htsparse: follow same-file redirects instead of self-pointing stubs (#159) (#471)

An http->https redirect (or any alias where the savename strips a
component the URL keeps) collapses the source and target onto one saved
file. hts_mirror_check_moved wrote a "Page has moved" stub linking to the
target's savename, but that equals the source's, so the stub pointed at
itself and the real content was never saved.

Detect a same-file alias with a new hts_redirect_same_savefile helper
(scheme and userinfo stripped, www kept, path slash/query-normalized per
the URL-hack dedup flags) and follow the redirect through: record the
moved link at the same savename so its content overwrites the placeholder.
Genuinely-different moves keep the stub. The old informational "URL Hack
identical" log is superseded by an actionable message on the followed
redirect. Covered by a redirect-samefile engine self-test.

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

src/htsparse.c

@@ -108,62 +108,47 @@ Please visit our Website: http://www.httrack.com
 
 #define HT_ADD_FOP
 
+/* Mutable extended-context fields: one source of truth so the DEFINE/SET/SAVE
+   load and store lists can't drift apart. */
+/* clang-format off */
+#define ENGINE_MUTABLE_FIELDS(X) \
+  X(int, error, stre->error_) \
+  X(int, store_errpage, stre->store_errpage_) \
+  X(int, makeindex_done, stre->makeindex_done_) \
+  X(FILE *, makeindex_fp, stre->makeindex_fp_) \
+  X(int, makeindex_links, stre->makeindex_links_) \
+  X(LLint, stat_fragment, stre->stat_fragment_)
+
+#define ENGINE_FIELD_DECLARE(type, name, src) type name = *(src);
+#define ENGINE_FIELD_LOAD(type, name, src) name = *(src);
+#define ENGINE_FIELD_STORE(type, name, src) *(src) = name;
+
 #define ENGINE_DEFINE_CONTEXT() \
   ENGINE_DEFINE_CONTEXT_BASE(); \
-  /* */ \
   htsblk* const r HTS_UNUSED = stre->r_; \
   hash_struct* const hash HTS_UNUSED = stre->hash_; \
   char* const codebase HTS_UNUSED = stre->codebase; \
   char* const base HTS_UNUSED = stre->base; \
-  /* */ \
   const char * const template_header HTS_UNUSED = stre->template_header_; \
   const char * const template_body HTS_UNUSED = stre->template_body_; \
   const char * const template_footer HTS_UNUSED = stre->template_footer_; \
-  /* */ \
   HTS_UNUSED char* const makeindex_firstlink = stre->makeindex_firstlink_; \
-  /* */ \
-  /* */ \
-  int error = * stre->error_; \
-  int store_errpage = * stre->store_errpage_; \
-  /* */ \
-  int makeindex_done = *stre->makeindex_done_; \
-  FILE* makeindex_fp = *stre->makeindex_fp_; \
-  int makeindex_links = *stre->makeindex_links_; \
-  /* */ \
-  LLint stat_fragment = *stre->stat_fragment_; \
+  ENGINE_MUTABLE_FIELDS(ENGINE_FIELD_DECLARE) \
+  /* load-once (kept out of SET/SAVE): re-reading would reset the throttle */ \
   HTS_UNUSED TStamp makestat_time = stre->makestat_time; \
   HTS_UNUSED FILE* makestat_fp = stre->makestat_fp
 
-/* clang-format off: an edit realigns all backslashes, churning the macro. */
-/* clang-format off */
-/* Load-once: re-reading resets makestat_time (mutated locally, never SAVEd). */
 #define ENGINE_SET_CONTEXT() \
   ENGINE_SET_CONTEXT_BASE(); \
-  /* */ \
-  error = * stre->error_; \
-  store_errpage = * stre->store_errpage_; \
-  /* */ \
-  makeindex_done = *stre->makeindex_done_; \
-  makeindex_fp = *stre->makeindex_fp_; \
-  makeindex_links = *stre->makeindex_links_; \
-  /* */ \
-  stat_fragment = *stre->stat_fragment_
-/* clang-format on */
+  ENGINE_MUTABLE_FIELDS(ENGINE_FIELD_LOAD)
 
 #define ENGINE_LOAD_CONTEXT() \
   ENGINE_DEFINE_CONTEXT()
 
 #define ENGINE_SAVE_CONTEXT() \
   ENGINE_SAVE_CONTEXT_BASE(); \
-  /* */ \
-  * stre->error_ = error; \
-  * stre->store_errpage_ = store_errpage; \
-  /* */ \
-  *stre->makeindex_done_ = makeindex_done; \
-  *stre->makeindex_fp_ = makeindex_fp; \
-  *stre->makeindex_links_ = makeindex_links; \
-  /* */ \
-  *stre->stat_fragment_ = stat_fragment
+  ENGINE_MUTABLE_FIELDS(ENGINE_FIELD_STORE)
+/* clang-format on */
 
 #define _ROBOTS      ((robots_wizard*)opt->robotsptr)
 
@@ -3503,6 +3488,24 @@ int htsparse(htsmoduleStruct * str, htsmoduleStructExtended * stre) {
   return 0;
 }
 
+/* Mirror the savename to tell whether a redirect saves to the same file (#159);
+ * contract in htsparse.h. */
+hts_boolean hts_redirect_same_savefile(httrackp *opt, const char *cur_adr,
+                                       const char *cur_fil,
+                                       const char *moved_adr,
+                                       const char *moved_fil) {
+  const int norm_slash = opt->urlhack && !opt->no_slash_dedup;
+  const int norm_query = opt->urlhack && !opt->no_query_dedup;
+  char BIGSTK n_fil[HTS_URLMAXSIZE * 2], pn_fil[HTS_URLMAXSIZE * 2];
+
+  if (strcasecmp(jump_identification_const(moved_adr),
+                 jump_identification_const(cur_adr)) != 0)
+    return HTS_FALSE;
+  fil_normalized_filtered_ex(moved_fil, n_fil, NULL, norm_slash, norm_query);
+  fil_normalized_filtered_ex(cur_fil, pn_fil, NULL, norm_slash, norm_query);
+  return strcasecmp(n_fil, pn_fil) == 0;
+}
+
 /*
 Check 301, 302, .. statuscodes (moved)
 */
@@ -3548,36 +3551,9 @@ int hts_mirror_check_moved(htsmoduleStruct * str,
         if ((reponse =
              ident_url_relatif(mov_url, urladr(), urlfil(), moved)) >= 0) {
           int set_prio_to = 0;  // pas de priotité fixéd par wizard
-
-          // check whether URLHack is harmless or not (per the effective
-          // sub-flags)
-          if (opt->urlhack && (!opt->no_www_dedup || !opt->no_slash_dedup ||
-                               !opt->no_query_dedup)) {
-            const int norm_host = !opt->no_www_dedup;
-            const int norm_slash = !opt->no_slash_dedup;
-            const int norm_query = !opt->no_query_dedup;
-            char BIGSTK n_adr[HTS_URLMAXSIZE * 2], n_fil[HTS_URLMAXSIZE * 2];
-            char BIGSTK pn_adr[HTS_URLMAXSIZE * 2], pn_fil[HTS_URLMAXSIZE * 2];
-
-            strlcpybuff(n_adr,
-                        norm_host ? jump_normalized_const(moved->adr)
-                                  : jump_identification_const(moved->adr),
-                        sizeof(n_adr));
-            strlcpybuff(pn_adr,
-                        norm_host ? jump_normalized_const(urladr())
-                                  : jump_identification_const(urladr()),
-                        sizeof(pn_adr));
-            fil_normalized_filtered_ex(moved->fil, n_fil, NULL, norm_slash,
-                                       norm_query);
-            fil_normalized_filtered_ex(urlfil(), pn_fil, NULL, norm_slash,
-                                       norm_query);
-            if (strcasecmp(n_adr, pn_adr) == 0
-                && strcasecmp(n_fil, pn_fil) == 0) {
-              hts_log_print(opt, LOG_WARNING,
-                            "Redirected link is identical because of 'URL Hack' option: %s%s and %s%s",
-                            urladr(), urlfil(), moved->adr, moved->fil);
-            }
-          }
+          // A same-file alias redirect must be followed, not stubbed (#159).
+          const hts_boolean same_savefile = hts_redirect_same_savefile(
+              opt, urladr(), urlfil(), moved->adr, moved->fil);
           //if (ident_url_absolute(mov_url,moved->adr,moved->fil)!=-1) {    // ok URL reconnue
           // c'est (en gros) la même URL..
           // si c'est un problème de casse dans le host c'est que le serveur est buggé
@@ -3605,7 +3581,17 @@ int hts_mirror_check_moved(htsmoduleStruct * str,
                 hts_log_print(opt, LOG_DEBUG, "moved link accepted: %s%s",
                               moved->adr, moved->fil);
               }
-            }                   /* sinon traité normalement */
+            } else if (same_savefile) {
+              // A stub would point at itself; follow the redirect instead.
+              if (hts_acceptlink(opt, ptr, moved->adr, moved->fil, NULL, NULL,
+                                 &set_prio_to, NULL) != 1) {
+                get_it = 1;
+                hts_log_print(opt, LOG_WARNING,
+                              "Redirect to a same-file alias, fetching real "
+                              "content: %s%s -> %s%s",
+                              urladr(), urlfil(), moved->adr, moved->fil);
+              }
+            } /* sinon traité normalement */
           }
 
           //if ((strfield2(moved->adr,urladr())!=0) && (strfield2(moved->fil,urlfil())!=0)) {  // identique à casse près
@@ -3628,7 +3614,11 @@ int hts_mirror_check_moved(htsmoduleStruct * str,
                    heap(heap(ptr)->precedent)->adr,
                    heap(heap(ptr)->precedent)->fil, opt,
                    sback, cache, hash, ptr, numero_passe, NULL) != -1) {
-                if (hash_read(hash, savedmoved.save, NULL, HASH_STRUCT_FILENAME) < 0) {   // n'existe pas déja
+                // Same-file alias: the reserved name is the invalidated source,
+                // so record anyway.
+                if (same_savefile ||
+                    hash_read(hash, savedmoved.save, NULL,
+                              HASH_STRUCT_FILENAME) < 0) { // n'existe pas déja
                   // enregistrer lien avec SAV IDENTIQUE
                   if (hts_record_link(opt, moved->adr, moved->fil, heap(ptr)->sav, "", "", NULL)) {
                     // mode test?
@@ -3652,7 +3642,6 @@ int hts_mirror_check_moved(htsmoduleStruct * str,
                                 "moving %s to an existing file %s",
                                 heap(ptr)->fil, urlfil());
                 }
-
               }
             }
 

src/htsparse.h

@@ -116,6 +116,19 @@ int htsparse(htsmoduleStruct * str, htsmoduleStructExtended * stre);
 int hts_mirror_check_moved(htsmoduleStruct * str,
                            htsmoduleStructExtended * stre);
 
+/*
+  Non-zero if a redirect (cur_adr,cur_fil)->(moved_adr,moved_fil) saves to the
+  same local file, so it must be followed rather than turned into a
+  self-pointing "moved" stub (#159). Mirrors the savename: scheme+userinfo
+  stripped, www kept (www dedup is the crawl layer's job), path
+  slash/query-normalized per the URL-hack flags. Not hash_url_equals: that keys
+  on the dedup hash, which folds www and never collapses http<->https.
+*/
+hts_boolean hts_redirect_same_savefile(httrackp *opt, const char *cur_adr,
+                                       const char *cur_fil,
+                                       const char *moved_adr,
+                                       const char *moved_fil);
+
 /*
   Process user intercations: pause, add link, delete link..
 */

src/htsselftest.c

@@ -45,6 +45,7 @@ Please visit our Website: http://www.httrack.com
 #include "htscore.h"
 #include "htsdefines.h"
 #include "htslib.h"
+#include "htsparse.h"
 #include "htscache_selftest.h"
 #include "htsdns_selftest.h"
 #include "htscharset.h"
@@ -1340,6 +1341,37 @@ static int st_urlhack(httrackp *opt, int argc, char **argv) {
   return 0;
 }
 
+/* #159: hts_redirect_same_savefile decides whether a redirect is a same-file
+ * alias. */
+static int st_redirect_samefile(httrackp *opt, int argc, char **argv) {
+  (void) argc;
+  (void) argv;
+#define SAME(aa, fa, ab, fb) hts_redirect_same_savefile(opt, aa, fa, ab, fb)
+  /* scheme and userinfo collapse (the #159 case); a different path does not */
+  assertf(SAME("http://foo.com", "/a/b", "https://foo.com", "/a/b"));
+  assertf(SAME("http://user@foo.com", "/a", "http://foo.com", "/a"));
+  assertf(!SAME("http://foo.com", "/a", "http://foo.com", "/b"));
+  /* www stays distinct here; the crawl's dedup layer folds www, not this helper
+   */
+  opt->urlhack = HTS_TRUE;
+  opt->no_www_dedup = opt->no_slash_dedup = opt->no_query_dedup = HTS_FALSE;
+  assertf(!SAME("http://www.foo.com", "/a", "http://foo.com", "/a"));
+  /* slash/query fold only when the dedup flag is on */
+  assertf(SAME("https://foo.com", "/a//b", "http://foo.com", "/a/b"));
+  assertf(
+      SAME("https://foo.com", "/p?b=2&a=1", "http://foo.com", "/p?a=1&b=2"));
+  opt->no_slash_dedup = opt->no_query_dedup = HTS_TRUE;
+  assertf(!SAME("https://foo.com", "/a//b", "http://foo.com", "/a/b"));
+  assertf(
+      !SAME("https://foo.com", "/p?b=2&a=1", "http://foo.com", "/p?a=1&b=2"));
+  /* but a pure scheme alias still collapses regardless of dedup opt-outs */
+  assertf(SAME("http://foo.com", "/a/b", "https://foo.com", "/a/b"));
+  opt->no_slash_dedup = opt->no_query_dedup = HTS_FALSE;
+#undef SAME
+  printf("redirect-samefile self-test OK\n");
+  return 0;
+}
+
 // hts_finish_makeindex writes the footer, emits the refresh meta only when
 // makeindex_links==1, and clears *fp / sets *done. argv[0] is a writable dir.
 static int st_makeindex(httrackp *opt, int argc, char **argv) {
@@ -1757,6 +1789,8 @@ static const struct selftest_entry {
      st_stripquery},
     {"urlhack", "", "-%u url-hack sub-flag (www/slash/query) self-test",
      st_urlhack},
+    {"redirect-samefile", "", "same-file redirect detection self-test (#159)",
+     st_redirect_samefile},
     {"mime", "<filename>", "MIME type for a filename", st_mime},
     {"charset", "<charset> <string>",
      "convert a string to UTF-8 from a charset", st_charset},

tests/01_engine-redirect.test

@@ -0,0 +1,9 @@
+#!/bin/bash
+#
+
+set -euo pipefail
+
+# #159: a redirect to a same-file alias (http<->https, user@host, ..) must be
+# followed through, not turned into a self-pointing "moved" stub. The decision
+# helper is exercised by the engine self-test.
+httrack -O /dev/null -#test=redirect-samefile run | grep -q "redirect-samefile self-test OK"

tests/30_local-fragment-link.test

@@ -0,0 +1,13 @@
+#!/bin/bash
+# Issue #279: an anchored link (target.html#sec, quoted or bare) fetches the
+# target with the fragment dropped (strict server 400s on a '#' in the request)
+# but keeps it in the rewritten local link so the anchor still works.
+set -e
+
+: "${top_srcdir:=..}"
+
+bash "$top_srcdir/tests/local-crawl.sh" --errors 0 \
+    --found 'fraglink/target.html' \
+    --file-matches 'fraglink/index.html' 'href=target\.html#sec' \
+    --file-matches 'fraglink/index.html' 'href="target\.html#sec2"' \
+    httrack 'BASEURL/fraglink/index.html'

tests/Makefile.am

@@ -6,6 +6,7 @@ EXTRA_DIST = $(TESTS) crawl-test.sh run-all-tests.sh check-network.sh \
 	local-crawl.sh local-server.py server.crt server.key \
 	server-root/simple/basic.html server-root/simple/link.html \
 	server-root/stripquery/index.html server-root/stripquery/a.html \
+	server-root/fraglink/index.html server-root/fraglink/target.html \
 	fixtures/cache-golden/hts-cache/new.zip
 
 TESTS_ENVIRONMENT =
@@ -43,6 +44,7 @@ TESTS = \
 	01_engine-parse.test \
 	01_engine-pause.test \
 	01_engine-rcfile.test \
+	01_engine-redirect.test \
 	01_engine-relative.test \
 	01_engine-robots.test \
 	01_engine-savename.test \
@@ -83,6 +85,7 @@ TESTS = \
 	26_local-strip-query.test \
 	27_local-cookies-file.test \
 	28_local-pause.test \
-	29_local-redirect-fragment.test
+	29_local-redirect-fragment.test \
+	30_local-fragment-link.test
 
 CLEANFILES = check-network_sh.cache

tests/local-crawl.sh

@@ -15,8 +15,11 @@
 #   bash local-crawl.sh [--tls] [--root DIR] [--cookie NAME=VALUE ...] \
 #       --errors N --files N --found PATH ... --directory PATH ... \
 #       --log-found REGEX ... --log-not-found REGEX ... \
+#       --file-matches PATH REGEX ... --file-not-matches PATH REGEX ... \
 #       httrack BASEURL/some/path [httrack-args...]
 # --log-found/--log-not-found grep (ERE) the crawl's hts-log.txt.
+# --file-matches/--file-not-matches grep (ERE) a mirrored file (PATH under the
+# host root), to assert rewritten link/content survived the crawl.
 # --cookie writes a Netscape cookies.txt (scoped to the discovered host:port,
 # which the ephemeral port forces into the cookie domain) and passes it to
 # httrack via --cookies-file, to exercise preloaded cookies.
@@ -121,6 +124,10 @@ while test "$pos" -lt "$nargs"; do
         audit+=("${args[$pos]}" "${args[$((pos + 1))]}")
         pos=$((pos + 1))
         ;;
+    --file-matches | --file-not-matches)
+        audit+=("${args[$pos]}" "${args[$((pos + 1))]}" "${args[$((pos + 2))]}")
+        pos=$((pos + 2))
+        ;;
     httrack)
         pos=$((pos + 1))
         break
@@ -294,6 +301,24 @@ while test "$i" -lt "${#audit[@]}"; do
             exit 1
         else result "OK"; fi
         ;;
+    --file-matches)
+        path="${audit[$((i + 1))]}"
+        i=$((i + 2))
+        info "checking ${path} matches ${audit[$i]}"
+        if grep -aqE "${audit[$i]}" "${hostroot}/${path}"; then result "OK"; else
+            result "no match"
+            exit 1
+        fi
+        ;;
+    --file-not-matches)
+        path="${audit[$((i + 1))]}"
+        i=$((i + 2))
+        info "checking ${path} lacks ${audit[$i]}"
+        if grep -aqE "${audit[$i]}" "${hostroot}/${path}"; then
+            result "matched"
+            exit 1
+        else result "OK"; fi
+        ;;
     esac
     i=$((i + 1))
 done

tests/server-root/fraglink/index.html

@@ -0,0 +1,4 @@
+<html><body>
+<a href=target.html#sec>unquoted fragment link</a>
+<a href="target.html#sec2">quoted fragment link</a>
+</body></html>

tests/server-root/fraglink/target.html

@@ -0,0 +1 @@
+<html><body><a name="sec"></a><a name="sec2"></a>target</body></html>