roche and other contributors
+
+SPDX-License-Identifier: GPL-3.0-or-later
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation, either version 3 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program. If not, see <http://www.gnu.org/licenses/>.
+
+Ethical use: we kindly ask that you NOT use this software to harvest email
+addresses or to collect any other private information about people. Doing so
+would dishonor our work and waste the many hours we have spent on it.
+
+Please visit our Website: http://www.httrack.com
+*/
+
+/* ------------------------------------------------------------ */
+/* File: htschanges.c subroutines:                              */
+/*       --changes: what this crawl changed vs. the previous    */
+/*       mirror (hts-changes.json)                              */
+/* Author: Xavier Roche                                         */
+/* ------------------------------------------------------------ */
+
+#define HTS_INTERNAL_BYTECODE
+
+#include "htschanges.h"
+
+#include "htscache.h"
+#include "htscharset.h"
+#include "htscore.h"
+#include "htslib.h"
+#include "htsmd5.h"
+#include "htssafe.h"
+#include "htsthread.h"
+#include "htstools.h"
+#include "coucal/coucal.h"
+#include "md5.h"
+
+#include <stdlib.h>
+#include <string.h>
+#include <time.h>
+
+#define DIGEST_SIZE 16
+
+typedef struct hts_changes hts_changes;
+
+/* One mirrored file. The mirror-relative path is the key, not the URL: a
+   redirect and its target can share a save name, and counting that file twice
+   would put it in two buckets. */
+typedef struct {
+  char *url;               /* absolute URL, "" for an engine-generated file */
+  char *file;              /* mirror-relative path, as listed in new.lst */
+  hts_boolean rewritten;   /* the crawl wrote over the local copy */
+  hts_boolean not_updated; /* transfer signal: the server reported no change */
+  hts_boolean existed;     /* the file was part of the previous mirror */
+  hts_boolean on_disk;     /* a copy was on disk when the crawl first saw it */
+  hts_boolean listed_prev; /* the previous mirror's index listed this file */
+  hts_boolean has_prev;    /* prev_digest holds the previous payload's digest */
+  hts_boolean has_new;     /* new_digest was taken from the payload, not disk */
+  unsigned char prev_digest[DIGEST_SIZE];
+  unsigned char new_digest[DIGEST_SIZE];
+  LLint prev_size; /* -1 when unknown */
+  LLint size;      /* size after the crawl, -1 when unknown */
+  hts_change_bucket bucket;
+} changes_entry;
+
+struct hts_changes {
+  coucal index; /* mirror-relative path -> entry slot + 1 */
+  changes_entry *entries;
+  size_t count;
+  size_t capacity;
+  hts_boolean has_index; /* the run keeps a mirror index (the cache is on) */
+  hts_boolean old_index; /* the previous mirror's index was read */
+  hts_boolean overflow;  /* an allocation failed; the report is partial */
+  hts_boolean closed;    /* reported: late notifies are dropped, not recorded */
+};
+
+/* ------------------------------------------------------------ */
+/* JSON                                                          */
+/* ------------------------------------------------------------ */
+
+/* Length of the UTF-8 sequence led by c, or 0 if c cannot lead one. */
+static size_t utf8_lead_length(unsigned char c) {
+  if (c >= 0xc2 && c <= 0xdf)
+    return 2;
+  if (c >= 0xe0 && c <= 0xef)
+    return 3;
+  if (c >= 0xf0 && c <= 0xf4)
+    return 4;
+  return 0;
+}
+
+void hts_changes_json_string(String *out, const char *s) {
+  const unsigned char *p = (const unsigned char *) s;
+
+  StringAddchar(*out, '"');
+  while (*p != '\0') {
+    const unsigned char c = *p;
+
+    if (c == '"' || c == '\\') {
+      StringAddchar(*out, '\\');
+      StringAddchar(*out, (char) c);
+      p++;
+    } else if (c < 0x20 || c == 0x7f) {
+      char esc[8];
+
+      snprintf(esc, sizeof(esc), "\\u%04x", (unsigned) c);
+      StringCat(*out, esc);
+      p++;
+    } else if (c < 0x80) {
+      StringAddchar(*out, (char) c);
+      p++;
+    } else {
+      /* Non-ASCII: emit it verbatim only if it is a well-formed sequence, so
+         a legacy-charset URL cannot produce unparseable JSON. */
+      const size_t len = utf8_lead_length(c);
+
+      if (len != 0 && strnlen((const char *) p, len) == len &&
+          hts_isStringUTF8((const char *) p, len)) {
+        StringMemcat(*out, (const char *) p, len);
+        p += len;
+      } else {
+        StringCat(*out, "\\ufffd");
+        p++;
+      }
+    }
+  }
+  StringAddchar(*out, '"');
+}
+
+/* ------------------------------------------------------------ */
+/* Accumulator                                                   */
+/* ------------------------------------------------------------ */
+
+static hts_changes *changes_new(void) {
+  hts_changes *changes = calloct(1, sizeof(*changes));
+
+  if (changes == NULL)
+    return NULL;
+  changes->index = coucal_new(0);
+  if (changes->index == NULL) {
+    freet(changes);
+    return NULL;
+  }
+  return changes;
+}
+
+static void changes_free(hts_changes **pchanges) {
+  hts_changes *changes = *pchanges;
+
+  if (changes == NULL)
+    return;
+  if (changes->entries != NULL) {
+    size_t i;
+
+    for (i = 0; i < changes->count; i++) {
+      freet(changes->entries[i].url);
+      freet(changes->entries[i].file);
+    }
+    freet(changes->entries);
+  }
+  if (changes->index != NULL)
+    coucal_delete(&changes->index);
+  freet(changes);
+  *pchanges = NULL;
+}
+
+/* Slot of `file`, or -1 when it has not been recorded. */
+static intptr_t changes_find(const hts_changes *changes, const char *file) {
+  intptr_t slot = 0;
+
+  return coucal_read(changes->index, file, &slot) ? slot - 1 : -1;
+}
+
+/* Slot of `file`, appending a fresh entry when it is new; -1 if allocation
+   failed, which flags the report partial. */
+static intptr_t changes_slot(hts_changes *changes, const char *file,
+                             hts_boolean *is_new) {
+  intptr_t slot = 0;
+
+  *is_new = HTS_FALSE;
+  if (coucal_read(changes->index, file, &slot))
+    return slot - 1;
+
+  if (changes->count == changes->capacity) {
+    const size_t capacity = changes->capacity != 0 ? changes->capacity * 2 : 64;
+    changes_entry *const entries =
+        realloct(changes->entries, capacity * sizeof(*entries));
+
+    if (entries == NULL) {
+      changes->overflow = HTS_TRUE;
+      return -1;
+    }
+    changes->entries = entries;
+    changes->capacity = capacity;
+  }
+  slot = (intptr_t) changes->count;
+  memset(&changes->entries[slot], 0, sizeof(changes->entries[slot]));
+  changes->entries[slot].file = strdupt(file);
+  changes->entries[slot].prev_size = -1;
+  changes->entries[slot].size = -1;
+  if (changes->entries[slot].file == NULL) {
+    changes->overflow = HTS_TRUE;
+    return -1;
+  }
+  changes->count++;
+  coucal_write(changes->index, file, slot + 1);
+  *is_new = HTS_TRUE;
+  return slot;
+}
+
+/* MD5 of the file at `path`; not a security boundary. Never call it holding
+   changes_mutex: hashing a large file would stall every other connection.
+   HTS_FALSE if it cannot be read. */
+static hts_boolean digest_file(const char *path,
+                               unsigned char digest[DIGEST_SIZE]) {
+  const int endian = 1;
+  struct MD5Context ctx;
+  char BIGSTK buffer[32768];
+  FILE *fp = FOPEN(path, "rb");
+  size_t nread;
+
+  if (fp == NULL)
+    return HTS_FALSE;
+  MD5Init(&ctx, *((const char *) &endian));
+  while ((nread = fread(buffer, 1, sizeof(buffer), fp)) > 0)
+    MD5Update(&ctx, (const unsigned char *) buffer, (unsigned int) nread);
+  if (ferror(fp) != 0) {
+    fclose(fp);
+    return HTS_FALSE;
+  }
+  fclose(fp);
+  MD5Final(digest, &ctx);
+  return HTS_TRUE;
+}
+
+static void digest_mem(const char *buffer, size_t len,
+                       unsigned char digest[DIGEST_SIZE]) {
+  domd5mem(buffer, len, (char *) digest, 0);
+}
+
+hts_change_bucket hts_changes_classify(hts_boolean rewritten,
+                                       hts_boolean existed,
+                                       hts_boolean not_updated,
+                                       hts_boolean have_digests,
+                                       hts_boolean digests_equal) {
+  if (!rewritten)
+    return HTS_CHANGE_UNCHANGED; /* the crawl left the copy alone */
+  if (!existed)
+    return HTS_CHANGE_NEW;
+  if (have_digests)
+    return digests_equal ? HTS_CHANGE_UNCHANGED : HTS_CHANGE_CHANGED;
+  /* No digest to compare: a server with no validators answers 200 with the
+     same bytes, so this signal alone over-reports. */
+  return not_updated ? HTS_CHANGE_UNCHANGED : HTS_CHANGE_CHANGED;
+}
+
+/* ------------------------------------------------------------ */
+/* Engine hooks                                                  */
+/* ------------------------------------------------------------ */
+
+/* FTP transfers reach file_notify() from a thread the crawl never joins, so
+   every entry point below, report and teardown included, takes this lock. */
+static htsmutex changes_mutex = HTSMUTEX_INIT;
+
+/* The live accumulator, created on first use; NULL when --changes is off or
+   the report is already written. Call under the lock. */
+static hts_changes *changes_get(httrackp *opt) {
+  hts_changes *changes = (hts_changes *) opt->changes_state;
+
+  if (!opt->changes)
+    return NULL;
+  if (changes == NULL) {
+    changes = changes_new();
+    opt->changes_state = changes;
+  }
+  return changes != NULL && !changes->closed ? changes : NULL;
+}
+
+void hts_changes_notify(httrackp *opt, const char *adr, const char *fil,
+                        const char *save, hts_boolean rewritten,
+                        hts_boolean not_updated) {
+  hts_changes *changes;
+  char BIGSTK file[HTS_URLMAXSIZE * 2];
+  char BIGSTK url[HTS_URLMAXSIZE * 4 + 8]; /* holds adr + fil + a scheme */
+  unsigned char prev_digest[DIGEST_SIZE];
+  hts_boolean has_prev = HTS_FALSE;
+  hts_boolean is_new = HTS_FALSE;
+  intptr_t slot = -1;
+  LLint prev_size;
+
+  if (!opt->changes)
+    return;
+  /* Engine-generated scaffolding (the top index) carries no URL and is not a
+     mirrored resource. */
+  if (save == NULL || !strnotempty(save) || adr == NULL || fil == NULL ||
+      (!strnotempty(adr) && !strnotempty(fil)))
+    return;
+
+  hts_savename_listed(StringBuff(opt->path_html), save, file, sizeof(file));
+
+  hts_mutexlock(&changes_mutex);
+  changes = changes_get(opt);
+  if (changes != NULL) {
+    slot = changes_slot(changes, file, &is_new);
+    if (slot >= 0 && !is_new) {
+      /* Retry, or a second call site for the same file: the pre-run state was
+         already sampled and the copy on disk may no longer be it. */
+      changes->entries[slot].rewritten =
+          changes->entries[slot].rewritten || rewritten;
+    }
+  }
+  hts_mutexrelease(&changes_mutex);
+  if (slot < 0 || !is_new)
+    return;
+
+  url[0] = '\0';
+  if (!link_has_authority(adr))
+    strlcatbuff(url, "http://", sizeof(url));
+  strlcatbuff(url, adr, sizeof(url));
+  strlcatbuff(url, fil, sizeof(url));
+
+  /* Unlocked: hashing a large file would stall every other connection. Hash
+     it only when it is about to be overwritten, the last moment it exists. */
+  prev_size = fsize_utf8(save);
+  if (prev_size >= 0 && rewritten)
+    has_prev = digest_file(save, prev_digest);
+
+  hts_mutexlock(&changes_mutex);
+  changes = changes_get(opt);
+  /* The slot index is stable, entries being append-only; the array is not. */
+  if (changes != NULL && (size_t) slot < changes->count) {
+    changes_entry *const entry = &changes->entries[slot];
+
+    entry->url = strdupt(url);
+    if (entry->url == NULL)
+      changes->overflow = HTS_TRUE;
+    entry->rewritten = entry->rewritten || rewritten;
+    entry->not_updated = not_updated;
+    entry->prev_size = prev_size;
+    entry->on_disk = prev_size >= 0;
+    /* hts_changes_html() compares payloads, and its digests win. */
+    if (has_prev && !entry->has_new) {
+      entry->has_prev = HTS_TRUE;
+      memcpy(entry->prev_digest, prev_digest, DIGEST_SIZE);
+    }
+  }
+  hts_mutexrelease(&changes_mutex);
+}
+
+void hts_changes_html(httrackp *opt, cache_back *cache, const htsblk *r,
+                      const char *adr, const char *fil, const char *save) {
+  hts_changes *changes;
+  char BIGSTK file[HTS_URLMAXSIZE * 2];
+  char BIGSTK location[HTS_URLMAXSIZE * 2];
+  unsigned char new_digest[DIGEST_SIZE];
+  unsigned char prev_digest[DIGEST_SIZE];
+  hts_boolean has_prev = HTS_FALSE;
+  intptr_t slot;
+  htsblk prev;
+
+  /* Ahead of the digest and the cache read: off must cost nothing. */
+  if (!opt->changes)
+    return;
+  if (r->adr == NULL || r->size < 0 || save == NULL || !strnotempty(save))
+    return;
+
+  /* On disk this is the payload plus rewritten links and a footer dated by the
+     crawl, so it differs every run; compare payloads, the previous one being
+     the body the cache kept. */
+  digest_mem(r->adr, (size_t) r->size, new_digest);
+  prev = cache_read_ro(opt, cache, adr, fil, "", location);
+  if (HTTP_IS_OK(prev.statuscode) && prev.adr != NULL && prev.size >= 0) {
+    digest_mem(prev.adr, (size_t) prev.size, prev_digest);
+    has_prev = HTS_TRUE;
+  }
+  freet(prev.adr);
+
+  /* Only now take the lock and look the entry up: cache_read_ro() can itself
+     reach file_notify(), which would move the entries array. */
+  hts_mutexlock(&changes_mutex);
+  changes = changes_get(opt);
+  /* file_notify() records the entry; this call only refines it. */
+  if (changes != NULL) {
+    hts_savename_listed(StringBuff(opt->path_html), save, file, sizeof(file));
+    slot = changes_find(changes, file);
+    if (slot >= 0) {
+      changes_entry *const entry = &changes->entries[slot];
+
+      entry->has_new = HTS_TRUE;
+      memcpy(entry->new_digest, new_digest, DIGEST_SIZE);
+      entry->has_prev = has_prev;
+      if (has_prev)
+        memcpy(entry->prev_digest, prev_digest, DIGEST_SIZE);
+    }
+  }
+  hts_mutexrelease(&changes_mutex);
+}
+
+void hts_changes_dropped(httrackp *opt, const char *file, hts_boolean kept) {
+  hts_changes *changes;
+  hts_boolean is_new;
+  intptr_t slot;
+
+  if (!opt->changes || file == NULL || !strnotempty(file))
+    return;
+  hts_mutexlock(&changes_mutex);
+  changes = changes_get(opt);
+  if (changes != NULL) {
+    slot = changes_slot(changes, file, &is_new);
+    if (slot >= 0 && is_new) {
+      changes->entries[slot].url = strdupt("");
+      changes->entries[slot].listed_prev = HTS_TRUE;
+      /* A kept file leaves rewritten clear, which resolves to unchanged. */
+      if (!kept)
+        changes->entries[slot].bucket = HTS_CHANGE_GONE;
+    }
+  }
+  hts_mutexrelease(&changes_mutex);
+}
+
+void hts_changes_previous(httrackp *opt, const char *file) {
+  hts_changes *changes;
+  intptr_t slot;
+
+  if (!opt->changes || file == NULL)
+    return;
+  hts_mutexlock(&changes_mutex);
+  changes = changes_get(opt);
+  if (changes != NULL) {
+    changes->has_index = HTS_TRUE;
+    changes->old_index = HTS_TRUE;
+    slot = changes_find(changes, file);
+    if (slot >= 0)
+      changes->entries[slot].listed_prev = HTS_TRUE;
+  }
+  hts_mutexrelease(&changes_mutex);
+}
+
+void hts_changes_indexed(httrackp *opt) {
+  hts_changes *changes;
+
+  if (!opt->changes)
+    return;
+  hts_mutexlock(&changes_mutex);
+  changes = changes_get(opt);
+  if (changes != NULL)
+    changes->has_index = HTS_TRUE;
+  hts_mutexrelease(&changes_mutex);
+}
+
+/* ------------------------------------------------------------ */
+/* Report                                                        */
+/* ------------------------------------------------------------ */
+
+/* Assign each entry its final bucket from the bytes now on disk. */
+static void changes_resolve(hts_changes *changes, httrackp *opt) {
+  char catbuff[CATBUFF_SIZE];
+  size_t i;
+
+  for (i = 0; i < changes->count; i++) {
+    changes_entry *const entry = &changes->entries[i];
+    const char *path;
+    unsigned char digest[DIGEST_SIZE];
+    hts_boolean have_digests = HTS_FALSE;
+    hts_boolean digests_equal = HTS_FALSE;
+
+    if (entry->bucket == HTS_CHANGE_GONE)
+      continue;
+
+    /* The previous mirror's index is the authority on what was there before:
+       a partial left by this crawl's own failed attempt is on disk but was
+       never part of the previous mirror. Without an index, fall back to what
+       the first notify saw on disk. */
+    entry->existed = changes->old_index ? entry->listed_prev : entry->on_disk;
+    path = fconcat(catbuff, sizeof(catbuff), StringBuff(opt->path_html),
+                   entry->file);
+    entry->size = fsize_utf8(path);
+    if (entry->rewritten && entry->existed) {
+      /* The size shortcut only holds when both digests describe the file on
+         disk; has_new means they are payload digests, and a parsed page's
+         rendered size moves with its links and footer. */
+      if (!entry->has_new && entry->size >= 0 && entry->prev_size >= 0 &&
+          entry->size != entry->prev_size) {
+        have_digests = HTS_TRUE; /* different lengths: no need to hash */
+        digests_equal = HTS_FALSE;
+      } else if (entry->has_prev &&
+                 (entry->has_new
+                      ? (memcpy(digest, entry->new_digest, DIGEST_SIZE), 1)
+                      : digest_file(path, digest))) {
+        have_digests = HTS_TRUE;
+        digests_equal = memcmp(digest, entry->prev_digest, DIGEST_SIZE) == 0
+                            ? HTS_TRUE
+                            : HTS_FALSE;
+      }
+    }
+    entry->bucket =
+        hts_changes_classify(entry->rewritten, entry->existed,
+                             entry->not_updated, have_digests, digests_equal);
+  }
+}
+
+static const char *const bucket_names[HTS_CHANGE_BUCKETS] = {
+    "new", "changed", "unchanged", "gone"};
+
+/* Serialize the report. Call under changes_mutex. */
+static void changes_serialize(httrackp *opt, String *out) {
+  hts_changes *const changes = (hts_changes *) opt->changes_state;
+  size_t counts[HTS_CHANGE_BUCKETS];
+  char date[32];
+  char scratch[64];
+  int bucket;
+  size_t i;
+
+  StringClear(*out);
+  if (changes == NULL)
+    return;
+  changes_resolve(changes, opt);
+
+  memset(counts, 0, sizeof(counts));
+  for (i = 0; i < changes->count; i++)
+    counts[changes->entries[i].bucket]++;
+
+  hts_now_iso8601(date);
+  StringCat(*out, "{\n  \"schema\": ");
+  snprintf(scratch, sizeof(scratch), "%d", HTS_CHANGES_SCHEMA);
+  StringCat(*out, scratch);
+  StringCat(*out, ",\n  \"generator\": ");
+  hts_changes_json_string(out, "HTTrack Website Copier/" HTTRACK_VERSION);
+  StringCat(*out, ",\n  \"date\": ");
+  hts_changes_json_string(out, date);
+  StringCat(*out, ",\n  \"first_crawl\": ");
+  StringCat(*out, !changes->has_index
+                      ? "null"
+                      : (changes->old_index ? "false" : "true"));
+  StringCat(*out, ",\n  \"partial\": ");
+  StringCat(*out, changes->overflow ? "true" : "false");
+  StringCat(*out, ",\n  \"purged\": ");
+  StringCat(*out, opt->delete_old ? "true" : "false");
+
+  StringCat(*out, ",\n  \"counts\": {");
+  for (bucket = 0; bucket < HTS_CHANGE_BUCKETS; bucket++) {
+    StringCat(*out, bucket != 0 ? ", " : " ");
+    hts_changes_json_string(out, bucket_names[bucket]);
+    snprintf(scratch, sizeof(scratch), ": %d", (int) counts[bucket]);
+    StringCat(*out, scratch);
+  }
+  StringCat(*out, " }");
+
+  for (bucket = 0; bucket < HTS_CHANGE_BUCKETS; bucket++) {
+    hts_boolean first = HTS_TRUE;
+
+    StringCat(*out, ",\n  ");
+    hts_changes_json_string(out, bucket_names[bucket]);
+    StringCat(*out, ": [");
+    for (i = 0; i < changes->count; i++) {
+      const changes_entry *const entry = &changes->entries[i];
+
+      if (entry->bucket != bucket)
+        continue;
+      StringCat(*out, first ? "\n    { \"url\": " : ",\n    { \"url\": ");
+      first = HTS_FALSE;
+      hts_changes_json_string(out, entry->url != NULL ? entry->url : "");
+      StringCat(*out, ", \"file\": ");
+      hts_changes_json_string(out, entry->file);
+      if (entry->size >= 0) {
+        snprintf(scratch, sizeof(scratch), ", \"size\": " LLintP,
+                 (LLint) entry->size);
+        StringCat(*out, scratch);
+      }
+      if (bucket == HTS_CHANGE_CHANGED && entry->prev_size >= 0) {
+        snprintf(scratch, sizeof(scratch), ", \"previous_size\": " LLintP,
+                 (LLint) entry->prev_size);
+        StringCat(*out, scratch);
+      }
+      StringCat(*out, " }");
+    }
+    StringCat(*out, first ? "]" : "\n  ]");
+  }
+  StringCat(*out, "\n}\n");
+}
+
+void hts_changes_report(httrackp *opt, String *out) {
+  hts_mutexlock(&changes_mutex);
+  changes_serialize(opt, out);
+  hts_mutexrelease(&changes_mutex);
+}
+
+void hts_changes_close_opt(httrackp *opt) {
+  char catbuff[CATBUFF_SIZE];
+  const char *path;
+  String report = STRING_EMPTY;
+  size_t counts[HTS_CHANGE_BUCKETS];
+  size_t total;
+  hts_boolean has_index, old_index;
+  FILE *fp;
+
+  hts_mutexlock(&changes_mutex);
+  {
+    /* changes_get(), not the raw field: a crawl that mirrored nothing still
+       owes the user a report, and a stale one from the previous run must not
+       survive on disk as if it described this one. */
+    hts_changes *const changes = changes_get(opt);
+    size_t i;
+
+    if (changes == NULL) {
+      hts_mutexrelease(&changes_mutex);
+      return;
+    }
+    changes_serialize(opt, &report);
+    memset(counts, 0, sizeof(counts));
+    for (i = 0; i < changes->count; i++)
+      counts[changes->entries[i].bucket]++;
+    total = changes->count;
+    has_index = changes->has_index;
+    old_index = changes->old_index;
+    /* Sticky: whatever the crawl's stragglers do next is not in this report. */
+    changes->closed = HTS_TRUE;
+  }
+  hts_mutexrelease(&changes_mutex);
+
+  path = fconcat(catbuff, sizeof(catbuff), StringBuff(opt->path_log),
+                 HTS_CHANGES_FILE);
+  fp = FOPEN(path, "wb");
+  if (fp != NULL) {
+    const size_t len = StringLength(report);
+
+    if (len != 0 && fwrite(StringBuff(report), 1, len, fp) != len)
+      hts_log_print(opt, LOG_ERROR | LOG_ERRNO,
+                    "Unable to write the change report %s", path);
+    fclose(fp);
+  } else {
+    hts_log_print(opt, LOG_ERROR | LOG_ERRNO,
+                  "Unable to create the change report %s", path);
+  }
+
+  if (!has_index) {
+    hts_log_print(opt, LOG_NOTICE,
+                  "Change report: no mirror index (the cache is off), %d files "
+                  "mirrored, deletions not detected (%s)",
+                  (int) total, HTS_CHANGES_FILE);
+  } else if (!old_index) {
+    hts_log_print(opt, LOG_NOTICE,
+                  "Change report: first crawl, %d files mirrored, nothing to "
+                  "compare against (%s)",
+                  (int) total, HTS_CHANGES_FILE);
+  } else {
+    hts_log_print(opt, LOG_NOTICE,
+                  "Change report: %d new, %d changed, %d unchanged, %d gone "
+                  "(%s)",
+                  (int) counts[HTS_CHANGE_NEW],
+                  (int) counts[HTS_CHANGE_CHANGED],
+                  (int) counts[HTS_CHANGE_UNCHANGED],
+                  (int) counts[HTS_CHANGE_GONE], HTS_CHANGES_FILE);
+  }
+
+  StringFree(report);
+}
+
+void hts_changes_free_opt(httrackp *opt) {
+  hts_changes *changes;
+
+  hts_mutexlock(&changes_mutex);
+  changes = (hts_changes *) opt->changes_state;
+  changes_free(&changes);
+  opt->changes_state = NULL;
+  hts_mutexrelease(&changes_mutex);
+}

src/htschanges.h

@@ -0,0 +1,125 @@
+/* ------------------------------------------------------------ */
+/*
+HTTrack Website Copier, Offline Browser for Windows and Unix
+Copyright (C) 2026 Xavier Roche and other contributors
+
+SPDX-License-Identifier: GPL-3.0-or-later
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation, either version 3 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program. If not, see <http://www.gnu.org/licenses/>.
+
+Ethical use: we kindly ask that you NOT use this software to harvest email
+addresses or to collect any other private information about people. Doing so
+would dishonor our work and waste the many hours we have spent on it.
+
+Please visit our Website: http://www.httrack.com
+*/
+
+/* ------------------------------------------------------------ */
+/* HTTrack change report (--changes). Internal, not installed.
+   Accumulates what the crawl did to each mirrored file, compares the bytes
+   against the copy the previous run left behind, and writes
+   hts-changes.json next to the log. */
+/* ------------------------------------------------------------ */
+
+#ifndef HTS_CHANGES_DEFH
+#define HTS_CHANGES_DEFH
+
+#include "htsopt.h"
+#include "htsstrings.h"
+
+#ifndef HTS_DEF_FWSTRUCT_cache_back
+#define HTS_DEF_FWSTRUCT_cache_back
+typedef struct cache_back cache_back;
+#endif
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/* Report file name, written under the project's log directory. */
+#define HTS_CHANGES_FILE "hts-changes.json"
+
+/* Schema version carried by the report; bump on an incompatible change. */
+#define HTS_CHANGES_SCHEMA 1
+
+/* Which side of the comparison a mirrored file ended up on. */
+typedef enum {
+  HTS_CHANGE_NEW = 0,   /* no local copy before this run */
+  HTS_CHANGE_CHANGED,   /* rewritten, and the bytes differ */
+  HTS_CHANGE_UNCHANGED, /* the bytes are those of the previous mirror */
+  HTS_CHANGE_GONE,      /* in the previous mirror, absent from this one */
+  HTS_CHANGE_BUCKETS
+} hts_change_bucket;
+
+/* Record what this crawl is doing to the local file `save` (absolute path;
+   adr/fil form its URL, either may be empty for an engine-generated file).
+   `rewritten` means the copy on disk is being written over, `not_updated` is
+   the 200-versus-304 signal, used only when no digest can be taken. Only the
+   first call for a file samples the previous copy, so a retried or
+   twice-notified resource is counted once. No-op unless --changes is on. */
+void hts_changes_notify(httrackp *opt, const char *adr, const char *fil,
+                        const char *save, hts_boolean rewritten,
+                        hts_boolean not_updated);
+
+/* Refine the entry for a parsed HTML file: its local copy carries rewritten
+   links and a footer dated by the crawl, so it differs every run. Compares the
+   payload `r` holds against the body the previous run left in the cache. */
+void hts_changes_html(httrackp *opt, cache_back *cache, const htsblk *r,
+                      const char *adr, const char *fil, const char *save);
+
+/* Record `file` (mirror-relative, as listed in new.lst) as listed by the
+   previous mirror's index and absent from this run's. `kept` means its local
+   copy survives the run because the crawl tried and failed to replace it: a
+   failed transfer is not a deletion, so it is reported unchanged, not gone. */
+void hts_changes_dropped(httrackp *opt, const char *file, hts_boolean kept);
+
+/* Record that the previous mirror's index listed `file`. That index, not the
+   file's presence on disk, decides what counts as already mirrored. */
+void hts_changes_previous(httrackp *opt, const char *file);
+
+/* Record that this run keeps a mirror index. Without one (the cache is off)
+   deletions and first_crawl are undecidable, and the report says so. */
+void hts_changes_indexed(httrackp *opt);
+
+/* Resolve every entry against the bytes now on disk and serialize the report
+   into `out` (replaced). Exposed for the self-tests. */
+void hts_changes_report(httrackp *opt, String *out);
+
+/* Write the report and log a one-line summary. Idempotent, and seals the
+   accumulator: a late notify is dropped rather than starting a second one. */
+void hts_changes_close_opt(httrackp *opt);
+
+/* Drop the accumulator, for a run that never reached its end and to start the
+   next one clean. Null-safe and idempotent. */
+void hts_changes_free_opt(httrackp *opt);
+
+/* Bucket for one entry, from what was observed. No filesystem access:
+   `have_digests` says both sides could be hashed, `digests_equal` compares
+   them, and `not_updated` is the fallback signal when they could not. */
+hts_change_bucket hts_changes_classify(hts_boolean rewritten,
+                                       hts_boolean existed,
+                                       hts_boolean not_updated,
+                                       hts_boolean have_digests,
+                                       hts_boolean digests_equal);
+
+/* Append `s` to `out` as a quoted JSON string. Byte sequences that are not
+   valid UTF-8 become U+FFFD, so a mirror carrying legacy-charset URLs still
+   produces parseable JSON. */
+void hts_changes_json_string(String *out, const char *s);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif

src/htscmdline.c

@@ -0,0 +1,95 @@
+/* ------------------------------------------------------------ */
+/*
+HTTrack Website Copier, Offline Browser for Windows and Unix
+Copyright (C) 1998 Xavier Roche and other contributors
+
+SPDX-License-Identifier: GPL-3.0-or-later
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation, either version 3 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program. If not, see <http://www.gnu.org/licenses/>.
+
+Ethical use: we kindly ask that you NOT use this software to harvest email
+addresses or to collect any other private information about people. Doing so
+would dishonor our work and waste the many hours we have spent on it.
+
+Please visit our Website: http://www.httrack.com
+*/
+
+/* ------------------------------------------------------------ */
+/* File: command line splitter, shared by the engine and         */
+/*       htsserver                                               */
+/* Author: Xavier Roche                                          */
+/* ------------------------------------------------------------ */
+
+#include "htscmdline.h"
+
+#include "htssafe.h"
+
+#include <limits.h>
+#include <stdint.h>
+
+char **hts_split_cmdline(char *cmd, int *nargs) {
+  size_t nsep = 0;
+  size_t capacity;
+  size_t r;
+  size_t w;
+  int argc = 0;
+  hts_boolean quoted = HTS_FALSE;
+  char **argv;
+
+  *nargs = 0;
+
+  /* fold the other separators, so counting them sizes the vector exactly */
+  for (r = 0; cmd[r] != '\0'; r++) {
+    if (cmd[r] == '\t' || cmd[r] == '\r' || cmd[r] == '\n') {
+      cmd[r] = ' ';
+    }
+    if (cmd[r] == ' ') {
+      nsep++;
+    }
+  }
+
+  /* at most one argument per separator, plus the leading one and the NULL */
+  if (nsep > (size_t) INT_MAX - 1 || nsep > SIZE_MAX / sizeof(char *) - 2) {
+    return NULL;
+  }
+  capacity = nsep + 2;
+  argv = (char **) malloct(capacity * sizeof(char *));
+  if (argv == NULL) {
+    return NULL;
+  }
+
+  argv[argc++] = cmd;
+  for (r = 0, w = 0; cmd[r] != '\0';) {
+    if (quoted && cmd[r] == '\\' &&
+        (cmd[r + 1] == '\\' || cmd[r + 1] == '\"')) {
+      r++;
+      cmd[w++] = cmd[r++];
+    } else if (cmd[r] == '\"') {
+      quoted = !quoted;
+      cmd[w++] = cmd[r++];
+    } else if (cmd[r] == ' ' && !quoted) {
+      cmd[w++] = '\0';
+      assertf((size_t) argc < capacity - 1); /* the last slot holds the NULL */
+      argv[argc++] = cmd + w;
+      r++;
+    } else {
+      cmd[w++] = cmd[r++];
+    }
+  }
+  cmd[w] = '\0';
+  argv[argc] = NULL; /* callers may rely on argv[argc] == NULL */
+
+  *nargs = argc;
+  return argv;
+}

src/htscmdline.h

@@ -0,0 +1,45 @@
+/* ------------------------------------------------------------ */
+/*
+HTTrack Website Copier, Offline Browser for Windows and Unix
+Copyright (C) 1998 Xavier Roche and other contributors
+
+SPDX-License-Identifier: GPL-3.0-or-later
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation, either version 3 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program. If not, see <http://www.gnu.org/licenses/>.
+
+Ethical use: we kindly ask that you NOT use this software to harvest email
+addresses or to collect any other private information about people. Doing so
+would dishonor our work and waste the many hours we have spent on it.
+
+Please visit our Website: http://www.httrack.com
+*/
+
+/* ------------------------------------------------------------ */
+/* File: command line splitter, shared by the engine and         */
+/*       htsserver                                               */
+/* Author: Xavier Roche                                          */
+/* ------------------------------------------------------------ */
+
+#ifndef HTSCMDLINE_DEFH
+#define HTSCMDLINE_DEFH
+
+#include "htsglobal.h"
+
+/* Split "cmd" in place into a NULL-terminated argv vector of *nargs entries,
+   argv[0] being the program name and quotes left for the engine to strip.
+   Returns a malloct'ed vector of pointers into cmd (freet the vector, never its
+   entries), or NULL when it cannot be sized or allocated. */
+char **hts_split_cmdline(char *cmd, int *nargs);
+
+#endif

src/htscore.c

@@ -40,6 +40,8 @@ Please visit our Website: http://www.httrack.com
 /* File defs */
 #include "htscore.h"
 #include "htswarc.h"
+#include "htschanges.h"
+#include "htssinglefile.h"
 
 /* specific definitions */
 #include "htsbase.h"
@@ -447,13 +449,22 @@ void hts_finish_makeindex(httrackp *opt, int *makeindex_done,
                           const char *fil) {
   if (!*makeindex_done) {
     if (*makeindex_fp) {
-      char BIGSTK tempo[1024];
+      /* sized off link_escaped below: at the old flat 1024 a long first link
+         produced a redirect to a clipped URL */
+      char BIGSTK tempo[HTS_URLMAXSIZE * 2 + 64];
       if (makeindex_links == 1) {
         char BIGSTK link_escaped[HTS_URLMAXSIZE * 2];
         escape_uri_utf(makeindex_firstlink, link_escaped, sizeof(link_escaped));
-        snprintf(tempo, sizeof(tempo),
-                 "<meta HTTP-EQUIV=\"Refresh\" CONTENT=\"0; URL=%s\">" CRLF,
-                 link_escaped);
+        /* no redirect beats one pointing at a clipped URL */
+        if (!sprintfbuff(
+                tempo,
+                "<meta HTTP-EQUIV=\"Refresh\" CONTENT=\"0; URL=%s\">" CRLF,
+                link_escaped)) {
+          hts_log_print(opt, LOG_WARNING,
+                        "index redirect omitted: first link too long (%s)",
+                        makeindex_firstlink);
+          tempo[0] = '\0';
+        }
       } else
         tempo[0] = '\0';
       hts_template_format(*makeindex_fp, template_footer,
@@ -479,6 +490,7 @@ void hts_finish_html_file(httrackp *opt, cache_back *cache, htsblk *r,
                           const char *adr, const char *fil, const char *save) {
   {
     file_notify(opt, adr, fil, save, 1, 1, r->notmodified);
+    hts_changes_html(opt, cache, r, adr, fil, save);
     *fp = filecreate(&opt->state.strc, save);
     if (*fp) {
       if (ht_len > 0 && fwrite(ht_buff, 1, ht_len, *fp) != ht_len) {
@@ -682,6 +694,9 @@ int httpmirror(char *url1, httrackp * opt) {
   // hash table
   opt->hash = &hash;
 
+  // a change report left by a previous crawl on this opt is not this one's
+  hts_changes_free_opt(opt);
+
   // initialize link heap
   hts_record_init(opt);
 
@@ -2080,9 +2095,12 @@ int httpmirror(char *url1, httrackp * opt) {
   if (cache.lst) {
     fclose(cache.lst);
     cache.lst = opt->state.strc.lst = NULL;
-    if (opt->delete_old) {
+    /* old.lst minus new.lst is the set of files the previous mirror had and
+       this one does not. --changes reports it; only --purge-old acts on it. */
+    if (opt->delete_old || opt->changes) {
       FILE *old_lst, *new_lst;
 
+      hts_changes_indexed(opt);
       //
       opt->state._hts_in_html_parsing = 3;
       //
@@ -2108,21 +2126,37 @@ int httpmirror(char *url1, httrackp * opt) {
               int purge = 0;
 
               while(!feof(old_lst)) {
-                linput(old_lst, line, 1000);
-                if (!strstr(adr, line)) { // not found in the new list?
-                  char BIGSTK file[HTS_URLMAXSIZE * 2];
+                char BIGSTK file[HTS_URLMAXSIZE * 2];
 
-                  strcpybuff(file, StringBuff(opt->path_html));
-                  strcatbuff(file, line + 1);
-                  file[strlen(file) - 1] = '\0';
-                  if (fexist_utf8(file)) { // toujours sur disque: virer
-                    hts_log_print(opt, LOG_INFO, "Purging %s", file);
-                    UNLINK(file);
-                    purge = 1;
+                linput(old_lst, line, 1000);
+                if (!strnotempty(line))
+                  continue;
+                strcpybuff(file, StringBuff(opt->path_html));
+                strcatbuff(file, line + 1);
+                file[strlen(file) - 1] = '\0';
+                hts_changes_previous(opt, file + StringLength(opt->path_html));
+                if (!strstr(adr, line)) { // not found in the new list?
+                  if (fexist_utf8(file)) { // still on disk
+                    /* A link this crawl did try but never wrote (a transfer
+                       killed mid-flight) also drops out of new.lst. Unless it
+                       is about to be purged, its previous copy stands and the
+                       file is not gone. */
+                    const hts_boolean kept =
+                        !opt->delete_old &&
+                        hash_read(opt->hash, file, NULL,
+                                  HASH_STRUCT_FILENAME) >= 0;
+
+                    hts_changes_dropped(
+                        opt, file + StringLength(opt->path_html), kept);
+                    if (opt->delete_old) {
+                      hts_log_print(opt, LOG_INFO, "Purging %s", file);
+                      UNLINK(file);
+                      purge = 1;
+                    }
                   }
                 }
               }
-              {
+              if (opt->delete_old) { // emptied directories go with the files
                 fseek(old_lst, 0, SEEK_SET);
                 while(!feof(old_lst)) {
                   linput(old_lst, line, 1000);
@@ -2157,7 +2191,7 @@ int httpmirror(char *url1, httrackp * opt) {
                 }
               }
               //
-              if (!purge) {
+              if (opt->delete_old && !purge) {
                 hts_log_print(opt, LOG_INFO, "No files purged");
               }
             }
@@ -2173,6 +2207,10 @@ int httpmirror(char *url1, httrackp * opt) {
   }
   // fin purge!
 
+  /* --single-file: inline each page's assets now the tree is final, after the
+     purge has deleted whatever this run dropped. */
+  singlefile_process_mirror(opt);
+
   // Indexation
   if (opt->kindex)
     index_finish(StringBuff(opt->path_html), opt->kindex);
@@ -2247,6 +2285,7 @@ int httpmirror(char *url1, httrackp * opt) {
   // ending
   usercommand(opt, 0, NULL, NULL, NULL, NULL);
   warc_close_opt(opt);
+  hts_changes_close_opt(opt);
 
   // désallocation mémoire & buffers
   XH_uninit;
@@ -2872,6 +2911,17 @@ int filecreateempty(filenote_strc * strc, const char *filename) {
     return 0;
 }
 
+void hts_savename_listed(const char *root, const char *s, char *dest,
+                         size_t destsize) {
+  char catbuff[CATBUFF_SIZE];
+
+  strlcpybuff(dest, fslash(catbuff, sizeof(catbuff), s), destsize);
+  if (strnotempty(root) && strncmp(fslash(catbuff, sizeof(catbuff), root), dest,
+                                   strlen(root)) == 0) {
+    strlcpybuff(dest, s + strlen(root), destsize);
+  }
+}
+
 // noter fichier
 int filenote(filenote_strc * strc, const char *s, filecreate_params * params) {
   // gestion du fichier liste liste
@@ -2881,15 +2931,8 @@ int filenote(filenote_strc * strc, const char *s, filecreate_params * params) {
     return 0;
   } else if (strc->lst) {
     char BIGSTK savelst[HTS_URLMAXSIZE * 2];
-    char catbuff[CATBUFF_SIZE];
 
-    strcpybuff(savelst, fslash(catbuff, sizeof(catbuff), s));
-    // couper chemin?
-    if (strnotempty(strc->path)) {
-      if (strncmp(fslash(catbuff, sizeof(catbuff), strc->path), savelst, strlen(strc->path)) == 0) {     // couper
-        strcpybuff(savelst, s + strlen(strc->path));
-      }
-    }
+    hts_savename_listed(strc->path, s, savelst, sizeof(savelst));
     fprintf(strc->lst, "[%s]" LF, savelst);
     fflush(strc->lst);
   }
@@ -2899,6 +2942,9 @@ int filenote(filenote_strc * strc, const char *s, filecreate_params * params) {
 /* Note: utf-8 */
 void file_notify(httrackp * opt, const char *adr, const char *fil,
                  const char *save, int create, int modify, int not_updated) {
+  hts_changes_notify(opt, adr, fil, save,
+                     (create || modify) ? HTS_TRUE : HTS_FALSE,
+                     not_updated ? HTS_TRUE : HTS_FALSE);
   RUN_CALLBACK6(opt, filesave2, adr, fil, save, create, modify, not_updated);
 }
 
@@ -3633,6 +3679,11 @@ HTSEXT_API int copy_htsopt(const httrackp * from, httrackp * to) {
   to->warc_max_size = from->warc_max_size;
   to->warc_cdx = from->warc_cdx;
   to->warc_wacz = from->warc_wacz;
+  to->changes = from->changes;
+
+  to->single_file = from->single_file;
+  if (from->single_file_max_size > 0)
+    to->single_file_max_size = from->single_file_max_size;
 
   if (from->pause_max_ms > 0) {
     to->pause_min_ms = from->pause_min_ms;

src/htscore.h

@@ -354,6 +354,12 @@ int filecreateempty(filenote_strc * strct, const char *filename);
 
 int filenote(filenote_strc * strct, const char *s, filecreate_params * params);
 
+/* Copy into dest (destsize bytes) the form under which the local path s is
+   listed in new.lst: forward slashes, with the mirror `root` stripped when s
+   sits under it. Also the change report's key, so the two must not drift. */
+void hts_savename_listed(const char *root, const char *s, char *dest,
+                         size_t destsize);
+
 void file_notify(httrackp * opt, const char *adr, const char *fil,
                  const char *save, int create, int modify, int wasupdated);
 
@@ -409,6 +415,8 @@ char *readfile2(const char *fil, LLint * size);
 
 char *readfile_utf8(const char *fil);
 
+char *readfile2_utf8(const char *fil, LLint *size);
+
 char *readfile_or(const char *fil, const char *defaultdata);
 
 /* Backing (download-slot) scheduler. Operate on the back[] ring (struct_back).

src/htscoremain.c

@@ -41,6 +41,7 @@ Please visit our Website: http://www.httrack.com
 #include "htsdefines.h"
 #include "htsalias.h"
 #include "htswarc.h"
+#include "htschanges.h"
 #include "htsbauth.h"
 #include "htswrap.h"
 #include "htsmodules.h"
@@ -1747,6 +1748,13 @@ static int hts_main_internal(int argc, char **argv, httrackp * opt) {
                   StringCopy(opt->cookies_file, argv[na]);
                 }
                 break;
+              case 'd': // --changes: report what this crawl changed
+                opt->changes = HTS_TRUE;
+                if (*(com + 1) == '0') {
+                  opt->changes = HTS_FALSE;
+                  com++;
+                }
+                break;
               case 'r': // warc / warc-file: write an ISO-28500 WARC archive
                 if (*(com + 1) == 'f') { // --warc-file NAME: explicit basename
                   com++;
@@ -1795,6 +1803,36 @@ static int hts_main_internal(int argc, char **argv, httrackp * opt) {
                   StringCopy(opt->warc_file, WARC_AUTONAME);
                 }
                 break;
+              case 'Z': // single-file: inline each page's assets as data: URIs
+                if (*(com + 1) == 's') { // --single-file-max-size N
+                  com++;
+                  if ((na + 1 >= argc) || (argv[na + 1][0] == '-')) {
+                    HTS_PANIC_PRINTF(
+                        "Option single-file-max-size needs a blank "
+                        "space and a size");
+                    htsmain_free();
+                    return -1;
+                  }
+                  na++;
+                  { // reject non-numeric/negative/overflow; keep the default
+                    char *end;
+                    LLint v;
+
+                    errno = 0;
+                    v = strtoll(argv[na], &end, 10);
+                    if (isdigit((unsigned char) argv[na][0]) && *end == '\0' &&
+                        errno != ERANGE && v > 0)
+                      opt->single_file_max_size = v;
+                  }
+                  opt->single_file = HTS_TRUE;
+                } else {
+                  opt->single_file = HTS_TRUE;
+                  if (*(com + 1) == '0') {
+                    opt->single_file = HTS_FALSE;
+                    com++;
+                  }
+                }
+                break;
               case 'Y': // why: explain the filter verdict for a URL, no crawl
                 if ((na + 1 >= argc) || (argv[na + 1][0] == '-')) {
                   HTS_PANIC_PRINTF("Option why needs a blank space and a URL");

src/htsftp.c

@@ -251,7 +251,7 @@ int run_launch_ftp(FTPDownloadStruct * pStruct) {
       // folding a nonsense port into 1..65535 fetches one the link never named;
       // an empty "host:" just means the default (#614)
       if (a[1] != '\0' && !hts_parse_url_port(a + 1, &port)) {
-        snprintf(back->r.msg, sizeof(back->r.msg), "Invalid port: %s", a + 1);
+        htsblk_failf(&back->r, "Invalid port: %s", a + 1);
         back->r.statuscode = STATUSCODE_INVALID; // permanent, unlike a DNS miss
         _HALT_FTP return 0;
       }
@@ -262,8 +262,7 @@ int run_launch_ftp(FTPDownloadStruct * pStruct) {
     // récupérer adresse résolue
     strcpybuff(back->info, "host name");
     if (hts_dns_resolve2(opt, _adr, &server, &error) == NULL) {
-      snprintf(back->r.msg, sizeof(back->r.msg),
-               "Unable to get server's address: %s", error);
+      htsblk_failf(&back->r, "Unable to get server's address: %s", error);
       back->r.statuscode = STATUSCODE_NON_FATAL;
       _HALT_FTP return 0;
     }
@@ -332,18 +331,15 @@ int run_launch_ftp(FTPDownloadStruct * pStruct) {
             }
 
           } else {
-            snprintf(back->r.msg, sizeof(back->r.msg), "Bad password: %s",
-                     linejmp(line));
+            htsblk_failf(&back->r, "Bad password: %s", linejmp(line));
             back->r.statuscode = STATUSCODE_INVALID;
           }
         } else {
-          snprintf(back->r.msg, sizeof(back->r.msg), "Bad user name: %s",
-                   linejmp(line));
+          htsblk_failf(&back->r, "Bad user name: %s", linejmp(line));
           back->r.statuscode = STATUSCODE_INVALID;
         }
       } else {
-        snprintf(back->r.msg, sizeof(back->r.msg), "Connection refused: %s",
-                 linejmp(line));
+        htsblk_failf(&back->r, "Connection refused: %s", linejmp(line));
         back->r.statuscode = STATUSCODE_INVALID;
       }
 
@@ -410,8 +406,7 @@ int run_launch_ftp(FTPDownloadStruct * pStruct) {
             }
             // -- fin analyse de l'adresse IP et du port --
           } else {
-            snprintf(back->r.msg, sizeof(back->r.msg), "PASV incorrect: %s",
-                     linejmp(line));
+            htsblk_failf(&back->r, "PASV incorrect: %s", linejmp(line));
             back->r.statuscode = STATUSCODE_INVALID;
           }                     // sinon on est prêts
         } else {
@@ -442,13 +437,11 @@ int run_launch_ftp(FTPDownloadStruct * pStruct) {
                 }
               }
             } else {
-              snprintf(back->r.msg, sizeof(back->r.msg), "EPSV incorrect: %s",
-                       linejmp(line));
+              htsblk_failf(&back->r, "EPSV incorrect: %s", linejmp(line));
               back->r.statuscode = STATUSCODE_INVALID;
             }
           } else {
-            snprintf(back->r.msg, sizeof(back->r.msg), "PASV/EPSV error: %s",
-                     linejmp(line));
+            htsblk_failf(&back->r, "PASV/EPSV error: %s", linejmp(line));
             back->r.statuscode = STATUSCODE_INVALID;
           }                     // sinon on est prêts
         }
@@ -554,8 +547,8 @@ int run_launch_ftp(FTPDownloadStruct * pStruct) {
                   deletesoc(soc_dat);
                   soc_dat = INVALID_SOCKET;
                   //
-                  snprintf(back->r.msg, sizeof(back->r.msg),
-                           "RETR command error: %s", linejmp(line));
+                  htsblk_failf(&back->r, "RETR command error: %s",
+                               linejmp(line));
                   back->r.statuscode = STATUSCODE_INVALID;
                 }               // sinon on est prêts
               } else {
@@ -573,13 +566,12 @@ int run_launch_ftp(FTPDownloadStruct * pStruct) {
               back->r.statuscode = STATUSCODE_INVALID;
             }                   // sinon on est prêts
           } else {
-            snprintf(back->r.msg, sizeof(back->r.msg),
-                     "Unable to resolve IP %s: %s", adr_ip, error);
+            htsblk_failf(&back->r, "Unable to resolve IP %s: %s", adr_ip,
+                         error);
             back->r.statuscode = STATUSCODE_INVALID;
           }                     // sinon on est prêts
         } else {
-          snprintf(back->r.msg, sizeof(back->r.msg), "PASV incorrect: %s",
-                   linejmp(line));
+          htsblk_failf(&back->r, "PASV incorrect: %s", linejmp(line));
           back->r.statuscode = STATUSCODE_INVALID;
         }                       // sinon on est prêts
 #else
@@ -603,13 +595,11 @@ int run_launch_ftp(FTPDownloadStruct * pStruct) {
                 back->r.statuscode = STATUSCODE_INVALID;
               }
             } else {
-              snprintf(back->r.msg, sizeof(back->r.msg),
-                       "RETR command error: %s", linejmp(line));
+              htsblk_failf(&back->r, "RETR command error: %s", linejmp(line));
               back->r.statuscode = STATUSCODE_INVALID;
             }
           } else {
-            snprintf(back->r.msg, sizeof(back->r.msg), "PORT command error: %s",
-                     linejmp(line));
+            htsblk_failf(&back->r, "PORT command error: %s", linejmp(line));
             back->r.statuscode = STATUSCODE_INVALID;
           }
 #ifdef _WIN32
@@ -652,8 +642,7 @@ int run_launch_ftp(FTPDownloadStruct * pStruct) {
                 len = 0;        // fin
                 break;
               case 0:
-                snprintf(back->r.msg, sizeof(back->r.msg), "Time out (%d)",
-                         timeout);
+                htsblk_failf(&back->r, "Time out (%d)", timeout);
                 back->r.statuscode = STATUSCODE_INVALID;
                 len = 0;        // fin
                 break;
@@ -716,8 +705,7 @@ int run_launch_ftp(FTPDownloadStruct * pStruct) {
                 strcpybuff(back->r.msg, "OK");
                 back->r.statuscode = HTTP_OK;
               } else {
-                snprintf(back->r.msg, sizeof(back->r.msg), "RETR incorrect: %s",
-                         linejmp(line));
+                htsblk_failf(&back->r, "RETR incorrect: %s", linejmp(line));
                 back->r.statuscode = STATUSCODE_INVALID;
               }
             } else {

src/htsglobal.h

@@ -72,7 +72,8 @@ Please visit our Website: http://www.httrack.com
    HTS_UNUSED: suppress unused-symbol warnings. HTS_STATIC: an unused-safe
    static. HTS_PRINTF_FUN(fmt, arg): mark a printf-like function so the
    compiler type-checks the format string at argument index fmt against the
-   varargs starting at arg. */
+   varargs starting at arg. HTS_CHECK_RESULT: the return value carries the only
+   error signal, so dropping it is a bug; a (void) cast does not silence it. */
 #ifndef HTS_UNUSED
 #ifdef __GNUC__
 #define HTS_UNUSED __attribute__((unused))
@@ -80,10 +81,13 @@ Please visit our Website: http://www.httrack.com
 #define HTS_STATIC static __attribute__((unused))
 
 #define HTS_PRINTF_FUN(fmt, arg) __attribute__((format(printf, fmt, arg)))
+
+#define HTS_CHECK_RESULT __attribute__((warn_unused_result))
 #else
 #define HTS_UNUSED
 #define HTS_STATIC static
 #define HTS_PRINTF_FUN(fmt, arg)
+#define HTS_CHECK_RESULT
 #endif
 #endif
 

src/htshelp.c

@@ -124,7 +124,9 @@ typedef struct help_wizard_buffers {
   char stropt[2048];        // options
   char stropt2[2048];       // options longues
   char strwild[2048];       // wildcards
-  char cmd[4096];
+  /* holds all four of the above plus separators: at 4096 a long answer set
+     clipped the filters off the command line */
+  char cmd[HTS_URLMAXSIZE * 2 + 3 * 2048 + 4];
   char str[256];
   char *argv[256];
 } help_wizard_buffers;
@@ -308,7 +310,14 @@ void help_wizard(httrackp * opt) {
       printf("\n");
       if (strlen(stropt) == 1)
         stropt[0] = '\0';       // aucune
-      snprintf(cmd, sizeof(cmd), "%s %s %s %s", urls, stropt, stropt2, strwild);
+      /* the tail is the filter list, and cmd is split into the argv handed to
+         hts_main() below: a clipped line would silently widen the crawl */
+      if (!sprintfbuff(cmd, "%s %s %s %s", urls, stropt, stropt2, strwild)) {
+        printf("* command line too long (%d bytes max)\n",
+               (int) sizeof(cmd) - 1);
+        freet(buffers);
+        return;
+      }
       printf("---> Wizard command line: httrack %s\n\n", cmd);
       printf("Ready to launch the mirror? (Y/n) :");
       fflush(stdout);
@@ -424,7 +433,8 @@ void help_catchurl(const char *dest_path) {
       }
       // former URL!
       {
-        char BIGSTK finalurl[HTS_URLMAXSIZE * 2];
+        /* url and dest are each HTS_URLMAXSIZE*2, plus the POSTTOK marker */
+        char BIGSTK finalurl[HTS_URLMAXSIZE * 4 + 32];
 
         inplace_escape_check_url(dest, sizeof(dest));
         snprintf(finalurl, sizeof(finalurl), "%s" POSTTOK "file:%s", url, dest);
@@ -525,6 +535,13 @@ void help(const char *app, int more) {
   infomsg
     (" %D  cached delayed type check, don't wait for remote type during updates, to speedup them (%D0 wait, * %D1 don't wait)");
   infomsg(" %M  generate a RFC MIME-encapsulated full-archive (.mht)");
+  infomsg(" %Z  after the mirror, rewrite each saved page with its "
+          "stylesheets, scripts, images and fonts inlined as data: URIs, so "
+          "any page opens by double-click anywhere (links between pages stay "
+          "relative; audio and video stay links); --single-file-max-size N "
+          "caps each asset (default 10485760 bytes). %M is the better "
+          "container where a Chromium-family browser is a given: one archive, "
+          "no base64 tax on text, a shared asset stored once");
   infomsg(" %t  keep the original file extension, don't rewrite it from the "
           "MIME type (%t0 rewrite)");
   infomsg
@@ -598,6 +615,8 @@ void help(const char *app, int more) {
           "output name, --warc-max-size N rotates segments past N bytes, "
           "--warc-cdx also writes a sorted CDXJ index, --wacz packages it all "
           "as a WACZ file");
+  infomsg(" %d  write hts-changes.json listing what this crawl left new, "
+          "changed, unchanged and gone compared to the previous mirror");
   infomsg(" %n  do not re-download locally erased files");
   infomsg
     (" %v  display on screen filenames downloaded (in realtime) - * %v1 short version - %v2 full animation");

src/htslib.c

@@ -37,6 +37,8 @@ Please visit our Website: http://www.httrack.com
 
 #include "htscore.h"
 #include "htswarc.h"
+#include "htschanges.h"
+#include "htssinglefile.h"
 
 /* specific definitions */
 #include "htsbase.h"
@@ -708,11 +710,11 @@ T_SOC http_xfopen(httrackp *opt, int mode, int treat, int waitconnect,
 #ifdef _WIN32
         int last_errno = WSAGetLastError();
 
-        sprintf(retour->msg, "Connect error: %s", strerror(last_errno));
+        htsblk_failf(retour, "Connect error: %s", strerror(last_errno));
 #else
         int last_errno = errno;
 
-        sprintf(retour->msg, "Connect error: %s", strerror(last_errno));
+        htsblk_failf(retour, "Connect error: %s", strerror(last_errno));
 #endif
       }
     }
@@ -2240,13 +2242,13 @@ T_SOC newhttp_addr(httrackp *opt, const char *_iadr, htsblk *retour, int port,
 #ifdef _WIN32
         int last_errno = WSAGetLastError();
 
-        sprintf(retour->msg, "Unable to create a socket: %s",
-                strerror(last_errno));
+        htsblk_failf(retour, "Unable to create a socket: %s",
+                     strerror(last_errno));
 #else
         int last_errno = errno;
 
-        sprintf(retour->msg, "Unable to create a socket: %s",
-                strerror(last_errno));
+        htsblk_failf(retour, "Unable to create a socket: %s",
+                     strerror(last_errno));
 #endif
       }
       return INVALID_SOCKET;    // erreur création socket impossible
@@ -2312,13 +2314,13 @@ T_SOC newhttp_addr(httrackp *opt, const char *_iadr, htsblk *retour, int port,
 #ifdef _WIN32
           const int last_errno = WSAGetLastError();
 
-          sprintf(retour->msg, "Unable to connect to the server: %s",
-                  strerror(last_errno));
+          htsblk_failf(retour, "Unable to connect to the server: %s",
+                       strerror(last_errno));
 #else
           const int last_errno = errno;
 
-          sprintf(retour->msg, "Unable to connect to the server: %s",
-                  strerror(last_errno));
+          htsblk_failf(retour, "Unable to connect to the server: %s",
+                       strerror(last_errno));
 #endif
         }
         /* Close the socket and notify the error!!! */
@@ -2698,6 +2700,24 @@ void time_gmt_rfc822(char *s) {
   time_rfc822(s, A);
 }
 
+void hts_now_iso8601(char out[32]) {
+  time_t t = time(NULL);
+  struct tm tmv;
+
+#if defined(_WIN32)
+  struct tm *g = gmtime(&t);
+
+  if (g != NULL)
+    tmv = *g;
+  else
+    memset(&tmv, 0, sizeof(tmv));
+#else
+  if (gmtime_r(&t, &tmv) == NULL)
+    memset(&tmv, 0, sizeof(tmv));
+#endif
+  strftime(out, 32, "%Y-%m-%dT%H:%M:%SZ", &tmv);
+}
+
 // heure actuelle, format rfc (taille buffer 256o)
 void time_local_rfc822(char *s) {
   time_t tt;
@@ -6011,6 +6031,10 @@ HTSEXT_API httrackp *hts_create_opt(void) {
   StringCopy(opt->cookies_file, "");
   StringCopy(opt->warc_file, "");
   opt->warc_max_size = 0; /* no rotation unless --warc-max-size sets it */
+  opt->changes = HTS_FALSE;
+  opt->changes_state = NULL;
+  opt->single_file = HTS_FALSE;
+  opt->single_file_max_size = SINGLEFILE_DEFAULT_MAX_SIZE;
   StringCopy(opt->why_url, "");
   opt->pause_min_ms = 0;
   opt->pause_max_ms = 0;
@@ -6164,6 +6188,8 @@ HTSEXT_API void hts_free_opt(httrackp * opt) {
     StringFree(opt->why_url);
     StringFree(opt->warc_file);
 
+    hts_changes_free_opt(opt);
+
     StringFree(opt->path_html);
     StringFree(opt->path_html_utf8);
     StringFree(opt->path_log);

src/htslib.h

@@ -199,6 +199,11 @@ T_SOC newhttp(httrackp * opt, const char *iadr, htsblk * retour, int port,
    etc.). */
 T_SOC newhttp_addr(httrackp *opt, const char *iadr, htsblk *retour, int port,
                    int waitconnect, int addr_index, int *addr_count);
+/* Clips the formatted failure reason into r->msg, which also round-trips
+   through the cache as X-StatusMessage. Leaves r->statuscode to the caller. */
+#define htsblk_failf(R, ...)                                                   \
+  slprintfbuff_clip((R)->msg, sizeof((R)->msg), __VA_ARGS__)
+
 HTS_INLINE void deletehttp(htsblk * r);
 HTS_INLINE int deleteaddr(htsblk * r);
 HTS_INLINE void deletesoc(T_SOC soc);
@@ -253,6 +258,10 @@ void sec2str(char *s, TStamp t);
 
 void time_gmt_rfc822(char *s);
 void time_local_rfc822(char *s);
+
+/* Current UTC time as "YYYY-MM-DDThh:mm:ssZ". */
+void hts_now_iso8601(char out[32]);
+
 struct tm *convert_time_rfc822(struct tm *buffer, const char *s);
 int set_filetime(const char *file, struct tm *tm_time);
 int set_filetime_rfc822(const char *file, const char *date);

src/htsopt.h

@@ -547,6 +547,16 @@ struct httrackp {
                              archive. Tail: ABI */
   hts_boolean warc_wacz; /**< --wacz: package archive+index+pages as a WACZ zip
                               (implies --warc + --warc-cdx). Tail: ABI */
+  hts_boolean changes;   /**< --changes: report what this crawl changed against
+                              the previous mirror. Tail: ABI */
+  void *changes_state;   /**< live change-report accumulator (hts_changes*),
+                              engine-owned. Tail: ABI */
+  hts_boolean single_file; /**< --single-file: once the mirror is done, rewrite
+                                each saved page with its assets inlined as
+                                data: URIs. Tail: ABI */
+  LLint single_file_max_size; /**< --single-file-max-size: per-asset cap in
+                                   bytes; a bigger asset stays a link.
+                                   Tail: ABI */
 };
 
 /* Running statistics for a mirror. */

src/htsproxy.c

@@ -173,8 +173,8 @@ int http_proxy_tunnel(httrackp *opt, htsblk *retour, const char *adr,
   if (sscanf(line, "HTTP/%*d.%*d %d", &code) < 1)
     code = 0;
   if (code < 200 || code >= 300) {
-    snprintf(retour->msg, sizeof(retour->msg), "proxy CONNECT refused: %s",
-             strnotempty(line) ? line : "(no status)");
+    htsblk_failf(retour, "proxy CONNECT refused: %s",
+                 strnotempty(line) ? line : "(no status)");
     return 0;
   }
 

src/htssafe.h

@@ -33,6 +33,7 @@ Please visit our Website: http://www.httrack.com
 #ifndef HTSSAFE_DEFH
 #define HTSSAFE_DEFH
 
+#include <stdarg.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
@@ -459,6 +460,85 @@ static HTS_INLINE HTS_UNUSED const char *htsbuff_str(const htsbuff *b) {
   return b->buf;
 }
 
+/**
+ * Copy src into dest (capacity size, NUL included), truncating to fit and
+ * always NUL-terminating. Unlike strlcpybuff() it never aborts, so it suits a
+ * value read back from a cache, a header or the wire, where refusing the whole
+ * record is worse than keeping a clipped one. Returns HTS_TRUE if it all fit;
+ * callers that clip on purpose ignore that, so it is not HTS_CHECK_RESULT.
+ */
+static HTS_INLINE HTS_UNUSED hts_boolean strclipbuff(char *dest, size_t size,
+                                                     const char *src) {
+  size_t len, copy;
+
+  assertf(dest != NULL && src != NULL && size != 0);
+  len = strlen(src);
+  copy = len < size ? len : size - 1;
+  memcpy(dest, src, copy);
+  dest[copy] = '\0';
+  return copy == len ? HTS_TRUE : HTS_FALSE;
+}
+
+/**
+ * Callers that deliberately ignore truncation use this instead of
+ * slprintfbuff(), so it is not HTS_CHECK_RESULT.
+ */
+static HTS_INLINE HTS_UNUSED HTS_PRINTF_FUN(3, 0) hts_boolean
+    vslprintfbuff(char *dest, size_t size, const char *fmt, va_list args) {
+  int ret;
+
+  assertf(dest != NULL && size != 0);
+  /* a vsnprintf failing outright may write nothing at all, leaving whatever
+     the caller had on the stack for it to publish */
+  dest[0] = '\0';
+  ret = vsnprintf(dest, size, fmt, args);
+  /* pre-C99 runtimes (msvcrt _vsnprintf) return -1 and do not terminate */
+  dest[size - 1] = '\0';
+  return ret >= 0 && (size_t) ret < size ? HTS_TRUE : HTS_FALSE;
+}
+
+/**
+ * Formatted print into dest (capacity size, NUL included), truncating to fit
+ * and always NUL-terminating. Returns HTS_TRUE if the whole output fit; the
+ * result is the only truncation signal, so it must be acted on. Unlike
+ * strcpybuff() it never aborts, so it suits text built from remote input.
+ */
+static HTS_INLINE HTS_UNUSED HTS_CHECK_RESULT HTS_PRINTF_FUN(3, 4) hts_boolean
+    slprintfbuff(char *dest, size_t size, const char *fmt, ...) {
+  va_list args;
+  hts_boolean ret;
+
+  va_start(args, fmt);
+  ret = vslprintfbuff(dest, size, fmt, args);
+  va_end(args);
+  return ret;
+}
+
+/**
+ * slprintfbuff() for diagnostics quoting remote or client text, which are
+ * meant to be clipped: nothing to act on, hence not HTS_CHECK_RESULT. A (void)
+ * cast on slprintfbuff() is no substitute, GCC warns through it.
+ */
+static HTS_INLINE HTS_UNUSED HTS_PRINTF_FUN(3, 4) void slprintfbuff_clip(
+    char *dest, size_t size, const char *fmt, ...) {
+  va_list args;
+
+  va_start(args, fmt);
+  (void) vslprintfbuff(dest, size, fmt, args);
+  va_end(args);
+}
+
+/**
+ * slprintfbuff() over the in-scope array ARR (capacity = sizeof(ARR)).
+ * On GCC/Clang a pointer is a compile error; use slprintfbuff() for those.
+ */
+#if (defined(__GNUC__) && !defined(__cplusplus))
+#define sprintfbuff(ARR, ...)                                                  \
+  slprintfbuff((ARR), sizeof(ARR) + htsbuff_must_be_array_(ARR), __VA_ARGS__)
+#else
+#define sprintfbuff(ARR, ...) slprintfbuff((ARR), sizeof(ARR), __VA_ARGS__)
+#endif
+
 /* Thin aliases over the libc allocator/memcpy (historical "t" suffix); no
    added bounds checking. freet() also NULLs the freed pointer and tolerates
    NULL. memcpybuff() despite the name is a raw memcpy: the caller owns the

src/htsserver.c

@@ -318,6 +318,18 @@ typedef struct {
   error_redirect = "/server/error.html"; \
 } while(0)
 
+/* Longest error message shown on the error page; the rest is clipped. */
+#define ERROR_MESSAGE_MAX 1024
+
+/* SET_ERROR() with a printf format. Clips: these messages quote posted fields,
+   whose length the client picks. */
+#define SET_ERRORF(...)                                                        \
+  do {                                                                         \
+    char errbuf[ERROR_MESSAGE_MAX];                                            \
+    slprintfbuff_clip(errbuf, sizeof(errbuf), __VA_ARGS__);                    \
+    SET_ERROR(errbuf);                                                         \
+  } while (0)
+
 /* Longest "sid" value worth unescaping: the expected one is an md5 hex digest,
    so anything near this is already invalid and is rejected unread. */
 #define SID_VALUE_MAX 64
@@ -363,6 +375,59 @@ static hts_boolean body_sid_is_valid(const char *body, const char *expected) {
   return seen;
 }
 
+/** Append src to the NUL-terminated dst of capacity size (NUL included).
+    False, leaving dst untouched, if it would not fit: unlike strcatbuff() this
+    never aborts, because every piece appended here is client-supplied. */
+static hts_boolean path_append(char *dst, size_t size, const char *src) {
+  const size_t used = strlen(dst);
+  const size_t len = strlen(src);
+
+  /* dst holds at most size-1 bytes, so "size - used" is >= 1 and the untrusted
+     len stays alone: "used + len < size" could wrap and pass. */
+  if (len >= size - used) {
+    return HTS_FALSE;
+  }
+  memcpy(dst + used, src, len + 1);
+  return HTS_TRUE;
+}
+
+/* Append c to dst as an HTML entity, or return HTS_FALSE if it needs none. */
+static hts_boolean cat_html_escaped(String *dst, char c) {
+  switch (c) {
+  case '<':
+    StringCat(*dst, "&lt;");
+    break;
+  case '>':
+    StringCat(*dst, "&gt;");
+    break;
+  case '&':
+    StringCat(*dst, "&amp;");
+    break;
+  case '\'':
+    StringCat(*dst, "&#39;");
+    break;
+  default:
+    return HTS_FALSE;
+  }
+  return HTS_TRUE;
+}
+
+/* Append the value of a double-quoted command-line argument: escaped for HTML,
+   which the browser undoes when it posts the command line back, and for the
+   argv splitter, which does not. */
+static void cat_cmdline_arg(String *output, const char *value) {
+  const char *a;
+
+  for (a = value; *a != '\0'; a++) {
+    if (*a == '\\' || *a == '\"') {
+      StringCat(*output, "\\");
+    }
+    if (!cat_html_escaped(output, *a)) {
+      StringMemcat(*output, a, 1);
+    }
+  }
+}
+
 int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
   int timeout = 30;
   int retour = 0;
@@ -374,6 +439,9 @@ int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
   String tmpbuff = STRING_EMPTY;
   String tmpbuff2 = STRING_EMPTY;
   String fspath = STRING_EMPTY;
+  /* Project directory this server set up; the only root /website/ serves from,
+     and deliberately not cleared between requests. */
+  String website = STRING_EMPTY;
   char catbuff[CATBUFF_SIZE];
 
   /* Load strings */
@@ -669,7 +737,7 @@ int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
           fp = fopen(StringBuff(fspath), "rb");
           if (fp) {
             /* Read file */
-            while(!feof(fp)) {
+            while (!feof(fp) && !ferror(fp)) {
               char *str = line;
               char *pos;
 
@@ -785,6 +853,11 @@ int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
                   if (!structcheck(StringBuff(tmpbuff))) {
                     FILE *fp;
 
+                    /* Both halves of fspath come from posted fields, so a ".."
+                       in them would escape the mirror once served. */
+                    if (strstr(StringBuff(fspath), "..") == NULL) {
+                      StringCopy(website, StringBuff(fspath));
+                    }
                     StringCat(tmpbuff, "winprofile.ini");
                     fp = fopen(StringBuff(tmpbuff), "wb");
                     if (fp != NULL) {
@@ -818,28 +891,18 @@ int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
                           commandEnd = 1;
                         }
                       } else {
-                        char tmp[1024];
-
-                        sprintf(tmp,
-                                "Unable to write %d bytes in the the init file %s",
-                                count, StringBuff(fspath));
-                        SET_ERROR(tmp);
+                        SET_ERRORF(
+                            "Unable to write %d bytes in the the init file %s",
+                            count, StringBuff(fspath));
                       }
                       fclose(fp);
                     } else {
-                      char tmp[1024];
-
-                      sprintf(tmp, "Unable to create the init file %s",
-                              StringBuff(fspath));
-                      SET_ERROR(tmp);
+                      SET_ERRORF("Unable to create the init file %s",
+                                 StringBuff(fspath));
                     }
                   } else {
-                    char tmp[1024];
-
-                    sprintf(tmp,
-                            "Unable to create the directory structure in %s",
-                            StringBuff(fspath));
-                    SET_ERROR(tmp);
+                    SET_ERRORF("Unable to create the directory structure in %s",
+                               StringBuff(fspath));
                   }
 
                 } else {
@@ -857,7 +920,7 @@ int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
 
       /* Response */
       if (meth) {
-        int virtualpath = 0;
+        hts_boolean virtualpath = HTS_FALSE;
         char *pos;
         char *url = strchr(line1, ' ');
 
@@ -868,11 +931,11 @@ int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
           char *qpos;
 
           /* get the URL */
+          fsfile[0] = '\0';
           if (error_redirect == NULL) {
             if ((qpos = strchr(url, '?'))) {
               *qpos = '\0';
             }
-            fsfile[0] = '\0';
             if (strcmp(url, "/") == 0) {
               file = "/server/index.html";
               meth = 2;
@@ -885,7 +948,7 @@ int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
           }
 
           if (strncmp(file, "/website/", 9) == 0) {
-            virtualpath = 1;
+            virtualpath = HTS_TRUE;
           }
 
           /* override */
@@ -899,20 +962,28 @@ int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
             }
           }
 
-          if (strlen(path) + strlen(file) + 32 < sizeof(fsfile)) {
-            if (strncmp(file, "/website/", 9) != 0) {
-              sprintf(fsfile, "%shtml%s", path, file);
-            } else {
-              intptr_t adr = 0;
+          /* the override above may have swapped a mirror path for a GUI page */
+          virtualpath = strncmp(file, "/website/", 9) == 0;
 
-              if (coucal_readptr(NewLangList, "projpath", &adr)) {
-                sprintf(fsfile, "%s%s", (char *) adr, file + 9);
-              }
+          if (!virtualpath) {
+            if (!path_append(fsfile, sizeof(fsfile), path) ||
+                !path_append(fsfile, sizeof(fsfile), "html") ||
+                !path_append(fsfile, sizeof(fsfile), file)) {
+              fsfile[0] = '\0';
+            }
+          } else if (StringNotEmpty(website)) {
+            /* Never the posted "projpath": a client root reads any file. */
+            if (!path_append(fsfile, sizeof(fsfile), StringBuff(website)) ||
+                !path_append(fsfile, sizeof(fsfile), "/") ||
+                !path_append(fsfile, sizeof(fsfile), file + 9)) {
+              fsfile[0] = '\0';
             }
           }
 
-          if (fsfile[0] && strstr(file, "..") == NULL
-              && (fp = fopen(fsfile, "rb"))) {
+          /* Regular files only: reading a directory or FIFO never ends, and
+             "path" may hold "..", so only the untrusted halves are checked. */
+          if (fsfile[0] && strstr(file, "..") == NULL && fexist(fsfile) &&
+              (fp = fopen(fsfile, "rb"))) {
             char ok[] =
               "HTTP/1.0 200 OK\r\n" "Connection: close\r\n"
               "Server: httrack-small-server\r\n" "Content-type: text/html\r\n"
@@ -965,11 +1036,13 @@ int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
                 StringCat(headers, "\r\n");
               }
               coucal_write(NewLangList, "redirect", (intptr_t) NULL);
-            } else if (is_html(file)) {
+            } else if (!virtualpath && is_html(file)) {
+              /* GUI templates only: ${_sid} in a mirrored page would hand the
+                 crawled site the session id that authenticates commands */
               int outputmode = 0;
 
               StringMemcat(headers, ok, sizeof(ok) - 1);
-              while(!feof(fp)) {
+              while (!feof(fp) && !ferror(fp)) {
                 char *str = line;
                 int prevlen = (int) StringLength(output);
                 int nocr = 0;
@@ -993,6 +1066,7 @@ int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
                     int p;
                     int format = 0;
                     int listDefault = 0;
+                    hts_boolean unquoted = HTS_FALSE;
 
                     name[0] = '\0';
                     strlncatbuff(name, str, sizeof(name_), n);
@@ -1002,6 +1076,12 @@ int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
                     } else if ((p = strfield(name, "html:"))) {
                       name += p;
                       format = 1;
+                    } else if ((p = strfield(name, "unquoted:"))) {
+                      name += p;
+                      unquoted = HTS_TRUE;
+                    } else if ((p = strfield(name, "arg:"))) {
+                      name += p;
+                      format = 5;
                     } else if ((p = strfield(name, "list:"))) {
                       name += p;
                       format = 2;
@@ -1138,8 +1218,8 @@ int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
                        test:<if ==0>:<if ==1>:<if == 2>..
                        ztest:<if == 0 || !exist>:<if == 1>:<if == 2>..
                      */
-                    else if ((p = strfield(name, "test:"))
-                             || (p = strfield(name, "ztest:"))) {
+                    else if ((p = strfield(name, "test:")) ||
+                             (p = strfield(name, "ztest:"))) {
                       intptr_t adr = 0;
                       char *pos2;
                       int ztest = (name[0] == 'z');
@@ -1242,6 +1322,12 @@ int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
                         }
                       }
                     }
+                    /* consumed here: it shares nothing with the list and
+                       option formats below */
+                    if (format == 5 && langstr != NULL && outputmode != -1) {
+                      cat_cmdline_arg(&output, langstr);
+                      langstr = NULL;
+                    }
                     if (langstr && outputmode != -1) {
                       switch (format) {
                       case 0:
@@ -1259,18 +1345,18 @@ int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
                                 StringMemcat(output, &c, 1);
                               }
                               a += 2;
-                            } else if (outputmode && a[0] == '<') {
-                              StringCat(output, "&lt;");
-                            } else if (outputmode && a[0] == '>') {
-                              StringCat(output, "&gt;");
-                            } else if (outputmode && a[0] == '&') {
-                              StringCat(output, "&amp;");
-                            } else if (outputmode && a[0] == '\'') {
-                              StringCat(output, "&#39;");
+                            } else if (unquoted && a[0] == '\"') {
+                              /* the browser posts an entity back as a raw
+                                 quote, which would open a quoted run in the
+                                 argv splitter; a URI cannot hold one anyway */
+                              StringCat(output, "%22");
+                            } else if (outputmode &&
+                                       cat_html_escaped(&output, a[0])) {
+                              /* appended as an entity */
                             } else if (outputmode == 3 && a[0] == ' ') {
                               StringCat(output, "%20");
-                            } else if (outputmode >= 2
-                                       && ((unsigned char) a[0]) < 32) {
+                            } else if (outputmode >= 2 &&
+                                       ((unsigned char) a[0]) < 32) {
                               char tmp[32];
 
                               sprintf(tmp, "%%%02x", (unsigned char) a[0]);
@@ -1331,20 +1417,10 @@ int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
                               }
                               StringClear(tmpbuff);
                               break;
-                            case '<':
-                              StringCat(tmpbuff, "&lt;");
-                              break;
-                            case '>':
-                              StringCat(tmpbuff, "&gt;");
-                              break;
-                            case '&':
-                              StringCat(tmpbuff, "&amp;");
-                              break;
-                            case '\'':
-                              StringCat(tmpbuff, "&#39;");
-                              break;
                             default:
-                              StringMemcat(tmpbuff, fstr, 1);
+                              if (!cat_html_escaped(&tmpbuff, *fstr)) {
+                                StringMemcat(tmpbuff, fstr, 1);
+                              }
                               break;
                             }
                             fstr++;
@@ -1384,7 +1460,9 @@ int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
               }
 #endif
             } else {
-              if (is_text(file)) {
+              if (is_html(file)) {
+                StringMemcat(headers, ok, sizeof(ok) - 1);
+              } else if (is_text(file)) {
                 StringMemcat(headers, ok_text, sizeof(ok_text) - 1);
               } else if (is_js(file)) {
                 StringMemcat(headers, ok_js, sizeof(ok_js) - 1);
@@ -1398,14 +1476,15 @@ int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
               while(!feof(fp)) {
                 int n = (int) fread(line, 1, sizeof(line) - 2, fp);
 
-                if (n > 0) {
-                  StringMemcat(output, line, n);
+                if (n <= 0) {
+                  break; /* short read: EOF or error, never a retry */
                 }
+                StringMemcat(output, line, n);
               }
             }
             fclose(fp);
-          } else if (strcmp(file, "/ping") == 0
-                     || strncmp(file, "/ping?", 6) == 0) {
+          } else if (strcmp(file, "/ping") == 0 ||
+                     strncmp(file, "/ping?", 6) == 0) {
             char error_hdr[] =
               "HTTP/1.0 200 Pong\r\n" "Server: httrack small server\r\n"
               "Content-type: text/html\r\n";
@@ -1491,6 +1570,7 @@ int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
   StringFree(tmpbuff);
   StringFree(tmpbuff2);
   StringFree(fspath);
+  StringFree(website);
 
   if (buffer)
     free(buffer);

src/htssinglefile.h

@@ -0,0 +1,84 @@
+/* ------------------------------------------------------------ */
+/*
+HTTrack Website Copier, Offline Browser for Windows and Unix
+Copyright (C) 2026 Xavier Roche and other contributors
+
+SPDX-License-Identifier: GPL-3.0-or-later
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation, either version 3 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program. If not, see <http://www.gnu.org/licenses/>.
+
+Ethical use: we kindly ask that you NOT use this software to harvest email
+addresses or to collect any other private information about people. Doing so
+would dishonor our work and waste the many hours we have spent on it.
+
+Please visit our Website: http://www.httrack.com
+*/
+
+/* ------------------------------------------------------------ */
+/* --single-file: post-mirror pass embedding each page's assets as data: URIs.
+   Internal, not installed. Unlike -%M (one MHT archive for the whole mirror)
+   it rewrites the saved pages in place and keeps page-to-page links relative,
+   so the mirror stays browsable and every page also stands alone.
+   Limitation: an inlined stylesheet becomes a data: URL, whose path is opaque,
+   so an asset inside it that stayed a link (over the cap) no longer resolves.
+   Raise --single-file-max-size past it to inline it too. */
+/* ------------------------------------------------------------ */
+
+#ifndef HTS_SINGLEFILE_DEFH
+#define HTS_SINGLEFILE_DEFH
+
+#include "htsopt.h"
+#include "htsstrings.h"
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/* --single-file-max-size default: a bigger asset stays an ordinary link. */
+#define SINGLEFILE_DEFAULT_MAX_SIZE (10 * 1024 * 1024)
+
+/* Never base64 more than this, whatever the cap: code64() sizes in int. */
+#define SINGLEFILE_HARD_MAX_SIZE (256 * 1024 * 1024)
+
+/* Total bytes one page may inline. Nested @import fans out multiplicatively,
+   so a few hundred bytes of hostile CSS can otherwise ask for gigabytes. */
+#define SINGLEFILE_MAX_PAGE_SIZE (64 * 1024 * 1024)
+
+/* Rewrite every HTML page the mirror produced. No-op unless opt->single_file;
+   call once the tree is final, after the update purge. */
+void singlefile_process_mirror(httrackp *opt);
+
+/* Rewrite one HTML document held in memory, appending the result to out.
+   root is the mirror directory that references may not escape; page_path is
+   the document's own path under it (both UTF-8, '/' or native separators).
+   page_budget caps the total inlined bytes, since nested @import fans out
+   multiplicatively; the mirror pass passes SINGLEFILE_MAX_PAGE_SIZE.
+   Returns HTS_TRUE if at least one reference was replaced; out may still
+   differ from the input when that is HTS_FALSE, since a style or srcset value
+   is re-serialized in place. */
+hts_boolean singlefile_rewrite_html(httrackp *opt, const char *root,
+                                    const char *page_path, const char *html,
+                                    size_t html_len, LLint page_budget,
+                                    String *out);
+
+/* singlefile_rewrite_html() over the file at page_path, rewritten in place.
+   Returns HTS_TRUE if the file changed. */
+hts_boolean singlefile_rewrite_file(httrackp *opt, const char *root,
+                                    const char *page_path);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif

src/htsstats.h

@@ -0,0 +1,63 @@
+/* ------------------------------------------------------------ */
+/*
+HTTrack Website Copier, Offline Browser for Windows and Unix
+Copyright (C) 1998 Xavier Roche and other contributors
+
+SPDX-License-Identifier: GPL-3.0-or-later
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation, either version 3 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program. If not, see <http://www.gnu.org/licenses/>.
+
+Ethical use: we kindly ask that you NOT use this software to harvest email
+addresses or to collect any other private information about people. Doing so
+would dishonor our work and waste the many hours we have spent on it.
+
+Please visit our Website: http://www.httrack.com
+*/
+
+/* ------------------------------------------------------------ */
+/* File: in-progress display rows, shared by httrack and htsserver .h */
+/* Author: Xavier Roche                                         */
+/* ------------------------------------------------------------ */
+
+#ifndef HTS_STATS_DEFH
+#define HTS_STATS_DEFH
+
+#include "htsglobal.h"
+
+/* One row of the "in progress" display, shared so the CLI (httrack) and the
+   web GUI (htsserver) cannot drift apart: each fills its own array. */
+
+#define NStatsBuffer 14
+
+#ifndef HTS_DEF_FWSTRUCT_t_StatsBuffer
+#define HTS_DEF_FWSTRUCT_t_StatsBuffer
+typedef struct t_StatsBuffer t_StatsBuffer;
+#endif
+struct t_StatsBuffer {
+  char name[1024];
+  char file[1024];
+  char state[288];                         // a short label plus back->info[256]
+  char BIGSTK url_sav[HTS_URLMAXSIZE * 2]; // pour cancel
+  char BIGSTK url_adr[HTS_URLMAXSIZE * 2];
+  char BIGSTK url_fil[HTS_URLMAXSIZE * 2];
+  LLint size;
+  LLint sizetot;
+  int offset;
+  //
+  int back;
+  //
+  int actived; // pour disabled
+};
+
+#endif

src/htsthread.c

@@ -44,9 +44,18 @@ Please visit our Website: http://www.httrack.com
 #endif
 #endif
 
+/* Outstanding threads, counted at spawn rather than by the child at entry, so
+   that a caller which spawns and immediately waits still joins them (#747). */
 static int process_chain = 0;
 static htsmutex process_chain_mutex = HTSMUTEX_INIT;
 
+static void process_chain_add(int delta) {
+  hts_mutexlock(&process_chain_mutex);
+  process_chain += delta;
+  assertf(process_chain >= 0);
+  hts_mutexrelease(&process_chain_mutex);
+}
+
 HTSEXT_API void htsthread_wait(void) {
   htsthread_wait_n(0);
 }
@@ -99,20 +108,12 @@ static void *hts_entry_point(void *tharg)
   void *const arg = s_args->arg;
   void (*fun) (void *arg) = s_args->fun;
 
-  free(tharg);
-
-  hts_mutexlock(&process_chain_mutex);
-  process_chain++;
-  assertf(process_chain > 0);
-  hts_mutexrelease(&process_chain_mutex);
+  freet(tharg);
 
   /* run */
   fun(arg);
 
-  hts_mutexlock(&process_chain_mutex);
-  process_chain--;
-  assertf(process_chain >= 0);
-  hts_mutexrelease(&process_chain_mutex);
+  process_chain_add(-1);
 #ifdef _WIN32
   return 0;
 #else
@@ -122,18 +123,20 @@ static void *hts_entry_point(void *tharg)
 
 /* create a thread */
 HTSEXT_API int hts_newthread(void (*fun) (void *arg), void *arg) {
-  hts_thread_s *s_args = malloc(sizeof(hts_thread_s));
+  hts_thread_s *s_args = malloct(sizeof(hts_thread_s));
 
   assertf(s_args != NULL);
   s_args->arg = arg;
   s_args->fun = fun;
+  process_chain_add(1);
 #ifdef _WIN32
   {
     unsigned int idt;
     HANDLE handle =
       (HANDLE) _beginthreadex(NULL, 0, hts_entry_point, s_args, 0, &idt);
     if (handle == 0) {
-      free(s_args);
+      process_chain_add(-1);
+      freet(s_args);
       return -1;
     } else {
       /* detach the thread from the main process so that is can be independent */
@@ -151,7 +154,8 @@ HTSEXT_API int hts_newthread(void (*fun) (void *arg), void *arg) {
         || pthread_attr_setstacksize(&attr, stackSize) != 0
         || (retcode =
             pthread_create(&handle, &attr, hts_entry_point, s_args)) != 0) {
-      free(s_args);
+      process_chain_add(-1);
+      freet(s_args);
       return -1;
     } else {
       /* detach the thread from the main process so that is can be independent */

src/htstools.c

@@ -308,8 +308,10 @@ int lienrelatif(char *s, size_t ssize, const char *link, const char *curr_fil) {
   // copy only the current path
   curr = _curr;
   strlcpybuff(curr, curr_fil, sizeof(_curr));
-  if ((a = strchr(curr, '?')) == NULL)  // couper au ? (params)
-    a = curr + strlen(curr) - 1;        // pas de params: aller à la fin
+  if ((a = strchr(curr, '?')) == NULL) { // cut at the ? (query parameters)
+    // an empty path has no last character: curr-1 would read before the buffer
+    a = curr[0] != '\0' ? curr + strlen(curr) - 1 : curr;
+  }
   while((*a != '/') && (a > curr))
     a--;                        // chercher dernier / du chemin courant
   if (*a == '/')

src/htswarc.c

@@ -462,22 +462,6 @@ static void warc_make_id(warc_writer *w, char out[64]) {
            b[11], b[12], b[13], b[14], b[15]);
 }
 
-static void warc_now_iso8601(char out[32]) {
-  time_t t = time(NULL);
-  struct tm tmv;
-#if defined(_WIN32)
-  struct tm *g = gmtime(&t);
-  if (g != NULL)
-    tmv = *g;
-  else
-    memset(&tmv, 0, sizeof(tmv));
-#else
-  if (gmtime_r(&t, &tmv) == NULL)
-    memset(&tmv, 0, sizeof(tmv));
-#endif
-  strftime(out, 32, "%Y-%m-%dT%H:%M:%SZ", &tmv);
-}
-
 /* Case-insensitive "is this the header named name?" test, tolerating optional
    whitespace before the ':' (non-compliant "Name : value" is still matched). */
 static int header_is(const char *line, size_t line_len, const char *name) {
@@ -1063,7 +1047,7 @@ static void warc_wacz_package(warc_writer *w) {
   wbuf_free(&pages);
 
   /* datapackage.json listing every stored file with its sha256 + size. */
-  warc_now_iso8601(created);
+  hts_now_iso8601(created);
   memset(&dp, 0, sizeof(dp));
   if (!err &&
       (wbuf_printf(&dp,
@@ -1166,7 +1150,7 @@ static int warc_emit(warc_writer *w, const char *type, const char *content_type,
 
   memset(&hdr, 0, sizeof(hdr));
   warc_make_id(w, id);
-  warc_now_iso8601(date);
+  hts_now_iso8601(date);
 
 #if HTS_USEOPENSSL
   /* Block digest over the whole block, in one streaming pass. */
@@ -1670,7 +1654,7 @@ int warc_write_transaction(warc_writer *w, const char *target_uri,
     char date[32];
     char *slot;
     size_t need;
-    warc_now_iso8601(date);
+    hts_now_iso8601(date);
     need = strlen(target_uri) + 1 + strlen(date) + 1;
     slot = malloct(need);
     if (slot != NULL) {

src/htsweb.c

@@ -62,6 +62,7 @@ Please visit our Website: http://www.httrack.com
 #include "htsmd5.c"
 #include "md5.c"
 
+#include "htscmdline.h"
 #include "htsserver.h"
 #include "htsurlport.h"
 #include "htsweb.h"
@@ -319,10 +320,8 @@ int main(int argc, char *argv[]) {
 static int webhttrack_runmain(httrackp * opt, int argc, char **argv);
 static void back_launch_cmd(void *pP) {
   char *cmd = (char *) pP;
-  char **argv = (char **) malloct(1024 * sizeof(char *));
+  char **argv;
   int argc = 0;
-  int i = 0;
-  int g = 0;
 
   //
   httrackp *opt;
@@ -333,28 +332,19 @@ static void back_launch_cmd(void *pP) {
   commandReturnCmdl = strdup(cmd);
 
   /* split */
-  argv[0] = strdup("webhttrack");
-  argv[1] = cmd;
-  argc++;
-  i = 0;
-  while(cmd[i]) {
-    if (cmd[i] == '\t' || cmd[i] == '\r' || cmd[i] == '\n') {
-      cmd[i] = ' ';
-    }
-    i++;
-  }
-  i = 0;
-  while(cmd[i]) {
-    if (cmd[i] == '\"')
-      g = !g;
-    if (cmd[i] == ' ') {
-      if (!g) {
-        cmd[i] = '\0';
-        argv[argc++] = cmd + i + 1;
-      }
-    }
-    i++;
+  argv = hts_split_cmdline(cmd, &argc);
+  if (argv == NULL) {
+    if (commandReturnMsg)
+      free(commandReturnMsg);
+    commandReturnMsg = strdup("could not parse the command line");
+    commandReturn = -1;
+    commandRunning = 0;
+    commandEnd = 1;
+    free(cmd);
+    return;
   }
+  /* drop the program name the posted command line carries */
+  argv[0] = strdupt("webhttrack");
 
   /* init */
   hts_init();
@@ -383,6 +373,7 @@ static void back_launch_cmd(void *pP) {
 
   /* free */
   free(cmd);
+  freet(argv[0]);
   freet(argv);
   return;
 }

src/htsweb.h

@@ -35,26 +35,10 @@ Please visit our Website: http://www.httrack.com
 
 #include "htsglobal.h"
 #include "htscore.h"
+#include "htsstats.h"
 
-#define NStatsBuffer     14
 #define MAX_LEN_INPROGRESS 40
 
-typedef struct t_StatsBuffer {
-  char name[1024];
-  char file[1024];
-  char state[256];
-  char url_sav[HTS_URLMAXSIZE * 2];     // pour cancel
-  char url_adr[HTS_URLMAXSIZE * 2];
-  char url_fil[HTS_URLMAXSIZE * 2];
-  LLint size;
-  LLint sizetot;
-  int offset;
-  //
-  int back;
-  //
-  int actived;                  // pour disabled
-} t_StatsBuffer;
-
 typedef struct t_InpInfo {
   int ask_refresh;
   int refresh;

src/httrack.c

@@ -46,6 +46,7 @@ Please visit our Website: http://www.httrack.com
 #include "httrack.h"
 #include "htslib.h"
 #include "htscharset.h" // after htslib.h: winsock2.h must precede windows.h
+#include "htsbacktrace.h"
 
 /* Static definitions */
 static int fexist(const char *s);
@@ -72,10 +73,6 @@ static int linput(FILE * fp, char *s, int max);
 #include <sys/ioctl.h>
 #endif
 #include <ctype.h>
-#if (defined(__linux) && defined(HAVE_EXECINFO_H))
-#include <execinfo.h>
-#define USES_BACKTRACE
-#endif
 /* END specific definitions */
 
 static void __cdecl htsshow_init(t_hts_callbackarg * carg);
@@ -233,8 +230,7 @@ static hts_boolean vt_size_refresh(void) {
 */
 #define STYLE_STATVALUES VT_BOLD
 #define STYLE_STATTEXT   VT_UNBOLD
-#define STYLE_STATRESET  VT_UNBOLD
-#define NStatsBuffer     14
+#define STYLE_STATRESET VT_UNBOLD
 /* Rows the stats block and "Current job" take above the in-progress list. */
 #define NStatsHeaderRows 7
 
@@ -310,8 +306,8 @@ int main(int argc, char **argv) {
     fprintf(stderr, "* %s\n", hts_errmsg(opt));
   }
   global_opt = NULL;
+  htsthread_wait(); /* pending threads still read opt */
   hts_free_opt(opt);
-  htsthread_wait();             /* wait for pending threads */
   hts_uninit();
 
 #ifdef _WIN32
@@ -880,21 +876,6 @@ static void sig_doback(int blind) {     // mettre en backing
 #undef FD_ERR
 #define FD_ERR 2
 
-static void print_backtrace(void) {
-#ifdef USES_BACKTRACE
-  void *stack[256];
-  const int size = backtrace(stack, sizeof(stack)/sizeof(stack[0]));
-  if (size != 0) {
-    backtrace_symbols_fd(stack, size, FD_ERR);
-  }
-#else
-  const char msg[] = "No stack trace available on this OS :(\n";
-  if (write(FD_ERR, msg, sizeof(msg) - 1) != sizeof(msg) - 1) {
-    /* sorry GCC */
-  }
-#endif
-}
-
 static size_t print_num(char *buffer, int num) {
   size_t i, j;
   if (num < 0) {
@@ -928,7 +909,7 @@ static void sig_fatal(int code) {
   size += print_num(&buffer[size], code);
   buffer[size++] = '\n';
   (void) (write(FD_ERR, buffer, size) == size);
-  print_backtrace();
+  hts_print_backtrace(FD_ERR);
   (void) (write(FD_ERR, msgreport, sizeof(msgreport) - 1)
     == sizeof(msgreport) - 1);
   abort();
@@ -951,6 +932,7 @@ static void sig_leave(int code) {
 }
 
 static void signal_handlers(void) {
+  hts_backtrace_init();
 #ifdef _WIN32
   signal(SIGINT, sig_leave);    // ^C
   signal(SIGTERM, sig_finish);  // kill <process>

src/httrack.h

@@ -36,26 +36,7 @@ Please visit our Website: http://www.httrack.com
 #include "htsglobal.h"
 #include "htscore.h"
 #include "htssafe.h"
-
-#ifndef HTS_DEF_FWSTRUCT_t_StatsBuffer
-#define HTS_DEF_FWSTRUCT_t_StatsBuffer
-typedef struct t_StatsBuffer t_StatsBuffer;
-#endif
-struct t_StatsBuffer {
-  char name[1024];
-  char file[1024];
-  char state[256];
-  char BIGSTK url_sav[HTS_URLMAXSIZE * 2];      // pour cancel
-  char BIGSTK url_adr[HTS_URLMAXSIZE * 2];
-  char BIGSTK url_fil[HTS_URLMAXSIZE * 2];
-  LLint size;
-  LLint sizetot;
-  int offset;
-  //
-  int back;
-  //
-  int actived;                  // pour disabled
-};
+#include "htsstats.h"
 
 #ifndef HTS_DEF_FWSTRUCT_t_InpInfo
 #define HTS_DEF_FWSTRUCT_t_InpInfo

src/httrack.vcxproj

@@ -100,6 +100,7 @@
 
   <ItemGroup>
     <ClCompile Include="httrack.c" />
+    <ClCompile Include="htsbacktrace.c" />
   </ItemGroup>
 
   <!-- Pulls in libhttrack.lib and builds it first. -->

src/libhttrack.vcxproj

@@ -122,6 +122,7 @@
     <ClCompile Include="htshelp.c" />
     <ClCompile Include="htsindex.c" />
     <ClCompile Include="htslib.c" />
+    <ClCompile Include="htscmdline.c" />
     <ClCompile Include="htsurlport.c" />
     <ClCompile Include="htsmd5.c" />
     <ClCompile Include="htsmodules.c" />
@@ -130,6 +131,7 @@
     <ClCompile Include="htsproxy.c" />
     <ClCompile Include="htsrobots.c" />
     <ClCompile Include="htsselftest.c" />
+    <ClCompile Include="htssinglefile.c" />
     <ClCompile Include="htssniff.c" />
     <ClCompile Include="htsthread.c" />
     <ClCompile Include="htstools.c" />
@@ -137,6 +139,7 @@
     <ClCompile Include="htswrap.c" />
     <ClCompile Include="htszlib.c" />
     <ClCompile Include="htswarc.c" />
+    <ClCompile Include="htschanges.c" />
     <ClCompile Include="md5.c" />
     <ClCompile Include="minizip\ioapi.c" />
     <ClCompile Include="minizip\iowin32.c" />

src/proxy/proxytrack.c

@@ -688,7 +688,7 @@ static PT_Element proxytrack_process_DAV_Request(PT_Indexes indexes,
     PT_Element elt = PT_ElementNew();
 
     elt->statuscode = 405;
-    strcpy(elt->msg, "Method Not Allowed");
+    strcpybuff(elt->msg, "Method Not Allowed");
     return elt;
   }
 
@@ -820,9 +820,9 @@ static PT_Element proxytrack_process_DAV_Request(PT_Indexes indexes,
     elt->size = StringLength(response);
     elt->adr = StringAcquire(&response);
     elt->statuscode = 207;      /* Multi-Status */
-    strcpy(elt->charset, "utf-8");
-    strcpy(elt->contenttype, "text/xml");
-    strcpy(elt->msg, "Multi-Status");
+    strcpybuff(elt->charset, "utf-8");
+    strcpybuff(elt->contenttype, "text/xml");
+    strcpybuff(elt->msg, "Multi-Status");
     StringFree(response);
 
     fprintf(stderr, "RESPONSE:\n%s\n", elt->adr);
@@ -877,9 +877,9 @@ static PT_Element proxytrack_process_HTTP_List(PT_Indexes indexes,
     elt->size = StringLength(html);
     elt->adr = StringAcquire(&html);
     elt->statuscode = HTTP_OK;
-    strcpy(elt->charset, "iso-8859-1");
-    strcpy(elt->contenttype, "text/html");
-    strcpy(elt->msg, "OK");
+    strcpybuff(elt->charset, "iso-8859-1");
+    strcpybuff(elt->contenttype, "text/html");
+    strcpybuff(elt->msg, "OK");
     StringFree(html);
     return elt;
   }

src/proxy/proxytrack.h

@@ -368,7 +368,11 @@ HTS_UNUSED static struct tm PT_GetTime(time_t t) {
   if (tm != NULL)
     return *tm;
   else {
+    /* an all-zero tm has tm_mday == 0, which the ARC date field prints as a
+       day of "00"; the epoch is the conventional "date unknown" */
     memset(&tmbuf, 0, sizeof(tmbuf));
+    tmbuf.tm_year = 70;
+    tmbuf.tm_mday = 1;
     return tmbuf;
   }
 }

src/proxy/store.c

@@ -405,9 +405,9 @@ PT_Element PT_Index_HTML_BuildRootInfo(PT_Indexes indexes) {
     elt->size = StringLength(html);
     elt->adr = StringAcquire(&html);
     elt->statuscode = HTTP_OK;
-    strcpy(elt->charset, "iso-8859-1");
-    strcpy(elt->contenttype, "text/html");
-    strcpy(elt->msg, "OK");
+    strcpybuff(elt->charset, "iso-8859-1");
+    strcpybuff(elt->contenttype, "text/html");
+    strcpybuff(elt->msg, "OK");
     StringFree(html);
     return elt;
   }
@@ -579,9 +579,10 @@ PT_Index PT_LoadCache(const char *filename) {
         if (chain != NULL) {
           const char *scheme = link_has_authority(chain->name) ? "" : "http://";
 
-          snprintf(index->slots.common.startUrl,
-                   sizeof(index->slots.common.startUrl), "%s%s", scheme,
-                   (const char *) chain->name);
+          /* dropped rather than truncated: empty already reads as "unset" */
+          if (!sprintfbuff(index->slots.common.startUrl, "%s%s", scheme,
+                           (const char *) chain->name))
+            index->slots.common.startUrl[0] = '\0';
         }
       }
     }
@@ -827,6 +828,10 @@ PT_Element PT_ElementNew(void) {
   return r;
 }
 
+/* ProxyTrack's htsblk_failf(): a clipped, diagnostic-only failure reason. */
+#define PT_Element_failf(R, ...)                                               \
+  slprintfbuff_clip((R)->msg, sizeof((R)->msg), __VA_ARGS__)
+
 PT_Element PT_ReadCache(PT_Index index, const char *url, int flags) {
   if (index != NULL && SAFE_INDEX(index)) {
     return _IndexFuncts[index->type].PT_ReadCache(index, url, flags);
@@ -865,12 +870,15 @@ static PT_Element PT_ReadCache__New(PT_Index index, const char *url, int flags)
   sprintf(headers + headersSize, "%s: "LLintP"\r\n", field, (LLint)(value)); \
   (headersSize) += (int) strlen(headers + headersSize); \
 } while(0)
-#define ZIP_READFIELD_STRING(line, value, refline, refvalue) do { \
-  if (line[0] != '\0' && strfield2(line, refline)) { \
-    strcpy(refvalue, value); \
-    line[0] = '\0'; \
-	} \
-} while(0)
+/* refvalue_size is mandatory: the cache line is bounded only by the line
+   buffer, not by the destination. */
+#define ZIP_READFIELD_STRING(line, value, refline, refvalue, refvalue_size)    \
+  do {                                                                         \
+    if (line[0] != '\0' && strfield2(line, refline)) {                         \
+      (void) strclipbuff(refvalue, refvalue_size, value);                      \
+      line[0] = '\0';                                                          \
+    }                                                                          \
+  } while (0)
 #define ZIP_READFIELD_INT(line, value, refline, refvalue) do { \
   if (line[0] != '\0' && strfield2(line, refline)) { \
     int intval = 0; \
@@ -972,9 +980,12 @@ int PT_LoadCache__New(PT_Index index_, const char *filename) {
                     const char *scheme =
                         link_has_authority(filenameIndex) ? "" : "http://";
 
-                    firstSeen = 1;
-                    snprintf(index->startUrl, sizeof(index->startUrl), "%s%s",
-                             scheme, filenameIndex);
+                    /* dropped rather than truncated; try the next entry */
+                    if (sprintfbuff(index->startUrl, "%s%s", scheme,
+                                    filenameIndex))
+                      firstSeen = 1;
+                    else
+                      index->startUrl[0] = '\0';
                   }
                 }
               } else {
@@ -1021,7 +1032,7 @@ static PT_Element PT_ReadCache__New_u(PT_Index index_, const char *url,
   previous_save[0] = previous_save_[0] = '\0';
   memset(r, 0, sizeof(_PT_Element));
   r->location = location_default;
-  strcpy(r->location, "");
+  r->location[0] = '\0';
   if (strncmp(url, "http://", 7) == 0)
     url += 7;
   hash_pos_return = coucal_read(index->hash, url, &hash_pos);
@@ -1070,16 +1081,23 @@ static PT_Element PT_ReadCache__New_u(PT_Index index_, const char *url,
                 value++;
               ZIP_READFIELD_INT(line, value, "X-In-Cache", dataincache);
               ZIP_READFIELD_INT(line, value, "X-Statuscode", r->statuscode);
-              ZIP_READFIELD_STRING(line, value, "X-StatusMessage", r->msg);     // msg
+              ZIP_READFIELD_STRING(line, value, "X-StatusMessage", r->msg,
+                                   sizeof(r->msg));
               ZIP_READFIELD_INT(line, value, "X-Size", r->size);        // size
-              ZIP_READFIELD_STRING(line, value, "Content-Type", r->contenttype);        // contenttype
-              ZIP_READFIELD_STRING(line, value, "X-Charset", r->charset);       // contenttype
-              ZIP_READFIELD_STRING(line, value, "Last-Modified", r->lastmodified);      // last-modified
-              ZIP_READFIELD_STRING(line, value, "Etag", r->etag);       // Etag
-              ZIP_READFIELD_STRING(line, value, "Location", r->location);       // 'location' pour moved
+              ZIP_READFIELD_STRING(line, value, "Content-Type", r->contenttype,
+                                   sizeof(r->contenttype));
+              ZIP_READFIELD_STRING(line, value, "X-Charset", r->charset,
+                                   sizeof(r->charset));
+              ZIP_READFIELD_STRING(line, value, "Last-Modified",
+                                   r->lastmodified, sizeof(r->lastmodified));
+              ZIP_READFIELD_STRING(line, value, "Etag", r->etag,
+                                   sizeof(r->etag));
+              ZIP_READFIELD_STRING(line, value, "Location", r->location,
+                                   sizeof(location_default));
               ZIP_READFIELD_STRING(line, value, "Content-Disposition",
-                                   r->cdispo); // Content-disposition
-              ZIP_READFIELD_STRING(line, value, "X-Save", previous_save_);      // Original save filename
+                                   r->cdispo, sizeof(r->cdispo));
+              ZIP_READFIELD_STRING(line, value, "X-Save", previous_save_,
+                                   sizeof(previous_save_));
               if (line[0] != '\0') {
                 int len = r->headers ? ((int) strlen(r->headers)) : 0;
                 int nlen =
@@ -1099,7 +1117,7 @@ static PT_Element PT_ReadCache__New_u(PT_Index index_, const char *url,
             int pathLen = (int) strlen(index->path);
 
             if (pathLen > 0 && strncmp(previous_save_, index->path, pathLen) == 0) {    // old (<3.40) buggy format
-              strcpy(previous_save, previous_save_);
+              strcpybuff(previous_save, previous_save_);
             }
             // relative ? (hack)
             else if (index->safeCache || (previous_save_[0] != '/'      // /home/foo/bar.gif
@@ -1135,8 +1153,9 @@ static PT_Element PT_ReadCache__New_u(PT_Index index_, const char *url,
                   snprintf(previous_save, sizeof(previous_save), "%s%s",
                            index->path, previous_save_ + index->fixedPath);
                 } else {
-                  snprintf(r->msg, sizeof(r->msg), "Bogus fixePath prefix for %s (prefixLen=%d)",
-                          previous_save_, (int) index->fixedPath);
+                  PT_Element_failf(
+                      r, "Bogus fixePath prefix for %s (prefixLen=%d)",
+                      previous_save_, (int) index->fixedPath);
                   r->statuscode = STATUSCODE_INVALID;
                 }
               } else {
@@ -1155,7 +1174,7 @@ static PT_Element PT_ReadCache__New_u(PT_Index index_, const char *url,
             // Peut-on stocker le fichier directement sur disque?
             if (ok) {
               if (r->msg[0] == '\0') {
-                strcpy(r->msg, "Cache Read Error : Unexpected error");
+                strcpybuff(r->msg, "Cache Read Error : Unexpected error");
               }
             } else {            // lire en mémoire
 
@@ -1173,24 +1192,27 @@ static PT_Element PT_ReadCache__New_u(PT_Index index_, const char *url,
                           int last_errno = errno;
 
                           r->statuscode = STATUSCODE_INVALID;
-                          sprintf(r->msg, "Read error in cache disk data: %s",
-                                  strerror(last_errno));
+                          PT_Element_failf(r,
+                                           "Read error in cache disk data: %s",
+                                           strerror(last_errno));
                         }
                         r->adr[r->size] = '\0';
                       } else {
                         r->statuscode = STATUSCODE_INVALID;
-                        strcpy(r->msg,
-                               "Read error (memory exhausted) from cache");
+                        strcpybuff(r->msg,
+                                   "Read error (memory exhausted) from cache");
                       }
                       fclose(fp);
                     } else {
                       r->statuscode = STATUSCODE_INVALID;
-                      snprintf(r->msg, sizeof(r->msg), "Read error (can't open '%s') from cache",
-                              file_convert(catbuff, sizeof(catbuff), previous_save));
+                      PT_Element_failf(
+                          r, "Read error (can't open '%s') from cache",
+                          file_convert(catbuff, sizeof(catbuff),
+                                       previous_save));
                     }
                   } else {
                     r->statuscode = STATUSCODE_INVALID;
-                    strcpy(r->msg, "Cached file name is invalid");
+                    strcpybuff(r->msg, "Cached file name is invalid");
                   }
                 }
               } else {
@@ -1202,12 +1224,12 @@ static PT_Element PT_ReadCache__New_u(PT_Index index_, const char *url,
                       free(r->adr);
                       r->adr = NULL;
                       r->statuscode = STATUSCODE_INVALID;
-                      strcpy(r->msg, "Cache Read Error : Read Data");
+                      strcpybuff(r->msg, "Cache Read Error : Read Data");
                     } else
                       *(r->adr + r->size) = '\0';
                   } else {      // erreur
                     r->statuscode = STATUSCODE_INVALID;
-                    strcpy(r->msg, "Cache Memory Error");
+                    strcpybuff(r->msg, "Cache Memory Error");
                   }
                 }
               }
@@ -1215,21 +1237,21 @@ static PT_Element PT_ReadCache__New_u(PT_Index index_, const char *url,
           }                     // si save==null, ne rien charger (juste en tête)
         } else {
           r->statuscode = STATUSCODE_INVALID;
-          strcpy(r->msg, "Cache Read Error : Read Header Data");
+          strcpybuff(r->msg, "Cache Read Error : Read Header Data");
         }
         unzCloseCurrentFile(index->zFile);
       } else {
         r->statuscode = STATUSCODE_INVALID;
-        strcpy(r->msg, "Cache Read Error : Open File");
+        strcpybuff(r->msg, "Cache Read Error : Open File");
       }
 
     } else {
       r->statuscode = STATUSCODE_INVALID;
-      strcpy(r->msg, "Cache Read Error : Bad Offset");
+      strcpybuff(r->msg, "Cache Read Error : Bad Offset");
     }
   } else {
     r->statuscode = STATUSCODE_INVALID;
-    strcpy(r->msg, "File Cache Entry Not Found");
+    strcpybuff(r->msg, "File Cache Entry Not Found");
   }
   if (r->location[0] != '\0') {
     r->location = strdup(r->location);
@@ -1522,7 +1544,7 @@ static int PT_LoadCache__Old(PT_Index index_, const char *filename) {
             cache->version = (int) (firstline[8] - '0');        // cache 1.x
             if (cache->version <= 5) {
               a += cache_brstr(a, firstline, sizeof(firstline));
-              strcpy(cache->lastmodified, firstline);
+              strcpybuff(cache->lastmodified, firstline);
             } else {
               fclose(cache->dat);
               cache->dat = NULL;
@@ -1539,7 +1561,7 @@ static int PT_LoadCache__Old(PT_Index index_, const char *filename) {
         } else {                // Vieille version du cache
           /* */
           cache->version = 0;   // cache 1.0
-          strcpy(cache->lastmodified, firstline);
+          strcpybuff(cache->lastmodified, firstline);
         }
 
         /* Create hash table for the cache (MUCH FASTER!) */
@@ -1555,7 +1577,14 @@ static int PT_LoadCache__Old(PT_Index index_, const char *filename) {
               a++;
               /* read "host/file" */
               a += binput(a, line, HTS_URLMAXSIZE);
-              a += binput(a, line + strlen(line), HTS_URLMAXSIZE);
+              {
+                /* binput writes its NUL at s[max], so the second field must be
+                   bounded by what is left of line[], not by the same constant
+                 */
+                const size_t used = strlen(line);
+
+                a += binput(a, line + used, (int) (sizeof(line) - used - 1));
+              }
               /* read position */
               a += binput(a, linepos, 200);
               sscanf(linepos, "%d", &pos);
@@ -1570,9 +1599,11 @@ static int PT_LoadCache__Old(PT_Index index_, const char *filename) {
                   const char *scheme =
                       link_has_authority(line) ? "" : "http://";
 
-                  firstSeen = 1;
-                  snprintf(index->startUrl, sizeof(index->startUrl), "%s%s",
-                           scheme, line);
+                  /* dropped rather than truncated; try the next entry */
+                  if (sprintfbuff(index->startUrl, "%s%s", scheme, line))
+                    firstSeen = 1;
+                  else
+                    index->startUrl[0] = '\0';
                 }
               }
 
@@ -1650,7 +1681,7 @@ static PT_Element PT_ReadCache__Old_u(PT_Index index_, const char *url,
   previous_save[0] = previous_save_[0] = '\0';
   memset(r, 0, sizeof(_PT_Element));
   r->location = location_default;
-  strcpy(r->location, "");
+  r->location[0] = '\0';
   if (strncmp(url, "http://", 7) == 0)
     url += 7;
   hash_pos_return = coucal_read(cache->hash, url, &hash_pos);
@@ -1757,7 +1788,7 @@ static PT_Element PT_ReadCache__Old_u(PT_Index index_, const char *url,
           int pathLen = (int) strlen(index->path);
 
           if (pathLen > 0 && strncmp(previous_save_, index->path, pathLen) == 0) {      // old (<3.40) buggy format
-            strcpy(previous_save, previous_save_);
+            strcpybuff(previous_save, previous_save_);
           }
           // relative ? (hack)
           else if (index->safeCache || (previous_save_[0] != '/'        // /home/foo/bar.gif
@@ -1793,8 +1824,9 @@ static PT_Element PT_ReadCache__Old_u(PT_Index index_, const char *url,
                 snprintf(previous_save, sizeof(previous_save), "%s%s",
                          index->path, previous_save_ + index->fixedPath);
               } else {
-                snprintf(r->msg, sizeof(r->msg), "Bogus fixePath prefix for %s (prefixLen=%d)",
-                        previous_save_, (int) index->fixedPath);
+                PT_Element_failf(r,
+                                 "Bogus fixePath prefix for %s (prefixLen=%d)",
+                                 previous_save_, (int) index->fixedPath);
                 r->statuscode = STATUSCODE_INVALID;
               }
             } else {
@@ -1820,17 +1852,18 @@ static PT_Element PT_ReadCache__Old_u(PT_Index index_, const char *url,
                 if (r->adr != NULL) {
                   if (r->size > 0 && fread(r->adr, 1, r->size, fp) != r->size) {
                     r->statuscode = STATUSCODE_INVALID;
-                    strcpy(r->msg, "Read error in cache disk data");
+                    strcpybuff(r->msg, "Read error in cache disk data");
                   }
                   r->adr[r->size] = '\0';
                 } else {
                   r->statuscode = STATUSCODE_INVALID;
-                  strcpy(r->msg, "Read error (memory exhausted) from cache");
+                  strcpybuff(r->msg,
+                             "Read error (memory exhausted) from cache");
                 }
                 fclose(fp);
               } else {
                 r->statuscode = STATUSCODE_INVALID;
-                strcpy(r->msg, "Previous cache file not found (2)");
+                strcpybuff(r->msg, "Previous cache file not found (2)");
               }
             }
           } else {
@@ -1842,30 +1875,30 @@ static PT_Element PT_ReadCache__Old_u(PT_Index index_, const char *url,
                   free(r->adr);
                   r->adr = NULL;
                   r->statuscode = STATUSCODE_INVALID;
-                  strcpy(r->msg, "Cache Read Error : Read Data");
+                  strcpybuff(r->msg, "Cache Read Error : Read Data");
                 } else
                   r->adr[r->size] = '\0';
               } else {          // erreur
                 r->statuscode = STATUSCODE_INVALID;
-                strcpy(r->msg, "Cache Memory Error");
+                strcpybuff(r->msg, "Cache Memory Error");
               }
             }
           }
         } else {
           r->statuscode = STATUSCODE_INVALID;
-          strcpy(r->msg, "Cache Read Error : Bad Data");
+          strcpybuff(r->msg, "Cache Read Error : Bad Data");
         }
       } else {                  // erreur
         r->statuscode = STATUSCODE_INVALID;
-        strcpy(r->msg, "Cache Read Error : Read Header");
+        strcpybuff(r->msg, "Cache Read Error : Read Header");
       }
     } else {
       r->statuscode = STATUSCODE_INVALID;
-      strcpy(r->msg, "Cache Read Error : Seek Failed");
+      strcpybuff(r->msg, "Cache Read Error : Seek Failed");
     }
   } else {
     r->statuscode = STATUSCODE_INVALID;
-    strcpy(r->msg, "File Cache Entry Not Found");
+    strcpybuff(r->msg, "File Cache Entry Not Found");
   }
   if (r->location[0] != '\0') {
     r->location = strdup(r->location);
@@ -2128,12 +2161,14 @@ int PT_LoadCache__Arc(PT_Index index_, const char *filename) {
   return 0;
 }
 
-#define HTTP_READFIELD_STRING(line, value, refline, refvalue) do { \
-  if (line[0] != '\0' && strfield2(line, refline)) { \
-    strcpy(refvalue, value); \
-    line[0] = '\0'; \
-	} \
-} while(0)
+/* Same contract as ZIP_READFIELD_STRING, for the ARC reader's header lines. */
+#define HTTP_READFIELD_STRING(line, value, refline, refvalue, refvalue_size)   \
+  do {                                                                         \
+    if (line[0] != '\0' && strfield2(line, refline)) {                         \
+      (void) strclipbuff(refvalue, refvalue_size, value);                      \
+      line[0] = '\0';                                                          \
+    }                                                                          \
+  } while (0)
 #define HTTP_READFIELD_INT(line, value, refline, refvalue) do { \
   if (line[0] != '\0' && strfield2(line, refline)) { \
     int intval = 0; \
@@ -2169,7 +2204,7 @@ static PT_Element PT_ReadCache__Arc_u(PT_Index index_, const char *url,
   location_default[0] = '\0';
   memset(r, 0, sizeof(_PT_Element));
   r->location = location_default;
-  strcpy(r->location, "");
+  r->location[0] = '\0';
   if (strncmp(url, "http://", 7) == 0)
     url += 7;
   hash_pos_return = coucal_read(index->hash, url, &hash_pos);
@@ -2202,11 +2237,16 @@ static PT_Element PT_ReadCache__Arc_u(PT_Index index_, const char *url,
               *value = '\0';
               for(value++; *value == ' ' || *value == '\t'; value++) ;
               HTTP_READFIELD_INT(line, value, "Content-Length", r->size);       // size
-              HTTP_READFIELD_STRING(line, value, "Content-Type", r->contenttype);       // contenttype
-              HTTP_READFIELD_STRING(line, value, "Last-Modified", r->lastmodified);     // last-modified
-              HTTP_READFIELD_STRING(line, value, "Etag", r->etag);      // Etag
-              HTTP_READFIELD_STRING(line, value, "Location", r->location);      // 'location' pour moved
-              HTTP_READFIELD_STRING(line, value, "Content-Disposition", r->cdispo);     // Content-disposition
+              HTTP_READFIELD_STRING(line, value, "Content-Type", r->contenttype,
+                                    sizeof(r->contenttype));
+              HTTP_READFIELD_STRING(line, value, "Last-Modified",
+                                    r->lastmodified, sizeof(r->lastmodified));
+              HTTP_READFIELD_STRING(line, value, "Etag", r->etag,
+                                    sizeof(r->etag));
+              HTTP_READFIELD_STRING(line, value, "Location", r->location,
+                                    sizeof(location_default));
+              HTTP_READFIELD_STRING(line, value, "Content-Disposition",
+                                    r->cdispo, sizeof(r->cdispo));
               if (line[0] != '\0') {
                 int len = r->headers ? ((int) strlen(r->headers)) : 0;
                 int nlen =
@@ -2246,7 +2286,7 @@ static PT_Element PT_ReadCache__Arc_u(PT_Index index_, const char *url,
                 fetchSize = dataLength - metaSize;
               } else if (fetchSize > dataLength - metaSize) {
                 r->statuscode = STATUSCODE_INVALID;
-                strcpy(r->msg, "Cache Read Error : Truncated Data");
+                strcpybuff(r->msg, "Cache Read Error : Truncated Data");
               }
               r->size = 0;
               if (r->statuscode != STATUSCODE_INVALID) {
@@ -2259,12 +2299,13 @@ static PT_Element PT_ReadCache__Arc_u(PT_Index index_, const char *url,
                     int last_errno = errno;
 
                     r->statuscode = STATUSCODE_INVALID;
-                    sprintf(r->msg, "Read error in cache disk data: %s",
-                            strerror(last_errno));
+                    PT_Element_failf(r, "Read error in cache disk data: %s",
+                                     strerror(last_errno));
                   }
                 } else {
                   r->statuscode = STATUSCODE_INVALID;
-                  strcpy(r->msg, "Read error (memory exhausted) from cache");
+                  strcpybuff(r->msg,
+                             "Read error (memory exhausted) from cache");
                 }
               }
             }
@@ -2272,21 +2313,21 @@ static PT_Element PT_ReadCache__Arc_u(PT_Index index_, const char *url,
 
         } else {
           r->statuscode = STATUSCODE_INVALID;
-          strcpy(r->msg, "Cache Read Error : Read Header Error");
+          strcpybuff(r->msg, "Cache Read Error : Read Header Error");
         }
 
       } else {
         r->statuscode = STATUSCODE_INVALID;
-        strcpy(r->msg, "Cache Read Error : Read Header Error");
+        strcpybuff(r->msg, "Cache Read Error : Read Header Error");
       }
     } else {
       r->statuscode = STATUSCODE_INVALID;
-      strcpy(r->msg, "Cache Read Error : Seek Error");
+      strcpybuff(r->msg, "Cache Read Error : Seek Error");
     }
 
   } else {
     r->statuscode = STATUSCODE_INVALID;
-    strcpy(r->msg, "File Cache Entry Not Found");
+    strcpybuff(r->msg, "File Cache Entry Not Found");
   }
   if (r->location[0] != '\0') {
     r->location = strdup(r->location);
@@ -2335,8 +2376,18 @@ static int PT_SaveCache__Arc_Fun(void *arg, const char *url, PT_Element element)
   PT_SaveCache__Arc_t *st = (PT_SaveCache__Arc_t *) arg;
   FILE *const fp = st->fp;
   struct tm *tm = convert_time_rfc822(&st->buff, element->lastmodified);
+  struct tm unknown_date;
   int size_headers;
 
+  /* a cached entry with no parseable Last-Modified must not take the writer
+     down; the epoch is the conventional "date unknown" */
+  if (tm == NULL) {
+    memset(&unknown_date, 0, sizeof(unknown_date));
+    unknown_date.tm_year = 70;
+    unknown_date.tm_mday = 1;
+    tm = &unknown_date;
+  }
+
   sprintf(st->headers,
           "HTTP/1.0 %d %s"
           "\r\n"
@@ -2372,7 +2423,7 @@ static int PT_SaveCache__Arc_Fun(void *arg, const char *url, PT_Element element)
   if (element->adr != NULL) {
     domd5mem(element->adr, element->size, st->md5, 1);
   } else {
-    strcpy(st->md5, "-");
+    strcpybuff(st->md5, "-");
   }
   fprintf(fp,
           /* nl */