Advertise deflate in Accept-Encoding and decode it

The request Accept-Encoding offered only gzip even though the response
parser already recognized deflate/x-deflate. But the actual decode path
(hts_zunpack) used zlib's gzread, which only inflates gzip and copies any
deflate body through verbatim, so a deflate response would have been
written out still compressed. Advertising deflate without fixing that
would corrupt files.

Rewrite hts_zunpack to inflate via inflateInit2 with format detection:
gzip and zlib (RFC1950) auto-detect with +32 windowBits, everything else
is treated as raw deflate (RFC1951). Then add deflate to the advertised
list through a small hts_acceptencoding() helper shared with the test.

A new -#test=acceptencoding self-test asserts the advertised header
carries both gzip and deflate, and round-trips gzip, zlib and raw-deflate
bodies through hts_zunpack on disk. Both halves fail on the old binary.

Brotli is intentionally out of scope (new dependency, larger change).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: Xavier Roche <roche@httrack.com>

.github/workflows/ci.yml

@@ -61,6 +61,50 @@ jobs:
         if: failure()
         run: cat tests/test-suite.log 2>/dev/null || true
 
+  # Reproduce the Debian buildds: they build in a minimal chroot with no
+  # python3, so the local-server tests must SKIP (exit 77), not fail. GitHub
+  # runners ship python3, so every other job hides this path; here we remove it
+  # before `make check`. This is the guard that would have caught the 3.49.10-1
+  # FTBFS (28_local-pause failed instead of skipping when python3 was absent).
+  buildd-no-python3:
+    name: build (no python3, Debian buildd)
+    runs-on: ubuntu-24.04
+    steps:
+      - uses: actions/checkout@v6
+        with:
+          submodules: recursive
+
+      - name: Install build dependencies
+        run: |
+          set -euo pipefail
+          sudo apt-get update
+          sudo apt-get install -y --no-install-recommends \
+            build-essential autoconf automake libtool autoconf-archive \
+            zlib1g-dev libssl-dev
+
+      - name: Configure
+        run: |
+          set -euo pipefail
+          autoreconf -fi
+          ./configure
+
+      - name: Build
+        run: make -j"$(nproc)"
+
+      - name: Test without python3
+        run: |
+          set -euo pipefail
+          # Hide every python3* so `command -v python3` fails like it does in the
+          # buildd chroot; masking with /bin/false would still resolve.
+          sudo find /usr/bin /usr/local/bin -maxdepth 1 -name 'python3*' \
+            -exec mv {} {}.hidden \;
+          ! command -v python3
+          make check
+
+      - name: Print the test log on failure
+        if: failure()
+        run: cat tests/test-suite.log 2>/dev/null || true
+
   # Portability: build and test on macOS (Darwin/clang) on a native runner --
   # no VM. The tree has no __APPLE__ branches, so Darwin exercises the
   # generic-Unix path on a second libc and kernel. brew's openssl@3 is keg-only,
@@ -225,8 +269,9 @@ jobs:
           MSAN_OPTIONS: abort_on_error=1:halt_on_error=1
         run: |
           set -euo pipefail
-          # Engine self-tests only; the cache trio pulls in uninstrumented zlib.
-          tests="$(cd tests && ls 01_engine-*.test | grep -v -- '-cache' | tr '\n' ' ')"
+          # 01_engine-* only; zlib-dependent self-tests are named 01_zlib-* and
+          # skipped here (uninstrumented libz floods MSan with false positives).
+          tests="$(cd tests && ls 01_engine-*.test | tr '\n' ' ')"
           make check TESTS="$tests"
 
       - name: Print the test log on failure

CONTRIBUTING.md

@@ -39,6 +39,10 @@ Welcome, and nothing to disclose. Two rules:
 
 The sign-off covers AI-assisted code too.
 
+## Translations
+
+Interface strings live in [`lang/`](lang/). See [lang/README.md](lang/README.md) for the file format and how to add or update a language.
+
 ## Bugs
 
 Open an issue with the version, OS, command used, and expected vs actual result.

configure.ac

@@ -1,6 +1,6 @@
 AC_PREREQ([2.71])
 
-AC_INIT([httrack], [3.49.9], [roche+packaging@httrack.com], [httrack], [http://www.httrack.com/])
+AC_INIT([httrack], [3.49.10], [roche+packaging@httrack.com], [httrack], [http://www.httrack.com/])
 AC_COPYRIGHT([
 HTTrack Website Copier, Offline Browser for Windows and Unix
 Copyright (C) 1998-2015 Xavier Roche and other contributors
@@ -29,10 +29,10 @@ AC_CONFIG_SRCDIR(src/httrack.c)
 AC_CONFIG_MACRO_DIR([m4])
 AC_CONFIG_HEADERS(config.h)
 AM_INIT_AUTOMAKE([subdir-objects])
-# 3:1:0: 3.49.9 changed code but not the exported interface vs 3.49.8 (same 164
-# symbols, no struct-layout change), so bump revision only. (3:0:0 was the htsblk
-# mime-buffer widening, an ABI break that moved the soname .so.2 -> .so.3.)
-VERSION_INFO="3:1:0"
+# 3:2:0: 3.49.10 only appends tail fields to the options struct (no existing
+# symbol or offset changed vs 3.49.9), so it stays soname .so.3; bump revision.
+# (3:0:0 was the htsblk mime-buffer widening, the ABI break that moved .so.2 -> .so.3.)
+VERSION_INFO="3:2:0"
 AM_MAINTAINER_MODE
 AC_USE_SYSTEM_EXTENSIONS
 

debian/changelog

@@ -1,3 +1,16 @@
+httrack (3.49.10-1) unstable; urgency=medium
+
+  * New upstream release: new download-pacing and URL-handling options plus a
+    batch of crawl and robustness fixes (full list in history.txt).
+  * Rewrite debian/copyright in machine-readable DEP-5 format, crediting the
+    bundled minizip, md5 and coucal sources (#415).
+  * Lead the webhttrack browser dependency with chromium so httrack is not
+    dragged into the firefox-esr autoremoval cascade (#436).
+  * Override the embedded-library lint for the bundled minizip (#419).
+  * Bump Standards-Version to 4.7.4 (no changes required).
+
+ -- Xavier Roche <xavier@debian.org>  Sun, 28 Jun 2026 14:01:53 +0200
+
 httrack (3.49.9-1) unstable; urgency=medium
 
   * New upstream release: Content-Type and file-type detection fixes (trust a

debian/control

@@ -2,7 +2,7 @@ Source: httrack
 Section: web
 Priority: optional
 Maintainer: Xavier Roche <roche@httrack.com>
-Standards-Version: 4.7.0
+Standards-Version: 4.7.4
 Build-Depends: debhelper-compat (= 13), autoconf, autoconf-archive, automake, libtool, zlib1g-dev, libssl-dev
 Rules-Requires-Root: no
 Homepage: http://www.httrack.com

history.txt

@@ -4,7 +4,25 @@ HTTrack Website Copier release history:
 
 This file lists all changes and fixes that have been made for HTTrack
 
-3.49-9
+3.49-10
++ New: --cookies-file to preload a Netscape cookies.txt before crawling (#215)
++ New: --pause to space out file downloads by a random delay (#185)
++ New: --strip-query to drop selected query keys from the dedup naming (#112)
++ Changed: split the -%u URL hacks into independent --keep-www-prefix, --keep-double-slashes and --keep-query-order toggles (#271)
++ Fixed: follow a redirect Location after dropping its #fragment, instead of requesting the fragment and polluting the saved name (#204)
++ Fixed: escaped brackets inside a *[...] filter character class (#148)
++ Fixed: honor the server's Content-Range when resuming a partial download, instead of appending overlapping bytes (#198)
++ Fixed: abort the download as soon as the response type is excluded by -mime:, instead of fetching then discarding the body (#58)
++ Fixed: keep size-based filter rules neutral until the file size is known (#143)
++ Fixed: stop the mirror with a clean fatal error on a cache write failure, instead of crashing (#174, #219)
++ Fixed: stop the 412/416 partial re-get loop on --continue and --update (#206)
++ Fixed: keep an unrecognized URL tail instead of mangling it to .html (#115)
++ Fixed: honor --tolerant (-%B) on a broken Content-Length, and fix an out-of-bounds read it exposed (#32, #41)
++ Fixed: fall back to the next resolved address when a connection fails or stalls, instead of hanging on a dead IPv6 address
++ Fixed: report why a -%L URL list could not be loaded (#49)
++ Changed: multiple internal hardening, build and CI improvements
+
+.49-9
 + Fixed: file-type detection from the Content-Type header: trust a declared type over a binary URL extension, honor --assume under the delayed type check, and keep a known extension against a bogus or empty Content-Type (#267, #29, #56)
 + Fixed: an uninitialized-buffer read when the Content-Type is empty (#411)
 + Fixed: restored C++ source-compatibility of the installed headers so reverse dependencies (httraqt) build again (#413)

html/filters.html

@@ -247,7 +247,7 @@ See also: The <a href="faq.html#VF1">FAQ</a><br>
         <td>the \ character</td>
       </tr>
       <tr>
-        <td nowrap><tt>*[\[\]]</tt></td>
+        <td nowrap><tt>*[\[,\]]</tt></td>
         <td>the [ or ] character</td>
       </tr>
       <tr>

lang/English.txt

@@ -295,7 +295,7 @@ Max Depth
 Maximum external depth:
 Maximum external depth:
 Filters (refuse/accept links) :
-Filters (refuse/accept links) :
+Filters (refuse/accept links):
 Paths
 Paths
 Save prefs

lang/README.md

@@ -0,0 +1,37 @@
+# Translating HTTrack
+
+Interface strings live here, one `.txt` file per language. `English.txt` is the reference: every other file maps each English string to its translation.
+
+## File format
+
+Plain text, entries in consecutive pairs of lines:
+
+```
+<English string>
+<translation>
+```
+
+The first line of a pair is the lookup key and must stay identical to the one in `English.txt`; translate only the second line. Missing entries fall back to the English text at runtime, so a partial translation works.
+
+Preserve any `\r\n`, `\t` and `printf` placeholders (`%s`, `%d`, ...) in the translation.
+
+A few `LANGUAGE_*` entries at the top describe the file itself:
+
+| Key | Meaning |
+| --- | --- |
+| `LANGUAGE_NAME` | Name shown in the language picker, in its own language (`Deutsch`, not `German`) |
+| `LANGUAGE_ISO` | ISO 639 code, with region if needed (`de`, `pt_BR`) |
+| `LANGUAGE_CHARSET` | Encoding the file is saved in (`ISO-8859-1`, `windows-1251`, `UTF-8`, ...) |
+| `LANGUAGE_AUTHOR` | Your name and contact |
+| `LANGUAGE_WINDOWSID` | Windows locale name used by WinHTTrack (`German (Standard)`) |
+
+Save the file in exactly its declared `LANGUAGE_CHARSET`; an editor that rewrites it as UTF-8 will corrupt the non-ASCII bytes.
+
+## Adding or updating a language
+
+1. Copy `English.txt` to `<Language>.txt`, or edit the existing file.
+2. Translate each second line; leave the English keys untouched.
+3. Fill in the `LANGUAGE_*` header for a new file.
+4. Open a pull request, or attach the file to a GitHub issue.
+
+When new strings land in `English.txt` they show up untranslated (as English) until a translator fills them in.

man/httrack.1

@@ -24,6 +24,7 @@ httrack \- offline browser : copy websites to a local directory
 [ \fB\-EN, \-\-max\-time[=N]\fR ]
 [ \fB\-AN, \-\-max\-rate[=N]\fR ]
 [ \fB\-%cN, \-\-connection\-per\-second[=N]\fR ]
+[ \fB\-%G, \-\-pause\fR ]
 [ \fB\-GN, \-\-max\-pause[=N]\fR ]
 [ \fB\-cN, \-\-sockets[=N]\fR ]
 [ \fB\-TN, \-\-timeout[=N]\fR ]
@@ -155,6 +156,8 @@ maximum mirror time in seconds (60=1 minute, 3600=1 hour) (\-\-max\-time[=N])
 maximum transfer rate in bytes/seconds (1000=1KB/s max) (\-\-max\-rate[=N])
 .IP \-%cN
 maximum number of connections/seconds (*%c10) (\-\-connection\-per\-second[=N])
+.IP \-%G
+random pause of MIN[:MAX] seconds between files (e.g. %G5:10) (\-\-pause <param>)
 .IP \-GN
 pause transfer if N bytes reached, and wait until lock file is deleted (\-\-max\-pause[=N])
 .SS Flow control:

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"},
+  {"pause", "-%G", "param1",
+   "random pause of MIN[:MAX] seconds between files"},
   {"generate-errors", "-o", "single", ""},
   {"do-not-generate-errors", "-o0", "single", ""},
   {"purge-old", "-X", "param", ""},

src/htsbasenet.h

@@ -129,6 +129,8 @@ typedef enum HTTPStatusCode {
   HTTP_UNSUPPORTED_MEDIA_TYPE = 415,
   HTTP_REQUESTED_RANGE_NOT_SATISFIABLE = 416,
   HTTP_EXPECTATION_FAILED = 417,
+  HTTP_TOO_MANY_REQUESTS = 429,
+  HTTP_UNAVAILABLE_FOR_LEGAL_REASONS = 451,
   HTTP_INTERNAL_SERVER_ERROR = 500,
   HTTP_NOT_IMPLEMENTED = 501,
   HTTP_BAD_GATEWAY = 502,

src/htsbasiccharsets.sh

@@ -3,12 +3,12 @@
 
 # Change this to download files
 if false; then
-    echo "mget ftp://ftp.unicode.org/Public/MAPPINGS/ISO8859/8859-*.TXT" | lftp
-    echo "mget ftp://ftp.unicode.org/Public/MAPPINGS/VENDORS/MICSFT/PC/CP*.TXT" | lftp
-    echo "mget ftp://ftp.unicode.org/Public/MAPPINGS/VENDORS/MICSFT/WINDOWS/CP*.TXT" | lftp
-    echo "mget ftp://ftp.unicode.org/Public/MAPPINGS/VENDORS/MICSFT/EBCDIC/CP*.TXT" | lftp
-    echo "mget ftp://ftp.unicode.org/Public/MAPPINGS/VENDORS/MISC/CP*.TXT" | lftp
-    echo "mget ftp://ftp.unicode.org/Public/MAPPINGS/VENDORS/MISC/KOI8*.TXT" | lftp
+    echo "mget https://www.unicode.org/Public/MAPPINGS/ISO8859/8859-*.TXT" | lftp
+    echo "mget https://www.unicode.org/Public/MAPPINGS/VENDORS/MICSFT/PC/CP*.TXT" | lftp
+    echo "mget https://www.unicode.org/Public/MAPPINGS/VENDORS/MICSFT/WINDOWS/CP*.TXT" | lftp
+    echo "mget https://www.unicode.org/Public/MAPPINGS/VENDORS/MICSFT/EBCDIC/CP*.TXT" | lftp
+    echo "mget https://www.unicode.org/Public/MAPPINGS/VENDORS/MISC/CP*.TXT" | lftp
+    echo "mget https://www.unicode.org/Public/MAPPINGS/VENDORS/MISC/KOI8*.TXT" | lftp
     rm -f CP932.TXT CP936.TXT CP949.TXT CP950.TXT
 fi
 

src/htscatchurl.c

@@ -64,7 +64,7 @@ Please visit our Website: http://www.httrack.com
 // catch_url_init(&port,&return_host);
 HTSEXT_API T_SOC catch_url_init_std(int *port_prox, char *adr_prox) {
   T_SOC soc;
-  int try_to_listen_to[] = { 8080, 3128, 80, 81, 82, 8081, 3129, 31337, 0, -1 };
+  int try_to_listen_to[] = {8080, 3128, 80, 81, 82, 8081, 3129, 0, -1};
   int i = 0;
 
   do {

src/htscore.c

@@ -35,6 +35,7 @@ Please visit our Website: http://www.httrack.com
 
 #include <fcntl.h>
 #include <ctype.h>
+#include <stdint.h> /* uint64_t for the pause mixer (already a hard dep via md5.h) */
 
 /* File defs */
 #include "htscore.h"
@@ -3314,6 +3315,21 @@ HTS_INLINE int back_fillmax(struct_back * sback, httrackp * opt,
   return -1;                    /* plus de place */
 }
 
+/* Seed-derived: stable within a gap, rerolls per launch; a per-call rand()
+   would bias the delay toward min_ms (see header). Jitter, not crypto. */
+int hts_pause_target_ms(TStamp seed, int min_ms, int max_ms) {
+  uint64_t z = (uint64_t) seed;
+
+  if (max_ms <= min_ms)
+    return min_ms;
+  /* SplitMix64 finalizer: scrambles the low-entropy ms timestamp. */
+  z += 0x9E3779B97F4A7C15ULL;
+  z = (z ^ (z >> 30)) * 0xBF58476D1CE4E5B9ULL;
+  z = (z ^ (z >> 27)) * 0x94D049BB133111EBULL;
+  z ^= z >> 31;
+  return min_ms + (int) (z % (uint64_t) (max_ms - min_ms + 1));
+}
+
 int back_pluggable_sockets_strict(struct_back * sback, httrackp * opt) {
   int n = opt->maxsoc - back_nsoc(sback);
 
@@ -3334,6 +3350,18 @@ int back_pluggable_sockets_strict(struct_back * sback, httrackp * opt) {
     }
   }
 
+  // #185 randomized inter-file pause: non-blocking, one launch per gap
+  if (n > 0 && opt->pause_max_ms > 0 && HTS_STAT.last_connect > 0) {
+    TStamp opTime =
+        HTS_STAT.last_request ? HTS_STAT.last_request : HTS_STAT.last_connect;
+    TStamp lap = mtime_local() - opTime;
+
+    if (lap < hts_pause_target_ms(opTime, opt->pause_min_ms, opt->pause_max_ms))
+      n = 0;
+    else
+      n = 1;
+  }
+
   return n;
 }
 
@@ -3748,6 +3776,11 @@ HTSEXT_API int copy_htsopt(const httrackp * from, httrackp * to) {
   if (StringNotEmpty(from->cookies_file))
     StringCopyS(to->cookies_file, from->cookies_file);
 
+  if (from->pause_max_ms > 0) {
+    to->pause_min_ms = from->pause_min_ms;
+    to->pause_max_ms = from->pause_max_ms;
+  }
+
   if (from->retry > -1)
     to->retry = from->retry;
 

src/htscore.h

@@ -418,6 +418,10 @@ int back_pluggable_sockets(struct_back * sback, httrackp * opt);
 
 int back_pluggable_sockets_strict(struct_back * sback, httrackp * opt);
 
+/* Randomized inter-file pause target in [min_ms,max_ms] (#185), derived from a
+   timestamp seed so it is stable within one gap and rerolls per launch. */
+int hts_pause_target_ms(TStamp seed, int min_ms, int max_ms);
+
 /* Schedule more links from the heap into free slots. Returns the number queued,
    or <=0 if none could be added (no free slot / paused / stopped). */
 int back_fill(struct_back * sback, httrackp * opt, cache_back * cache,

src/htscoremain.c

@@ -1994,6 +1994,33 @@ static int hts_main_internal(int argc, char **argv, httrackp * opt) {
                   StringCopy(opt->cookies_file, argv[na]);
                 }
                 break;
+              case 'G': // pause: randomized inter-file delay MIN[:MAX] seconds
+                if ((na + 1 >= argc) || (argv[na + 1][0] == '-')) {
+                  HTS_PANIC_PRINTF("Option pause needs a blank space and a "
+                                   "delay in seconds (MIN[:MAX])");
+                  printf("Example: --pause 5:10\n");
+                  htsmain_free();
+                  return -1;
+                } else {
+                  double pmin = 0, pmax = 0;
+                  int nf;
+
+                  na++;
+                  nf = sscanf(argv[na], "%lf:%lf", &pmin, &pmax);
+                  if (nf < 2)
+                    pmax = pmin; /* a single value means a fixed delay */
+                  /* positive-form bounds: NaN fails every comparison, so this
+                     rejects it before the undefined (int)(NaN*1000) cast */
+                  if (nf < 1 || !(pmin >= 0 && pmax >= pmin && pmax <= 86400)) {
+                    HTS_PANIC_PRINTF("Invalid --pause range (expected "
+                                     "MIN[:MAX] seconds, 0<=MIN<=MAX<=86400)");
+                    htsmain_free();
+                    return -1;
+                  }
+                  opt->pause_min_ms = (int) (pmin * 1000.0);
+                  opt->pause_max_ms = (int) (pmax * 1000.0);
+                }
+                break;
               case 't':        /* do not change type (ending) of filenames according to the MIME type */
                 opt->no_type_change = 1;
                 if (*(com+1)=='0') { opt->no_type_change = 0; com++; }

src/htsencoding.c

@@ -30,12 +30,14 @@ Please visit our Website: http://www.httrack.com
 /* Author: Xavier Roche                                         */
 /* ------------------------------------------------------------ */
 
+#include <stdint.h>
+
 #include "htscharset.h"
 #include "htsencoding.h"
 #include "htssafe.h"
 
-/* static int decode_entity(const unsigned int hash, const size_t len);
-*/
+/* static int decode_entity(const uint64_t hash, const size_t len);
+ */
 #include "htsentities.h"
 
 /* hexadecimal conversion */
@@ -50,30 +52,31 @@ static int get_hex_value(char c) {
     return -1;
 }
 
-/* Numerical Recipes,
-   see <http://en.wikipedia.org/wiki/Linear_congruential_generator> */
-#define HASH_PRIME ( 1664525 )
-#define HASH_CONST ( 1013904223 )
-#define HASH_ADD(HASH, C) do {                  \
-    (HASH) *= HASH_PRIME;                       \
-    (HASH) += HASH_CONST;                       \
-    (HASH) += (C);                              \
-  } while(0)
+/* 64-bit FNV-1a; must match htsentities.sh, which keys the entity table on it.
+ */
+#define HASH_INIT 0xcbf29ce484222325ULL
+#define HASH_PRIME 0x100000001b3ULL
+#define HASH_ADD(HASH, C)                                                      \
+  do {                                                                         \
+    (HASH) ^= (unsigned char) (C);                                             \
+    (HASH) *= HASH_PRIME;                                                      \
+  } while (0)
 
 int hts_unescapeEntitiesWithCharset(const char *src, char *dest, const size_t max, const char *charset) {
   size_t i, j, ampStart, ampStartDest;
   int uc;
   int hex;
-  unsigned int hash;
+  uint64_t hash;
 
   assertf(max != 0);
-  for(i = 0, j = 0, ampStart = (size_t) -1, ampStartDest = 0,
-        uc = -1, hex = 0, hash = 0 ; src[i] != '\0' ; i++) {
+  for (i = 0, j = 0, ampStart = (size_t) -1, ampStartDest = 0, uc = -1, hex = 0,
+      hash = HASH_INIT;
+       src[i] != '\0'; i++) {
     /* start of entity */
     if (src[i] == '&') {
       ampStart = i;
       ampStartDest = j;
-      hash = 0;
+      hash = HASH_INIT;
       uc = -1;
     }
     /* inside a potential entity */
@@ -174,14 +177,11 @@ int hts_unescapeEntitiesWithCharset(const char *src, char *dest, const size_t ma
       }
       /* alphanumerical entity */
       else {
-        /* alphanum and not too far ('&thetasym;' is the longest) */
-        if (i <= ampStart + 10 &&
-            (
-             (src[i] >= '0' && src[i] <= '9')
-             || (src[i] >= 'A' && src[i] <= 'Z')
-             || (src[i] >= 'a' && src[i] <= 'z')
-             )
-            ) {
+        /* alphanum, capped at the longest name
+         * '&CounterClockwiseContourIntegral;' (31) */
+        if (i <= ampStart + 31 && ((src[i] >= '0' && src[i] <= '9') ||
+                                   (src[i] >= 'A' && src[i] <= 'Z') ||
+                                   (src[i] >= 'a' && src[i] <= 'z'))) {
           /* compute hash */
           HASH_ADD(hash, (unsigned char) src[i]);
         } else {

src/htsentities.sh

@@ -1,75 +1,92 @@
 #!/bin/bash
 #
+# Regenerate htsentities.h from the WHATWG named character references.
 
-src=html40.txt
-url=http://www.w3.org/TR/1998/REC-html40-19980424/html40.txt
+set -euo pipefail
+
+src=entities.json
+url=https://html.spec.whatwg.org/entities.json
 dest=htsentities.h
 
-(
-    cat <<EOF
-/*
-  -- ${dest} --
-  FILE GENERATED BY $0, DO NOT MODIFY
+# 64-bit FNV-1a of $1, printed as a C constant. Must match the hash in
+# htsencoding.c. The offset basis is stored as its wrapped (signed) bit pattern;
+# bash arithmetic is 64-bit two's complement, so the result is bit-exact.
+fnv1a() {
+    local s=$1 i c h=$((0xcbf29ce484222325))
+    for ((i = 0; i < ${#s}; i++)); do
+        printf -v c '%d' "'${s:i:1}"
+        h=$(((h ^ (c & 0xff)) * 0x100000001b3))
+    done
+    printf '0x%016xULL' "$h"
+}
 
-  We compute the LCG hash
-  (see <http://en.wikipedia.org/wiki/Linear_congruential_generator>)
-  for each entity. We should in theory check using strncmp() that we
-  actually have the correct entity, but this is actually statistically
-  not needed.
+if [ ! -f "$src" ]; then
+    curl -fsS "$url" -o "$src"
+fi
 
-  We may want to do better, but we expect the hash function to be uniform, and
-  let the compiler be smart enough to optimize the switch (for example by
-  checking in log2() intervals)
-  
-  This code has been generated using the evil $0 script.
-*/
+# Keep ';'-terminated single-codepoint names; the ~93 multi-codepoint refs can't
+# fit decode_entity's single-codepoint return and are skipped (left verbatim).
+pairs=$(jq -r '
+    to_entries
+    | map(select((.key | endswith(";")) and (.value.codepoints | length == 1)))
+    | sort_by(.key)
+    | .[] | "\(.key | ltrimstr("&") | rtrimstr(";"))\t\(.value.codepoints[0])"' "$src")
 
-static int decode_entity(const unsigned int hash, const size_t len) {
+# Skipped multi-codepoint names, kept to prove none aliases an emitted hash.
+skipped=$(jq -r '
+    to_entries
+    | map(select((.key | endswith(";")) and (.value.codepoints | length > 1)))
+    | .[] | .key | ltrimstr("&") | rtrimstr(";")' "$src")
+
+cases=""
+emit_hashes=""
+while IFS=$'\t' read -r name cp; do
+    hash=$(fnv1a "$name")
+    cases+="    /* $name */"$'\n'
+    cases+="  case $hash:"$'\n'
+    cases+="    if (len == ${#name}) {"$'\n'
+    cases+="      return $cp;"$'\n'
+    cases+="    }"$'\n'
+    cases+="    break;"$'\n'
+    emit_hashes+="$hash"$'\n'
+done <<<"$pairs"
+
+skip_hashes=""
+while IFS= read -r name; do
+    [ -n "$name" ] && skip_hashes+="$(fnv1a "$name")"$'\n'
+done <<<"$skipped"
+
+# The switch keys on the hash alone, so the dispatch is correct only while every
+# emitted name hashes uniquely; prove it here, no runtime name compare needed.
+dups=$(printf '%s' "$emit_hashes" | sort | uniq -d || true)
+if [ -n "$dups" ]; then
+    echo "FATAL: two entity names share a hash (duplicate switch case); change the hash:" >&2
+    echo "$dups" >&2
+    exit 1
+fi
+# A skipped name colliding with an emitted hash would mis-decode instead of
+# staying verbatim; forbid that too.
+aliased=$(comm -12 <(printf '%s' "$emit_hashes" | sort -u) <(printf '%s' "$skip_hashes" | sort -u) || true)
+if [ -n "$aliased" ]; then
+    echo "FATAL: a skipped multi-codepoint name aliases an emitted hash:" >&2
+    echo "$aliased" >&2
+    exit 1
+fi
+
+cat >"$dest" <<EOF
+/* GENERATED by $0 from the WHATWG named character references
+   (${url}). DO NOT EDIT.
+   Dispatch keys on a 64-bit FNV-1a hash of the entity name; the generator
+   aborts on any hash collision, so no runtime name compare is needed. */
+
+#include <stdint.h>
+
+static int decode_entity(const uint64_t hash, const size_t len) {
   switch(hash) {
-EOF
-    (
-        if test -f ${src}; then
-            cat ${src}
-        else
-            GET "${url}"
-        fi
-    ) |
-        grep -E '^<!ENTITY [a-zA-Z0-9_]' |
-        sed \
-            -e 's/<!ENTITY //' -e "s/[[:space:]][[:space:]]*/ /g" \
-            -e 's/-->$//' \
-            -e 's/\([^ ]*\) CDATA "&#\([^\"]*\);" -- \(.*\)/\1 \2 \3/' |
-        (
-            read -r A
-            while test -n "$A"; do
-                ent="${A%% *}"
-                code=$(echo "$A" | cut -f2 -d' ')
-                # compute hash
-                hash=0
-                i=0
-                a=1664525
-                c=1013904223
-                m="$((1 << 32))"
-                while test "$i" -lt ${#ent}; do
-                    d="$(echo -n "${ent:${i}:1}" | hexdump -v -e '/1 "%d"')"
-                    hash="$((((hash * a) % (m) + d + c) % (m)))"
-                    i=$((i + 1))
-                done
-                echo -e "    /* $A */"
-                echo -e "  case ${hash}u:"
-                echo -e "    if (len == ${#ent} /* && strncmp(ent, \"${ent}\") == 0 */) {"
-                echo -e "      return ${code};"
-                echo -e "    }"
-                echo -e "    break;"
-
-                # next
-                read -r A
-            done
-        )
-    cat <<EOF
-  }
+${cases}  }
   /* unknown */
   return -1;
 }
 EOF
-) >${dest}
+
+echo "wrote $dest ($(grep -c '^  case ' "$dest") entities)" >&2

src/htsfilters.c

@@ -193,7 +193,12 @@ HTS_INLINE const char *strjoker(const char *chaine, const char *joker, LLint * s
           int len = (int) strlen(joker);
 
           while((joker[i] != RIGHT) && (joker[i]) && (i < len)) {
-            if ((joker[i] == '<') || (joker[i] == '>')) {       // *[<10]
+            // '\' escapes the next char as a literal member, e.g. *[\[\]]
+            if (joker[i] == '\\' && joker[i + 1] != '\0') {
+              i++;
+              pass[(int) (unsigned char) joker[i]] = 1;
+              i++;
+            } else if ((joker[i] == '<') || (joker[i] == '>')) { // *[<10]
               int lsize = 0;
               int lverdict;
 
@@ -221,7 +226,9 @@ HTS_INLINE const char *strjoker(const char *chaine, const char *joker, LLint * s
                 while(isdigit((unsigned char) joker[i]))
                   i++;
               }
-            } else if (joker[i + 1] == '-') {   // 2 car, ex: *[A-Z]
+            } else if (joker[i + 1] == '-' && joker[i + 2] != '\0') {
+              // range *[A-Z]; the '\0' guard rejects a truncated *[a- (else
+              // i+=3 overshoots the NUL)
               if ((int) (unsigned char) joker[i + 2] >
                   (int) (unsigned char) joker[i]) {
                 int j;
@@ -233,10 +240,7 @@ HTS_INLINE const char *strjoker(const char *chaine, const char *joker, LLint * s
               }
               // else err=1;
               i += 3;
-            } else {            // 1 car, ex: *[ ]
-              if (joker[i + 2] == '\\' && joker[i + 3] != 0) {  // escaped char, such as *[\[] or *[\]]
-                i++;
-              }
+            } else { // 1 car, ex: *[ ]
               pass[(int) (unsigned char) joker[i]] = 1;
               i++;
             }

src/htsglobal.h

@@ -43,8 +43,8 @@ Please visit our Website: http://www.httrack.com
    configure.ac, decoupled from these). VERSION is the display form, VERSIONID
    the dotted numeric form, AFF_VERSION the short form shown in footers,
    LIB_VERSION the data/cache format generation. */
-#define HTTRACK_VERSION "3.49-9"
-#define HTTRACK_VERSIONID "3.49.9"
+#define HTTRACK_VERSION "3.49-10"
+#define HTTRACK_VERSIONID "3.49.10"
 #define HTTRACK_AFF_VERSION "3.x"
 #define HTTRACK_LIB_VERSION "2.0"
 
@@ -229,6 +229,10 @@ Please visit our Website: http://www.httrack.com
 #define HTS_DEFAULT_FOOTER                                                     \
   "<!-- Mirrored from %s%s by HTTrack Website Copier/" HTTRACK_AFF_VERSION     \
   " " HTTRACK_AFF_AUTHORS ", %s -->"
+/* Honest crawler User-Agent; no fake OS/browser to go stale. */
+#define HTS_DEFAULT_USER_AGENT                                                 \
+  "Mozilla/5.0 (compatible; HTTrack/" HTTRACK_AFF_VERSION                      \
+  "; +https://www.httrack.com/)"
 #define HTTRACK_WEB "http://www.httrack.com"
 #define HTS_UPDATE_WEBSITE                                                     \
   "http://www.httrack.com/"                                                    \

src/htshelp.c

@@ -521,6 +521,7 @@ void help(const char *app, int more) {
   infomsg("  EN maximum mirror time in seconds (60=1 minute, 3600=1 hour)");
   infomsg("  AN maximum transfer rate in bytes/seconds (1000=1KB/s max)");
   infomsg(" %cN maximum number of connections/seconds (*%c10)");
+  infomsg(" %G  random pause of MIN[:MAX] seconds between files (e.g. %G5:10)");
   infomsg
     ("  GN pause transfer if N bytes reached, and wait until lock file is deleted");
   infomsg("");

src/htslib.c

@@ -563,6 +563,39 @@ const char *hts_mime[][2] = {
   {"", ""}
 };
 
+/* Modern web formats (post-2010), kept in their own table: appending to the
+   legacy hts_mime[] above makes clang-format reflow its whole initializer.
+   Scanned after hts_mime[], so it never shadows a legacy mapping. */
+static const char *hts_mime_modern[][2] = {
+    {"image/webp", "webp"},
+    {"image/avif", "avif"},
+    {"image/heic", "heic"},
+    {"font/woff", "woff"},
+    {"font/woff2", "woff2"},
+    {"font/ttf", "ttf"},
+    {"font/otf", "otf"},
+    {"application/json", "json"},
+    {"application/ld+json", "jsonld"},
+    {"application/manifest+json", "webmanifest"},
+    {"application/wasm", "wasm"},
+    {"text/javascript", "js"},
+    {"text/javascript", "mjs"},
+    {"text/markdown", "md"},
+    {"video/mp4", "mp4"},
+    {"video/webm", "webm"},
+    {"video/ogg", "ogv"},
+    {"video/mp2t", "ts"},
+    {"audio/mp4", "m4a"},
+    {"audio/aac", "aac"},
+    {"audio/ogg", "oga"},
+    {"audio/opus", "opus"},
+    {"audio/flac", "flac"},
+    {"audio/webm", "weba"},
+    {"application/x-7z-compressed", "7z"},
+    {"application/x-rar-compressed", "rar"},
+    {"application/zstd", "zst"},
+    {"", ""}};
+
 // Reserved (RFC2396)
 #define CIS(c,ch) ( ((unsigned char)(c)) == (ch) )
 #define CHAR_RESERVED(c)  ( CIS(c,';') \
@@ -1293,16 +1326,12 @@ int http_sendhead(httrackp * opt, t_cookie * cookie, int mode,
 
       // Compression accepted ?
       if (retour->req.http11) {
+        hts_boolean compressible = HTS_FALSE;
 #if HTS_USEZLIB
-        if ((!retour->req.range_used)
-            && (!retour->req.nocompression))
-          print_buffer(&bstr, "Accept-Encoding: " "gzip" /* gzip if the preffered encoding */
-                     ", " "identity;q=0.9" H_CRLF);
-        else
-          print_buffer(&bstr, "Accept-Encoding: identity" H_CRLF);       /* no compression */
-#else
-        print_buffer(&bstr, "Accept-Encoding: identity" H_CRLF); /* no compression */
+        compressible = (!retour->req.range_used && !retour->req.nocompression);
 #endif
+        print_buffer(&bstr, "Accept-Encoding: %s" H_CRLF,
+                     hts_acceptencoding(compressible));
       }
 
       /* Authentification */
@@ -1918,6 +1947,10 @@ HTSEXT_API const char *infostatuscode_const(int statuscode) {
     return "Requested Range Not Satisfiable";
   case 417:
     return "Expectation Failed";
+  case 429:
+    return "Too Many Requests";
+  case 451:
+    return "Unavailable For Legal Reasons";
   case 500:
     return "Internal Server Error";
   case 501:
@@ -4308,6 +4341,20 @@ void guess_httptype(httrackp * opt, char *s, const char *fil) {
   (void) get_httptype_sized(opt, s, HTS_MIMETYPE_SIZE, fil, 1);
 }
 
+// first match in a NUL-terminated {mime,ext} table. key selects the lookup
+// column (0=mime, 1=ext); returns the other column, or NULL if no row matches
+// (a "*" partner means the row carries no value).
+static const char *hts_mime_lookup(const char *(*table)[2], int key,
+                                   const char *needle) {
+  int j;
+
+  for (j = 0; strnotempty(table[j][1]); j++) {
+    if (strfield2(table[j][key], needle) && table[j][!key][0] != '*')
+      return table[j][!key];
+  }
+  return NULL;
+}
+
 // write the mime type for fil into s (capacity ssize)
 // flag: 1 to always return a type (the "application/..." / octet-stream
 // fallback) returns 1 if a type was written to s, 0 otherwise
@@ -4331,17 +4378,15 @@ HTSEXT_API hts_boolean get_httptype_sized(httrackp *opt, char *s, size_t ssize,
     while ((a > fil) && (*a != '.') && (*a != '/'))
       a--;
     if (a >= fil && *a == '.' && strlen(a) < 32) {
-      int j = 0;
+      const char *mime;
 
       a++;
-      while(strnotempty(hts_mime[j][1])) {
-        if (strfield2(hts_mime[j][1], a)) {
-          if (hts_mime[j][0][0] != '*') { // a match exists
-            strlcpybuff(s, hts_mime[j][0], ssize);
-            return 1;
-          }
-        }
-        j++;
+      mime = hts_mime_lookup(hts_mime, 1, a);
+      if (mime == NULL)
+        mime = hts_mime_lookup(hts_mime_modern, 1, a);
+      if (mime != NULL) {
+        strlcpybuff(s, mime, ssize);
+        return 1;
       }
 
       if (flag) {
@@ -4365,6 +4410,11 @@ HTSEXT_API void get_httptype(httrackp *opt, char *s, const char *fil,
   (void) get_httptype_sized(opt, s, HTS_MIMETYPE_SIZE, fil, flag);
 }
 
+/* Advertised Accept-Encoding; gzip and deflate both decode via hts_zunpack */
+const char *hts_acceptencoding(hts_boolean compressible) {
+  return compressible ? "gzip, deflate, identity;q=0.9" : "identity";
+}
+
 // get type of fil (php)
 // s: buffer (text/html) or NULL
 // return: 1 if known by user
@@ -4476,18 +4526,16 @@ int get_userhttptype(httrackp * opt, char *s, const char *fil) {
 // returns 1 if an extension was found (and written to s), 0 otherwise
 int give_mimext(char *s, size_t ssize, const char *st) {
   int ok = 0;
-  int j = 0;
+  const char *ext;
 
   st = hts_effective_mime(st); /* no declared type: derive an html ext */
   s[0] = '\0';
-  while((!ok) && (strnotempty(hts_mime[j][1]))) {
-    if (strfield2(hts_mime[j][0], st)) {
-      if (hts_mime[j][1][0] != '*') { // a match exists
-        strlcpybuff(s, hts_mime[j][1], ssize);
-        ok = 1;
-      }
-    }
-    j++;
+  ext = hts_mime_lookup(hts_mime, 0, st);
+  if (ext == NULL)
+    ext = hts_mime_lookup(hts_mime_modern, 0, st);
+  if (ext != NULL) {
+    strlcpybuff(s, ext, ssize);
+    ok = 1;
   }
   // wrap "x" mimetypes, such as:
   // application/x-mp3
@@ -5754,6 +5802,13 @@ HTSEXT_API int hts_init(void) {
       abortLog("unable to initialize TLS: SSL_CTX_new()");
       assertf("unable to initialize TLS" == NULL);
     }
+    /* Pin a TLS floor (no SSLv3/TLS1.0/1.1); no cert verify, by design. */
+#if OPENSSL_VERSION_NUMBER >= 0x10100000L
+    SSL_CTX_set_min_proto_version(openssl_ctx, TLS1_2_VERSION);
+#else
+    SSL_CTX_set_options(openssl_ctx, SSL_OP_NO_SSLv2 | SSL_OP_NO_SSLv3 |
+                                         SSL_OP_NO_TLSv1 | SSL_OP_NO_TLSv1_1);
+#endif
   }
 #endif
 
@@ -6005,8 +6060,7 @@ HTSEXT_API httrackp *hts_create_opt(void) {
   opt->shell = HTS_FALSE;
   opt->proxy.active = 0;        // pas de proxy
   opt->user_agent_send = HTS_TRUE;
-  StringCopy(opt->user_agent,
-             "Mozilla/4.5 (compatible; HTTrack 3.0x; Windows 98)");
+  StringCopy(opt->user_agent, HTS_DEFAULT_USER_AGENT);
   StringCopy(opt->referer, "");
   StringCopy(opt->from, "");
   opt->savename_83 = HTS_SAVENAME_83_LONG; // long names by default
@@ -6046,6 +6100,8 @@ HTSEXT_API httrackp *hts_create_opt(void) {
   StringCopy(opt->footer, HTS_DEFAULT_FOOTER);
   StringCopy(opt->strip_query, "");
   StringCopy(opt->cookies_file, "");
+  opt->pause_min_ms = 0;
+  opt->pause_max_ms = 0;
   opt->ftp_proxy = HTS_TRUE;
   opt->convert_utf8 = HTS_TRUE;
   StringCopy(opt->filelist, "");

src/htslib.h

@@ -285,6 +285,9 @@ int ishttperror(int err);
 int get_userhttptype(httrackp * opt, char *s, const char *fil);
 int give_mimext(char *s, size_t ssize, const char *st);
 
+/* Advertised Accept-Encoding value (no header name/CRLF); see htslib.c. */
+const char *hts_acceptencoding(hts_boolean compressible);
+
 int may_bogus_multiple(httrackp * opt, const char *mime, const char *filename);
 int may_unknown2(httrackp * opt, const char *mime, const char *filename);
 

src/htsopt.h

@@ -537,6 +537,8 @@ struct httrackp {
   hts_boolean no_query_dedup; /**< with urlhack, keep query-argument order */
   String cookies_file;        /**< extra Netscape cookies.txt to preload
                                  (--cookies-file) */
+  int pause_min_ms; /**< inter-file pause lower bound, ms (0=off, #185) */
+  int pause_max_ms; /**< inter-file pause upper bound, ms */
 };
 
 /* Running statistics for a mirror. */

src/htsparse.c

@@ -302,6 +302,14 @@ static HTS_INLINE char html_prevc(const char *html, const char *start) {
   return html > start ? html[-1] : ' ';
 }
 
+/* Drop a redirect Location's #fragment: a UA anchor, never part of the fetched
+ * resource (#204). */
+static void url_drop_fragment(char *const url) {
+  char *const frag = strchr(url, '#');
+  if (frag != NULL)
+    *frag = '\0';
+}
+
 /* True if [s, s+len) is exactly an HTTP method token (XHR.open's first
    argument is a method, not a URL: #218). Case-insensitive. */
 static int is_http_method(const char *s, size_t len) {
@@ -3596,6 +3604,7 @@ int hts_mirror_check_moved(htsmoduleStruct * str,
         //
 
         strcpybuff(mov_url, r->location);
+        url_drop_fragment(mov_url);
 
         // url qque -> adresse+fichier
         if ((reponse =
@@ -4803,6 +4812,7 @@ int hts_wait_delayed(htsmoduleStruct * str, lien_adrfilsave *afs,
 
             mov_url[0] = '\0';
             strcpybuff(mov_url, back[b].r.location);    // copier URL
+            url_drop_fragment(mov_url);
 
             /* Remove (temporarily created) file if it was created */
             UNLINK(fconv(OPT_GET_BUFF(opt), OPT_GET_BUFF_SIZE(opt), back[b].url_sav));

src/htsselftest.c

@@ -50,6 +50,9 @@ Please visit our Website: http://www.httrack.com
 #include "htscharset.h"
 #include "htsencoding.h"
 #include "htsmd5.h"
+#if HTS_USEZLIB
+#include "htszlib.h"
+#endif
 #include "coucal/coucal.h"
 
 #include <ctype.h>
@@ -239,6 +242,14 @@ static void basic_selftests(void) {
     assertf(strcmp(ext, "html") == 0);
     assertf(give_mimext(ext, sizeof(ext), "no/such-mime-type") == 0);
     assertf(ext[0] == '\0');
+    // modern web formats -> extension. Avoid MIME types the
+    // application/<=4-char-subtype fallback could fabricate without a row.
+    assertf(give_mimext(ext, sizeof(ext), "image/webp") == 1);
+    assertf(strcmp(ext, "webp") == 0);
+    assertf(give_mimext(ext, sizeof(ext), "application/manifest+json") == 1);
+    assertf(strcmp(ext, "webmanifest") == 0);
+    assertf(give_mimext(ext, sizeof(ext), "font/woff2") == 1);
+    assertf(strcmp(ext, "woff2") == 0);
   }
   // convtolower(): lower-cases into the caller buffer (bounded by its size).
   {
@@ -293,6 +304,16 @@ static void basic_selftests(void) {
     assertf(get_httptype_sized(opt, r.contenttype, sizeof(r.contenttype),
                                "x.gif", 0) == 1);
     assertf(strcmp(r.contenttype, "image/gif") == 0);
+    // modern extensions map back to their MIME type
+    assertf(get_httptype_sized(opt, r.contenttype, sizeof(r.contenttype),
+                               "x.webp", 0) == 1);
+    assertf(strcmp(r.contenttype, "image/webp") == 0);
+    assertf(get_httptype_sized(opt, r.contenttype, sizeof(r.contenttype),
+                               "app.wasm", 0) == 1);
+    assertf(strcmp(r.contenttype, "application/wasm") == 0);
+    assertf(get_httptype_sized(opt, r.contenttype, sizeof(r.contenttype),
+                               "mod.mjs", 0) == 1);
+    assertf(strcmp(r.contenttype, "text/javascript") == 0);
     // no extension and flag==0: nothing written, returns 0
     assertf(get_httptype_sized(opt, r.contenttype, sizeof(r.contenttype),
                                "noextfile", 0) == 0);
@@ -512,15 +533,21 @@ static int string_safety_selftests(void) {
 /* ------------------------------------------------------------ */
 
 static int st_filter(httrackp *opt, int argc, char **argv) {
+  char *str, *pat;
+  int matched;
+
   (void) opt;
   if (argc < 2) {
     fprintf(stderr, "filter: needs a filter pattern and a string\n");
     return 1;
   }
-  if (strjoker(argv[1], argv[0], NULL, NULL))
-    printf("%s does match %s\n", argv[1], argv[0]);
-  else
-    printf("%s does NOT match %s\n", argv[1], argv[0]);
+  /* exact-size heap copies so a sanitizer traps any over-read of the pattern */
+  str = strdupt(argv[1]);
+  pat = strdupt(argv[0]);
+  matched = strjoker(str, pat, NULL, NULL) != NULL;
+  printf("%s does %s %s\n", argv[1], matched ? "match" : "NOT match", argv[0]);
+  freet(str);
+  freet(pat);
   return 0;
 }
 
@@ -912,12 +939,58 @@ static int st_copyopt(httrackp *opt, int argc, char **argv) {
   if (strcmp(StringBuff(to->cookies_file), "/tmp/jar.txt") != 0)
     err = 1;
 
+  /* #185 pause pair: copied when enabled (max>0), the 0 sentinel skips */
+  from->pause_min_ms = 5000;
+  from->pause_max_ms = 10000;
+  to->pause_min_ms = to->pause_max_ms = 0;
+  copy_htsopt(from, to);
+  if (to->pause_min_ms != 5000 || to->pause_max_ms != 10000)
+    err = 1;
+  from->pause_min_ms = from->pause_max_ms = 0;
+  copy_htsopt(from, to);
+  if (to->pause_min_ms != 5000 || to->pause_max_ms != 10000)
+    err = 1;
+
   hts_free_opt(from);
   hts_free_opt(to);
   printf("copy-htsopt: %s\n", err ? "FAIL" : "OK");
   return err;
 }
 
+static int st_pause(httrackp *opt, int argc, char **argv) {
+  int err = 0, i, seen_low = 0, seen_high = 0;
+
+  (void) opt;
+  (void) argc;
+  (void) argv;
+  /* Consecutive-ms seeds (production shape: launch timestamps a few ms apart)
+     must stay in range and spread, not collapse to a bound -- worst case for a
+     weak low-bit mixer. */
+  for (i = 0; i < 10000; i++) {
+    int t = hts_pause_target_ms((TStamp) (1719500000000LL + i), 5000, 10000);
+
+    if (t < 5000 || t > 10000)
+      err = 1;
+    seen_low |= (t < 6000);
+    seen_high |= (t > 9000);
+  }
+  if (!seen_low || !seen_high)
+    err = 1;
+  if (hts_pause_target_ms(12345, 8000, 8000) != 8000) /* equal bounds = fixed */
+    err = 1;
+  /* deterministic: a seed yields the same target even after an intervening call
+     with another seed (no global PRNG state to perturb it) */
+  {
+    int a = hts_pause_target_ms(99, 5000, 10000);
+
+    (void) hts_pause_target_ms(54321, 5000, 10000);
+    if (hts_pause_target_ms(99, 5000, 10000) != a)
+      err = 1;
+  }
+  printf("pause: %s\n", err ? "FAIL" : "OK");
+  return err;
+}
+
 static int st_relative(httrackp *opt, int argc, char **argv) {
   char s[HTS_URLMAXSIZE * 2];
 
@@ -1232,6 +1305,192 @@ static int st_urlhack(httrackp *opt, int argc, char **argv) {
   return 0;
 }
 
+/* Default User-Agent: honest HTTrack token, no resurrected Windows 98. */
+static int st_useragent(httrackp *opt, int argc, char **argv) {
+  const char *ua = StringBuff(opt->user_agent);
+  (void) argc;
+  (void) argv;
+  assertf(ua != NULL);
+  assertf(strcmp(ua, HTS_DEFAULT_USER_AGENT) == 0);
+  /* Teeth independent of the macro: honest token + self-identifier, and no
+     legacy Mozilla/4.x fake-browser string (rejects the whole relic family). */
+  assertf(strstr(ua, "HTTrack/") != NULL);
+  assertf(strstr(ua, "+https://www.httrack.com/") != NULL);
+  assertf(strstr(ua, "Mozilla/4.") == NULL);
+  printf("useragent self-test OK: %s\n", ua);
+  return 0;
+}
+
+/* HTTP status code -> reason phrase, including the modern 429/451. */
+static int st_status(httrackp *opt, int argc, char **argv) {
+  const char *s;
+  (void) opt;
+  (void) argc;
+  (void) argv;
+  s = infostatuscode_const(429);
+  assertf(s != NULL && strcmp(s, "Too Many Requests") == 0);
+  s = infostatuscode_const(451);
+  assertf(s != NULL && strcmp(s, "Unavailable For Legal Reasons") == 0);
+  /* A spot-check of a long-standing code, and an unknown one. */
+  s = infostatuscode_const(404);
+  assertf(s != NULL && strcmp(s, "Not Found") == 0);
+  assertf(infostatuscode_const(799) == NULL);
+  printf("status self-test OK\n");
+  return 0;
+}
+
+#if HTS_USEZLIB
+/* Deflate src->path at windowBits (16+ gzip, + zlib, - raw); 0 on success. */
+static int ae_write_packed(const char *path, int windowBits,
+                           const unsigned char *src, size_t len) {
+  unsigned char out[8192];
+  z_stream strm;
+  FILE *f;
+  int zerr;
+
+  memset(&strm, 0, sizeof(strm));
+  if (deflateInit2(&strm, Z_DEFAULT_COMPRESSION, Z_DEFLATED, windowBits, 8,
+                   Z_DEFAULT_STRATEGY) != Z_OK)
+    return 1;
+  if ((f = FOPEN(path, "wb")) == NULL) {
+    deflateEnd(&strm);
+    return 1;
+  }
+  strm.next_in = (Bytef *) src;
+  strm.avail_in = (uInt) len;
+  do {
+    size_t n;
+
+    strm.next_out = out;
+    strm.avail_out = sizeof(out);
+    zerr = deflate(&strm, Z_FINISH);
+    n = sizeof(out) - strm.avail_out;
+    if (n > 0 && fwrite(out, 1, n, f) != n) {
+      deflateEnd(&strm);
+      fclose(f);
+      return 1;
+    }
+  } while (zerr == Z_OK);
+  deflateEnd(&strm);
+  fclose(f);
+  return (zerr == Z_STREAM_END) ? 0 : 1;
+}
+
+/* Forged raw deflate (08 1D) that misdetects as zlib; only fallback decodes */
+static int ae_write_collision(const char *path, const unsigned char *src,
+                              size_t len) {
+  /* block-1 LEN low byte 0x1D: with 0x08, (0x081D)%31==0 */
+  const size_t n1 = 29;
+  size_t n2, p = 0;
+  unsigned char *buf;
+  FILE *f;
+  int ok;
+
+  if (len < n1 || len - n1 > 0xFFFF)
+    return 1;
+  n2 = len - n1;
+  buf = malloct(10 + len);
+  if (buf == NULL)
+    return 1;
+  buf[p++] = 0x08; /* BFINAL=0, BTYPE=00, forged padding -> zlib CMF nibble */
+  buf[p++] = (unsigned char) (n1 & 0xff);
+  buf[p++] = (unsigned char) (n1 >> 8);
+  buf[p++] = (unsigned char) (~n1 & 0xff);
+  buf[p++] = (unsigned char) ((~n1 >> 8) & 0xff);
+  memcpy(buf + p, src, n1);
+  p += n1;
+  buf[p++] = 0x01; /* BFINAL=1, BTYPE=00 */
+  buf[p++] = (unsigned char) (n2 & 0xff);
+  buf[p++] = (unsigned char) (n2 >> 8);
+  buf[p++] = (unsigned char) (~n2 & 0xff);
+  buf[p++] = (unsigned char) ((~n2 >> 8) & 0xff);
+  memcpy(buf + p, src + n1, n2);
+  p += n2;
+  f = FOPEN(path, "wb");
+  ok = (f != NULL && fwrite(buf, 1, p, f) == p);
+  if (f != NULL)
+    fclose(f);
+  freet(buf);
+  return ok ? 0 : 1;
+}
+
+/* Compare path's bytes to expect[0..len); 0 if equal. Streams (large files). */
+static int ae_check_decoded(const char *path, const unsigned char *expect,
+                            size_t len) {
+  unsigned char buf[8192];
+  FILE *f = FOPEN(path, "rb");
+  size_t off = 0, n;
+
+  if (f == NULL)
+    return 1;
+  while ((n = fread(buf, 1, sizeof(buf), f)) > 0) {
+    if (n > len - off || memcmp(buf, expect + off, n) != 0) {
+      fclose(f);
+      return 1;
+    }
+    off += n;
+  }
+  fclose(f);
+  return (off == len) ? 0 : 1;
+}
+#endif
+
+/* Accept-Encoding (#450): advertise gzip+deflate; both decode (hts_zunpack) */
+static int st_acceptencoding(httrackp *opt, int argc, char **argv) {
+  const char *off = hts_acceptencoding(HTS_FALSE);
+  const char *on = hts_acceptencoding(HTS_TRUE);
+
+  (void) opt;
+  assertf(strcmp(off, "identity") == 0);
+  assertf(strstr(on, "gzip") != NULL);
+  assertf(strstr(on, "deflate") != NULL); /* fails on the old gzip-only list */
+#if HTS_USEZLIB
+  if (argc >= 1) {
+    static const int windowBits[] = {16 + MAX_WBITS, MAX_WBITS, -MAX_WBITS};
+    const unsigned char small[] =
+        "deflate round-trip: HTTrack decodes gzip and deflate alike. "
+        "deflate round-trip: HTTrack decodes gzip and deflate alike.";
+    const size_t slen = sizeof(small) - 1;
+    /* 64 KiB of varied (LCG) bytes: forces the multi-fread loop */
+    const size_t blen = 64 * 1024;
+    unsigned char *body = malloct(blen);
+    uint32_t x = 0x1234567u;
+    char inpath[HTS_URLMAXSIZE], outpath[HTS_URLMAXSIZE];
+    size_t i;
+
+    assertf(body != NULL);
+    for (i = 0; i < blen; i++) {
+      x = x * 1103515245u + 12345u;
+      body[i] = (unsigned char) (x >> 16);
+    }
+    /* gzip, zlib (RFC1950) and raw deflate (RFC1951), both small and large. */
+    for (i = 0; i < sizeof(windowBits) / sizeof(windowBits[0]); i++) {
+      snprintf(inpath, sizeof(inpath), "%s/ae-in-%d.z", argv[0], windowBits[i]);
+      snprintf(outpath, sizeof(outpath), "%s/ae-out-%d", argv[0],
+               windowBits[i]);
+      assertf(ae_write_packed(inpath, windowBits[i], small, slen) == 0);
+      assertf(hts_zunpack(inpath, outpath) == (int) slen);
+      assertf(ae_check_decoded(outpath, small, slen) == 0);
+      assertf(ae_write_packed(inpath, windowBits[i], body, blen) == 0);
+      assertf(hts_zunpack(inpath, outpath) == (int) blen);
+      assertf(ae_check_decoded(outpath, body, blen) == 0);
+    }
+    /* Fallback teeth: raw deflate misdetected as zlib; -1 without the retry. */
+    snprintf(inpath, sizeof(inpath), "%s/ae-collide.z", argv[0]);
+    snprintf(outpath, sizeof(outpath), "%s/ae-collide.out", argv[0]);
+    assertf(ae_write_collision(inpath, body, 64) == 0);
+    assertf(hts_zunpack(inpath, outpath) == 64);
+    assertf(ae_check_decoded(outpath, body, 64) == 0);
+    freet(body);
+  }
+#else
+  (void) argc;
+  (void) argv;
+#endif
+  printf("acceptencoding self-test OK: %s\n", on);
+  return 0;
+}
+
 /* ------------------------------------------------------------ */
 /* Registry: name -> handler, with a usage hint and a one-line description. */
 /* ------------------------------------------------------------ */
@@ -1264,6 +1523,7 @@ static const struct selftest_entry {
     {"strsafe", "[overflow|overflow-buff [str]]", "bounded string-op self-test",
      st_strsafe},
     {"copyopt", "", "copy_htsopt option-copy self-test", st_copyopt},
+    {"pause", "", "randomized inter-file pause target self-test", st_pause},
     {"relative", "<link> <curr-file>", "relative link between two paths",
      st_relative},
     {"resolve", "<link> <adr> <fil>", "resolve a link against an origin",
@@ -1277,6 +1537,10 @@ static const struct selftest_entry {
      st_cache_writefail},
     {"dns", "", "DNS resolver/cache self-test", st_dns},
     {"cookies", "", "cookie request-header self-test", st_cookies},
+    {"useragent", "", "default User-Agent self-test", st_useragent},
+    {"status", "", "HTTP status code -> reason phrase self-test", st_status},
+    {"acceptencoding", "[dir]",
+     "Accept-Encoding advertises gzip+deflate, both decode", st_acceptencoding},
 };
 
 static void list_selftests(void) {

src/htsserver.c

@@ -358,12 +358,12 @@ int smallserver(T_SOC soc, char *url, char *method, char *data, char *path) {
       {NULL, 0}
     };
     initStrElt initStr[] = {
-      {"user", "Mozilla/4.5 (compatible; HTTrack 3.0x; Windows 98)"},
-      {"footer",
-       "<!-- Mirrored from %s%s by HTTrack Website Copier/3.x [XR&CO'2014], %s -->"},
-      {"url2", "+*.png +*.gif +*.jpg +*.jpeg +*.css +*.js -ad.doubleclick.net/*"},
-      {NULL, NULL}
-    };
+        {"user", HTS_DEFAULT_USER_AGENT},
+        {"footer", "<!-- Mirrored from %s%s by HTTrack Website Copier/3.x "
+                   "[XR&CO'2014], %s -->"},
+        {"url2",
+         "+*.png +*.gif +*.jpg +*.jpeg +*.css +*.js -ad.doubleclick.net/*"},
+        {NULL, NULL}};
     int i = 0;
 
     for(i = 0; initInt[i].name; i++) {

src/htswizard.c

@@ -80,6 +80,10 @@ htspair_t hts_detect_embed[] = {
   {NULL, NULL}
 };
 
+/* HTML5 media siblings of <img src>: same near-link treatment (#451) */
+static const htspair_t hts_detect_embed_html5[] = {
+    {"source", "src"}, {"source", "srcset"}, {"track", "src"}, {NULL, NULL}};
+
 /* Internal */
 static int hts_acceptlink_(httrackp * opt, int ptr, const char *adr,
                            const char *fil, const char *tag,
@@ -136,6 +140,17 @@ static int cmp_token(const char *tag, const char *cmp) {
           && !isalnum((unsigned char) tag[p]));
 }
 
+/* TRUE if (tag, attribute) matches an embedded-asset pair in the table */
+static hts_boolean is_embed_pair(const htspair_t *table, const char *tag,
+                                 const char *attribute) {
+  int i;
+  for (i = 0; table[i].tag != NULL; i++) {
+    if (cmp_token(tag, table[i].tag) && cmp_token(attribute, table[i].attr))
+      return HTS_TRUE;
+  }
+  return HTS_FALSE;
+}
+
 static int hts_acceptlink_(httrackp * opt, int ptr,
                            const char *adr, const char *fil, const char *tag,
                            const char *attribute, int *set_prio_to,
@@ -163,15 +178,9 @@ static int hts_acceptlink_(httrackp * opt, int ptr,
 
   /* Built-in known tags (<img src=..>, ..) */
   if (forbidden_url != 0 && opt->nearlink && tag != NULL && attribute != NULL) {
-    int i;
-
-    for(i = 0; hts_detect_embed[i].tag != NULL; i++) {
-      if (cmp_token(tag, hts_detect_embed[i].tag)
-          && cmp_token(attribute, hts_detect_embed[i].attr)
-        ) {
-        embedded_triggered = 1;
-        break;
-      }
+    if (is_embed_pair(hts_detect_embed, tag, attribute) ||
+        is_embed_pair(hts_detect_embed_html5, tag, attribute)) {
+      embedded_triggered = 1;
     }
   }
 

src/htszlib.c

@@ -47,48 +47,89 @@ Please visit our Website: http://www.httrack.com
 */
 
 /*
-  Unpack file into a new file
+  Unpack file into a new file (gzip, zlib RFC1950 or raw deflate RFC1951).
   Return value: size of the new file, or -1 if an error occurred
 */
 /* Note: utf-8 */
 int hts_zunpack(char *filename, char *newfile) {
   int ret = -1;
 
-  if (filename != NULL && newfile != NULL) {
-    if (filename[0] && newfile[0]) {
-      char catbuff[CATBUFF_SIZE];
-      FILE *const in = FOPEN(fconv(catbuff, sizeof(catbuff), filename), "rb");
-      const int fd = in != NULL ? fileno(in) : -1;
-      const int dup_fd = fd != -1 ? dup(fd) : -1;
-      // Note: we must dup to be able to flose cleanly.
-      const gzFile gz = dup_fd != -1 ? gzdopen(dup_fd, "rb") : NULL;
+  if (filename != NULL && newfile != NULL && filename[0] && newfile[0]) {
+    char catbuff[CATBUFF_SIZE];
+    FILE *const in = FOPEN(fconv(catbuff, sizeof(catbuff), filename), "rb");
 
-      if (gz) {
-        FILE *const fpout = FOPEN(fconv(catbuff, sizeof(catbuff), newfile), "wb");
-        int size = 0;
+    if (in != NULL) {
+      unsigned char BIGSTK inbuf[8192];
+      size_t navail = fread(inbuf, 1, sizeof(inbuf), in);
+      /* gzip/zlib headers -> +32 windowBits; else raw deflate (RFC1951) */
+      const hts_boolean wrapped =
+          (navail >= 2 &&
+           ((inbuf[0] == 0x1f && inbuf[1] == 0x8b) ||
+            ((inbuf[0] & 0x0f) == Z_DEFLATED &&
+             (((unsigned) inbuf[0] << 8 | inbuf[1]) % 31) == 0)));
+      int attempt;
 
-        if (fpout) {
-          int nr;
+      /* deflate is ambiguous; on failure retry with the other windowBits */
+      for (attempt = 0; attempt < 2 && ret < 0; attempt++) {
+        const int windowBits =
+            (attempt == 0 ? wrapped : !wrapped) ? (32 + MAX_WBITS) : -MAX_WBITS;
+        FILE *fpout;
+        z_stream strm;
 
-          do {
-            char BIGSTK buff[1024];
-
-            nr = gzread(gz, buff, sizeof(buff));
-            if (nr > 0) {
-              size += nr;
-              if (fwrite(buff, 1, nr, fpout) != nr)
-                nr = size = -1;
-            }
-          } while(nr > 0);
+        if (attempt > 0) {
+          /* rewind input; reopening fpout "wb" discards the partial output */
+          if (fseek(in, 0, SEEK_SET) != 0)
+            break;
+          navail = fread(inbuf, 1, sizeof(inbuf), in);
+        }
+        fpout = FOPEN(fconv(catbuff, sizeof(catbuff), newfile), "wb");
+        if (fpout == NULL)
+          break;
+        memset(&strm, 0, sizeof(strm));
+        if (inflateInit2(&strm, windowBits) != Z_OK) {
           fclose(fpout);
-        } else
-          size = -1;
-        gzclose(gz);
-        ret = (int) size;
-      }
-      if (in != NULL) {
-        fclose(in);
+          break;
+        }
+        {
+          hts_boolean ok = HTS_TRUE;
+          int size = 0;
+          int zerr = Z_OK;
+
+          /* chunked inflate; first chunk in inbuf, single member */
+          do {
+            strm.next_in = inbuf;
+            strm.avail_in = (uInt) navail;
+            do {
+              unsigned char BIGSTK outbuf[8192];
+              size_t produced;
+
+              strm.next_out = outbuf;
+              strm.avail_out = sizeof(outbuf);
+              zerr = inflate(&strm, Z_NO_FLUSH);
+              if (zerr == Z_NEED_DICT || zerr == Z_DATA_ERROR ||
+                  zerr == Z_MEM_ERROR || zerr == Z_STREAM_ERROR) {
+                ok = HTS_FALSE;
+                break;
+              }
+              produced = sizeof(outbuf) - strm.avail_out;
+              if (produced > 0 &&
+                  fwrite(outbuf, 1, produced, fpout) != produced) {
+                ok = HTS_FALSE;
+                break;
+              }
+              size += (int) produced;
+            } while (strm.avail_out == 0);
+            if (!ok || zerr == Z_STREAM_END)
+              break;
+            navail = fread(inbuf, 1, sizeof(inbuf), in);
+          } while (navail > 0);
+          if (ok && zerr == Z_STREAM_END)
+            ret = size;
+        }
+        inflateEnd(&strm);
+        fclose(fpout);
       }
+      fclose(in);
     }
   }
   return ret;

src/proxy/proxytrack.c

@@ -497,6 +497,12 @@ static const char *GetHttpMessage(int statuscode) {
   case 417:
     return "Expectation Failed";
     break;
+  case 429:
+    return "Too Many Requests";
+    break;
+  case 451:
+    return "Unavailable For Legal Reasons";
+    break;
   case 500:
     return "Internal Server Error";
     break;

tests/01_engine-cmdline.test

@@ -90,4 +90,16 @@ refused "dangling-quote argument not refused cleanly"
 run_only "$tmp/q-lone" '"'
 refused "lone-quote argument not refused cleanly"
 
+# --pause (#185): valid MIN[:MAX] accepted; malformed, reversed, over-range and
+# non-finite values refused cleanly. NaN defeats naive `<`/`>` checks (it
+# compares false to everything), so it must not slip through to the int cast.
+run "$tmp/pause-ok" --pause 0.2:0.4
+accepted "$tmp/pause-ok" "#185: valid --pause range rejected"
+run "$tmp/pause-fix" --pause 0.2
+accepted "$tmp/pause-fix" "#185: valid fixed --pause rejected"
+for bad in nan nan:5 5:nan inf 10:5 99999; do
+    run "$tmp/pause-bad" --pause "$bad"
+    refused "#185: invalid --pause '$bad' not refused cleanly"
+done
+
 exit 0

tests/01_engine-entities.test

@@ -18,6 +18,21 @@ ent '&' '&'
 ent '&lt;&gt;' '<>'
 ent '&eacute;' 'é'
 
+# HTML5 names from the WHATWG set
+ent '&hellip;' '…'
+ent '&bigcup;' '⋃'
+# longest name (31 chars) exercises the name-length cap
+ent '&CounterClockwiseContourIntegral;' '∳'
+# astral codepoint -> 4-byte UTF-8
+ent '&Aopf;' '𝔸'
+# multi-codepoint refs are skipped at generation, so left verbatim
+ent '&fjlig;' '&fjlig;'
+
+# common HTML4 names still decode (regression guard against accidental drops)
+ent '&copy;&reg;&trade;' '©®™'
+ent '&mdash;&ndash;' '—–'
+ent '&alpha;&beta;' 'αβ'
+
 # numeric: decimal and hex
 ent '&#65;&#66;' 'AB'
 ent '&#x41;' 'A'

tests/01_engine-filter.test

@@ -50,27 +50,54 @@ match '*foo*bar' 'foozbar'
 # '?' is the query-string marker, not a single-char wildcard
 nomatch 'a?c' 'abc'
 
-# backslash escapes a metacharacter inside a class so it is matched literally.
-# Quirk: the decoder also adds the backslash itself to the set, so '\X' matches
-# both X and '\'. These assertions pin that behavior.
+# Inside a class, backslash escapes the next char as a literal member (#148):
+# '\X' matches X only (not '\'), and an escaped ']' is a member, not the terminator.
 match '*[\*]' '*'
-match '*[\*]' "\\"
-nomatch '*[\*]' 'a'
+nomatch '*[\*]' "\\"
 match '*[\\]' "\\"
-nomatch '*[\\]' 'a'
+nomatch '*[\\]' '*'
 match '*[\[]' '['
-match '*[\[]' "\\"
-nomatch '*[\[]' 'a'
+nomatch '*[\[]' "\\"
+match '*[\]]' ']'
+nomatch '*[\]]' "\\"
 
-# A literal ']' cannot be a class member: the class parser stops at the first
-# ']', escaped or not. So '*[\[\]]' does NOT mean "the [ or ] character" as the
-# filter guide claims (GitHub #148); it parses as the class {'[','\'} followed
-# by a trailing literal ']'. These assertions document the current (buggy)
-# behavior so any future matcher fix is a deliberate, visible change.
-nomatch '*[\[\]]' '[' # not matched, despite the docs
-match '*[\[\]]' ']'   # only via the empty class-match + trailing ']'
-match '*[\[\]]' '[]'  # one of {'[','\'} then the trailing ']'
-nomatch '*[\[\]]' '[]x'
+# '*[\[\]]' is "the [ or ] character", as the filter guide documents.
+match '*[\[\]]' '['
+match '*[\[\]]' ']'
+nomatch '*[\[\]]' 'a'
+match '*[\[,\]]' '[' # comma between members is optional
+match '*[\[,\]]' ']'
+match '*[a,\[]' 'a' # an escaped member no longer eats the preceding one
+match '*[a,\[]' '['
+
+# Escape is decoded before the range/separator/size checks, so '\-' '\,' '\<'
+# are literal members, not operators.
+match '*[a\-z]' 'a'
+match '*[a\-z]' 'z'
+nomatch '*[a\-z]' 'b' # not the a..z range
+match '*[\,]' ','
+nomatch '*[\,]' "\\" # the escape must not leak '\' into the class
+match '*[\<]' '<'
+nomatch '*[\<]' "\\"
+match '*[\[,\],a]' '['
+match '*[\[,\],a]' ']'
+match '*[\[,\],a]' 'a'
+
+# A truncated range '*[a-' is the literal members {a,-}; the parser must not
+# read past the end decoding it (was a 1-byte heap over-read in the range arm).
+match '*[a-' 'a'
+nomatch '*[a-' 'b'
+
+# *(...) matches exactly one char from the class; *[...] matches a run.
+match '*(a,b)' 'a'
+nomatch '*(a,b)' 'aa'
+nomatch '*(a,b)' 'c'
+
+# documented composite filters (filters.html)
+match 'www.*[path].com/*[path].zip' 'www.foo.com/a/b.zip'
+nomatch 'www.*[path].com/*[path].zip' 'www.foo.com/a/b.tar'
+match '*.html*[]' 'page.html'
+nomatch '*.html*[]' 'page.html?x=1' # *[] forbids the trailing query
 
 # Size-based rules (-#test=filtersize <size> <string> <filter...>): a negative size
 # means the size is still unknown (scan time). A size exclusion must stay neutral

tests/01_engine-parse.test

@@ -323,4 +323,33 @@ grep -Fq 'href="ahref%20(4).gif"' "$saved9" ||
 ! grep -Eq '(src|href)="[^"]*%28' "$saved9" ||
     ! echo "FAIL #163: gate over-fired onto a non-url() attribute link" || exit 1
 
+# HTML5 <source>/<track> follow as embedded near-links past the -r2 depth boundary (#451).
+# img.gif positive control; plain.gif (bare <a href>) negative control proves the gate is selective.
+site10="$tmp/html5media"
+mkdir -p "$site10"
+for f in img ss plain; do gif "$site10/$f.gif"; done
+printf 'x' >"$site10/v.webm"
+printf 'x' >"$site10/subs.vtt"
+cat >"$site10/index.html" <<EOF
+<html><body><a href="leaf.html">leaf</a></body></html>
+EOF
+cat >"$site10/leaf.html" <<EOF
+<html><body>
+<img src="img.gif">
+<picture><source srcset="ss.gif 2x"></picture>
+<video><source src="v.webm"></video>
+<video><track src="subs.vtt"></video>
+<a href="plain.gif">plain link past the boundary</a>
+</body></html>
+EOF
+out10="$tmp/html5media-out"
+rm -rf "$out10"
+mkdir -p "$out10"
+httrack "file://$site10/index.html" -O "$out10" --quiet --near -r2 >"$out10/.log" 2>&1 || true
+found "img.gif" "$out10"
+found "ss.gif" "$out10"
+found "v.webm" "$out10"
+found "subs.vtt" "$out10"
+notfound "plain.gif" "$out10"
+
 exit 0

tests/01_engine-pause.test

@@ -0,0 +1,15 @@
+#!/bin/bash
+#
+# --pause (#185): the inter-file pause target must stay in [min,max] and spread
+# across it (a per-call rand() would collapse it toward min). Driven by the
+# in-process 'httrack -#test=pause' test. POSIX-portable ($(BASH) is /bin/sh on macOS).
+
+set -eu
+
+# 'run' is an ignored placeholder argument.
+out=$(httrack -#test=pause run)
+
+test "$out" = "pause: OK" || {
+    echo "expected 'pause: OK', got: $out" >&2
+    exit 1
+}

tests/01_engine-status.test

@@ -0,0 +1,7 @@
+#!/bin/bash
+#
+
+set -euo pipefail
+
+# HTTP status -> reason phrase, including the modern 429/451 (#453).
+httrack -O /dev/null -#test=status run | grep -q "status self-test OK"

tests/01_engine-useragent.test

@@ -0,0 +1,7 @@
+#!/bin/bash
+#
+
+set -euo pipefail
+
+# Default User-Agent (#449): honest HTTrack token, no Windows 98 relic.
+httrack -O /dev/null -#test=useragent run | grep -q "useragent self-test OK"

tests/01_zlib-acceptencoding.test

@@ -0,0 +1,11 @@
+#!/bin/bash
+#
+
+set -euo pipefail
+
+# Accept-Encoding (#450): advertise gzip+deflate; decode gzip/zlib/raw-deflate.
+dir=$(mktemp -d)
+trap 'rm -rf "$dir"' EXIT
+
+httrack -O /dev/null -#test=acceptencoding "$dir" run |
+    grep -q "acceptencoding self-test OK"

tests/01_zlib-cache-golden.test

@@ -6,7 +6,7 @@
 
 # Golden cache-format regression test (driven by 'httrack -#test=cache-golden <dir>').
 #
-# 01_engine-cache.test writes the cache with the same build it reads back (a
+# 01_zlib-cache.test writes the cache with the same build it reads back (a
 # round-trip), so it cannot catch a read-path or ZIP-format regression where
 # writer and reader drift together. This reads a *committed* cache frozen by an
 # earlier build and asserts a fixed set of entries still decodes field- and

tests/28_local-pause.test

@@ -0,0 +1,36 @@
+#!/bin/bash
+#
+# --pause (#185): a fixed inter-file delay must slow a multi-file crawl. Measure
+# the same crawl with and without --pause and compare: the harness overhead
+# cancels, leaving only the pause. Integer seconds keep it portable (BSD date
+# has no %N); a lower bound is not timing-flaky since a pause only adds time.
+
+set -e
+
+: "${top_srcdir:=..}"
+
+# python3 runs the local server (mirror local-crawl.sh); skip when absent, else
+# run() swallows its exit-77 and the serverless 0s/0s crawl looks like a fail.
+command -v python3 >/dev/null || {
+    echo "python3 not found; skipping local crawl tests"
+    exit 77
+}
+
+run() { # echoes the wall-clock seconds of one crawl
+    local t0 t1
+    t0=$(date +%s)
+    bash "$top_srcdir/tests/local-crawl.sh" --errors 0 \
+        httrack 'BASEURL/types/index.html' -c1 "$@" >/dev/null 2>&1
+    t1=$(date +%s)
+    echo $((t1 - t0))
+}
+
+base=$(run)
+paused=$(run --pause 0.5)
+delta=$((paused - base))
+
+echo "crawl: ${base}s, with --pause 0.5: ${paused}s (delta ${delta}s)"
+if [ "$delta" -lt 2 ]; then
+    echo "FAIL: --pause did not delay the crawl (delta ${delta}s)" >&2
+    exit 1
+fi

tests/29_local-redirect-fragment.test

@@ -0,0 +1,11 @@
+#!/bin/bash
+# Issue #204: a 302 Location with a #fragment must drop the fragment before the
+# target is fetched. The server is strict (400 on a '#' in the request-target),
+# so a leaked fragment logs an error and the target is never saved.
+set -e
+
+: "${top_srcdir:=..}"
+
+bash "$top_srcdir/tests/local-crawl.sh" --errors 0 \
+    --found 'redir/target.html' \
+    httrack 'BASEURL/redir/index.html'

tests/Makefile.am

@@ -1,4 +1,4 @@
-# Committed binary fixture read by 01_engine-cache-golden.test. List it
+# Committed binary fixture read by 01_zlib-cache-golden.test. List it
 # explicitly: automake does not expand wildcards in EXTRA_DIST, so a glob would
 # silently drop it from the dist tarball and break "make distcheck".
 EXTRA_DIST = $(TESTS) crawl-test.sh run-all-tests.sh check-network.sh \
@@ -25,9 +25,6 @@ TEST_EXTENSIONS = .test
 TEST_LOG_COMPILER = $(BASH)
 TESTS = \
 	00_runnable.test \
-	01_engine-cache.test \
-	01_engine-cache-golden.test \
-	01_engine-cache-writefail.test \
 	01_engine-charset.test \
 	01_engine-cmdline.test \
 	01_engine-cookies.test \
@@ -41,14 +38,21 @@ TESTS = \
 	01_engine-idna.test \
 	01_engine-mime.test \
 	01_engine-parse.test \
+	01_engine-pause.test \
 	01_engine-rcfile.test \
 	01_engine-relative.test \
 	01_engine-savename.test \
 	01_engine-selftest-dispatch.test \
 	01_engine-simplify.test \
+	01_engine-status.test \
 	01_engine-stripquery.test \
 	01_engine-strsafe.test \
 	01_engine-urlhack.test \
+	01_engine-useragent.test \
+	01_zlib-acceptencoding.test \
+	01_zlib-cache.test \
+	01_zlib-cache-golden.test \
+	01_zlib-cache-writefail.test \
 	02_manpage-regen.test \
 	02_update-cache.test \
 	10_crawl-simple.test \
@@ -73,6 +77,8 @@ TESTS = \
 	24_local-resume-overlap.test \
 	25_local-mime-exclude.test \
 	26_local-strip-query.test \
-	27_local-cookies-file.test
+	27_local-cookies-file.test \
+	28_local-pause.test \
+	29_local-redirect-fragment.test
 
 CLEANFILES = check-network_sh.cache

tests/local-server.py

@@ -354,6 +354,21 @@ class Handler(SimpleHTTPRequestHandler):
         if self.command != "HEAD":
             self.wfile.write(body)
 
+    # 302 whose Location carries a #fragment (#204): the fragment is a UA anchor
+    # that must be dropped before the target is fetched. A leaked '#' reaches the
+    # strict-server guard below and 400s.
+    def route_redir_index(self):
+        self.send_html('\t<a href="go.php">go</a>')
+
+    def route_redir_go(self):
+        self.send_response(302, "Found")
+        self.send_header("Location", "target.html#section")
+        self.send_header("Content-Length", "0")
+        self.end_headers()
+
+    def route_redir_target(self):
+        self.send_raw(b"<html><body>redirect target</body></html>\n", "text/html")
+
     ROUTES = {
         "/cookies/entrance.php": route_entrance,
         "/cookies/second.php": route_second,
@@ -391,10 +406,23 @@ class Handler(SimpleHTTPRequestHandler):
         "/mimex/index.html": route_mimex_index,
         "/mimex/blob.pdf": route_mimex_blob,
         "/mimex/real.html": route_mimex_real,
+        "/redir/index.html": route_redir_index,
+        "/redir/go.php": route_redir_go,
+        "/redir/target.html": route_redir_target,
     }
 
     # --- dispatch ----------------------------------------------------------
 
+    def reject_fragment(self):
+        # Strict server: a '#' in the request-target is the client failing to
+        # drop a fragment (#204). RFC 3986 forbids it on the wire; answer 400.
+        if "#" in self.path:
+            self.send_response(400, "Bad Request")
+            self.send_header("Content-Length", "0")
+            self.end_headers()
+            return True
+        return False
+
     def dispatch(self):
         self._set_cookies = []
         path = urlsplit(self.path).path
@@ -406,10 +434,14 @@ class Handler(SimpleHTTPRequestHandler):
         return False
 
     def do_GET(self):
+        if self.reject_fragment():
+            return
         if not self.dispatch():
             super().do_GET()
 
     def do_HEAD(self):
+        if self.reject_fragment():
+            return
         if not self.dispatch():
             super().do_HEAD()