Keep the vcxproj source list alphabetical

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Signed-off-by: Xavier Roche <roche@httrack.com>

.gitattributes

@@ -1,8 +1,3 @@
 # Resource scripts are Windows tooling input and must stay CRLF: the engine has no
 # encoding guard, so an autocrlf checkout could otherwise flatten them silently.
 *.rc text eol=crlf
-
-# The test scripts run under Git Bash on the Windows CI; a CRLF checkout makes
-# bash die on $'\r' on every line of them.
-*.test text eol=lf
-*.sh text eol=lf

.github/workflows/windows-build.yml

@@ -119,61 +119,22 @@ jobs:
             throw "CRT mismatch across the boundary: libhttrack.dll=$a httrack.exe=$b"
           }
 
-      # The engine ships ~90 tests and none had ever run on Windows: "make check"
+      # The engine self-tests exercise the codecs the DLL was linked with: a
-      # is Linux/macOS only. These are the offline ones, driven from Git Bash
+      # missing brotli/zstd decoder shows up here, not just at link time.
-      # against the native httrack.exe. They subsume the self-tests this step
+      - name: Run the content-coding self-tests
-      # used to run inline (codecs, cache, fsize).
+        shell: pwsh
-      - name: Run the engine test suite (offline tests)
-        shell: bash
-        working-directory: tests
         run: |
-          set -u
+          $ErrorActionPreference = "Stop"
-          bin="$(cygpath -u "$GITHUB_WORKSPACE")/src/${{ matrix.platform }}/${{ matrix.configuration }}"
+          $exe = "src\${{ matrix.platform }}\${{ matrix.configuration }}\httrack.exe"
-          export PATH="$bin:$PATH"
+          $tmp = New-Item -ItemType Directory -Path (Join-Path $env:RUNNER_TEMP "st")
-          command -v httrack >/dev/null || { echo "::error::no httrack.exe in $bin"; exit 1; }
+          foreach ($t in @("acceptencoding", "contentcodings")) {
-
+            $out = & $exe -O NUL "-#test=$t" $tmp.FullName run
-          # httrack.exe is native, so MSYS rewrites any argument shaped like a
+            if ($LASTEXITCODE -ne 0) { throw "$t failed" }
-          # POSIX path, and a URL path is shaped exactly like one: "/a/b.html"
+            if (-not ($out | Select-String -SimpleMatch "$t self-test OK")) {
-          # reached the engine as "C:/Program Files/Git/a/b.html". Switch that
+              throw "$t did not report OK: $out"
-          # off, and hand the tests a TMPDIR that is already a Windows path.
+            }
-          export MSYS_NO_PATHCONV=1
+            Write-Host ($out -join "`n")
-          export MSYS2_ARG_CONV_EXCL='*'
+          }
-          TMPDIR="$(cygpath -m "$RUNNER_TEMP")"
-          export TMPDIR
-
-          pass=0 fail=0 skip=0 failed=""
-          for t in 00_runnable.test 01_engine-*.test 01_zlib-*.test; do
-            rc=0
-            bash "$t" >"$t.log" 2>&1 || rc=$?
-            case "$rc" in
-              0) pass=$((pass + 1)); echo "PASS $t" ;;
-              77) skip=$((skip + 1)); echo "SKIP $t" ;;
-              *)
-                fail=$((fail + 1)) failed="$failed $t"
-                echo "FAIL $t (exit $rc)"
-                # These assert with `test "$(...)" == "..." || exit 1`, which
-                # says nothing at all on failure. Re-run traced.
-                bash -x "$t" >>"$t.log" 2>&1 || true
-                tail -n 25 "$t.log" | sed 's/^/      /'
-                ;;
-            esac
-          done
-          echo "ran=$((pass + fail + skip)) pass=$pass fail=$fail skip=$skip" |
-            tee -a "$GITHUB_STEP_SUMMARY"
-
-          # Every gate in these scripts exits 77, so a suite that degraded to
-          # all-skipped would report green having tested nothing: assert a floor
-          # on what actually ran, not just the absence of failures.
-          [ "$pass" -ge 45 ] || { echo "::error::only $pass tests passed ($skip skipped)"; exit 1; }
-          [ "$fail" -eq 0 ] || { echo "::error::failing:$failed"; exit 1; }
-
-      - name: Upload the test logs
-        if: always()
-        uses: actions/upload-artifact@v4
-        with:
-          name: engine-tests-${{ matrix.platform }}-${{ matrix.configuration }}
-          path: tests/*.log
-          if-no-files-found: warn
 
       - name: Upload MSBuild logs
         if: always()

src/htsback.c

@@ -140,9 +140,8 @@ int back_connect_fallback_due(int addr_index, int addr_count, int elapsed,
 
 /* Pending-connect result for a non-blocking socket reported ready by select():
    0 = connected, >0 = the connect errno (refused, unreachable, ...), -1 if the
-   probe itself failed. A failed connect is reported ready as well (writable on
+   probe itself failed. A failed connect is reported writable too, so this is
-   posix, exception set on winsock), so this is how success is told from failure
+   how success is told from failure without blocking. */
-   without blocking. */
 static int connect_socket_error(T_SOC soc) {
   int soerr = 0;
   socklen_t len = (socklen_t) sizeof(soerr);
@@ -777,7 +776,7 @@ int back_finalize(httrackp * opt, cache_back * cache, struct_back * sback,
         if (back[p].r.statuscode == HTTP_OK) {  // OK (ou 304 en backing)
           if (back[p].r.is_write) {     // Written file
             if (may_be_hypertext_mime(opt, back[p].r.contenttype, back[p].url_fil)) {   // to parse!
-              LLint sz;
+              off_t sz;
 
               sz = fsize_utf8(back[p].url_sav);
               if (sz > 0) {     // ok, exists!
@@ -1954,7 +1953,7 @@ int back_add(struct_back * sback, httrackp * opt, cache_back * cache, const char
       /* Not in cache ; maybe in temporary cache ? Warning: non-movable
          "url_sav" */
       else if (back_unserialize_ref(opt, adr, fil, &itemback) == 0) {
-        const LLint file_size = fsize_utf8(itemback->url_sav);
+        const off_t file_size = fsize_utf8(itemback->url_sav);
 
         /* Found file on disk */
         if (file_size > 0) {
@@ -1989,7 +1988,7 @@ int back_add(struct_back * sback, httrackp * opt, cache_back * cache, const char
       }
       /* Not in cache or temporary cache ; found on disk ? (hack) */
       else if (fexist_utf8(save)) {
-        const LLint sz = fsize_utf8(save);
+        const off_t sz = fsize_utf8(save);
 
         // Bon, là il est possible que le fichier ait été partiellement transféré
         // (s'il l'avait été en totalité il aurait été inscrit dans le cache ET existerait sur disque)
@@ -2659,9 +2658,7 @@ void back_wait(struct_back * sback, httrackp * opt, cache_back * cache,
     for (i_mod = 0; i_mod < (unsigned int) back_max; i_mod++) {
       unsigned int i = (i_mod + mod_random) % (back_max);
 
-      // winsock flags a failed connect in the exception set only: leave a
+      if (back[i].status > 0) {
-      // connecting slot to the connect handler, which can still fall back
-      if (back[i].status > 0 && back[i].status != STATUS_CONNECTING) {
         if (!back[i].r.is_file) {       // not file..
           if (back[i].r.soc != INVALID_SOCKET) {        // hey, you never know..
             int err = FD_ISSET(back[i].r.soc, &fds_e);
@@ -2675,7 +2672,10 @@ void back_wait(struct_back * sback, httrackp * opt, cache_back * cache,
               }
               back[i].r.soc = INVALID_SOCKET;
               back[i].r.statuscode = STATUSCODE_CONNERROR;
-              strcpybuff(back[i].r.msg, "Receive Error");
+              if (back[i].status == STATUS_CONNECTING)
+                strcpybuff(back[i].r.msg, "Connect Error");
+              else
+                strcpybuff(back[i].r.msg, "Receive Error");
               if (back[i].status == STATUS_ALIVE) {     /* Keep-alive socket */
                 back_delete(opt, cache, sback, i);
               } else {
@@ -2708,11 +2708,10 @@ void back_wait(struct_back * sback, httrackp * opt, cache_back * cache,
             gestion_timeout = 1;
 
         // connecté?
-        dispo = back[i].r.soc != INVALID_SOCKET &&
+        dispo = FD_ISSET(back[i].r.soc, &fds_c);
-                (FD_ISSET(back[i].r.soc, &fds_c) ||
-                 FD_ISSET(back[i].r.soc, &fds_e));
         if (dispo) { // socket ready: connect() finished (ok or failed)
-          // probe SO_ERROR and, on failure, fall back to the next address
+          // a refused/failed connect is reported writable too; probe SO_ERROR
+          // and, on failure, fall back to the next address (or fail the slot)
           if (connect_socket_error(back[i].r.soc) != 0) {
             if (!back_connect_next(opt, sback, i)) {
               deletehttp(&back[i].r);
@@ -3696,7 +3695,7 @@ void back_wait(struct_back * sback, httrackp * opt, cache_back * cache,
                       if (back[i].r.statuscode == HTTP_OK && !back[i].testmode) {       // 'OK'
                         if (!is_hypertext_mime(opt, back[i].r.contenttype, back[i].url_fil)) {  // not HTML
                           if (strnotempty(back[i].url_sav)) {   // target found
-                            LLint size = fsize_utf8(back[i].url_sav);
+                            off_t size = fsize_utf8(back[i].url_sav);
 
                             if (size >= 0) {
                               if (back[i].r.totalsize == size) {        // same size!
@@ -3915,7 +3914,7 @@ void back_wait(struct_back * sback, httrackp * opt, cache_back * cache,
                     // traiter 206 (partial content)
                     // xxc SI CHUNK VERIFIER QUE CA MARCHE??
                     if (back[i].r.statuscode == 206) {  // on nous envoie un morceau (la fin) coz une partie sur disque!
-                      LLint sz = fsize_utf8(back[i].url_sav);
+                      off_t sz = fsize_utf8(back[i].url_sav);
                       /* RFC 7233: resume at the server's Content-Range start,
                          not the offset we requested; a server may resume
                          earlier and appending the overlap duplicates bytes

src/htscache.c

@@ -356,7 +356,7 @@ void cache_add(httrackp * opt, cache_back * cache, const htsblk * r,
       FILE *fp;
 
       // On recopie le fichier->.
-      LLint file_size = fsize_utf8(fconv(catbuff, sizeof(catbuff), url_save));
+      off_t file_size = fsize_utf8(fconv(catbuff, sizeof(catbuff), url_save));
 
       if (file_size >= 0) {
         fp = FOPEN(fconv(catbuff, sizeof(catbuff), url_save), "rb");
@@ -1241,27 +1241,23 @@ char *readfile(const char *fil) {
 char *readfile2(const char *fil, LLint * size) {
   char *adr = NULL;
   char catbuff[CATBUFF_SIZE];
-  const LLint len = fsize(fil);
+  INTsys len = 0;
 
-  /* a size too large for size_t (32-bit Windows/i386) must fail closed: it
+  len = fsize(fil);
-     would wrap malloct() short while fread() still read the untruncated len */
+  if (len >= 0) {               // exists
-  const size_t buflen = len >= 0 ? llint_to_size_t(len) : (size_t) -1;
-
-  if (buflen != (size_t) -1) { // exists, and is addressable
     FILE *fp;
 
     fp = fopen(fconv(catbuff, sizeof(catbuff), fil), "rb");
     if (fp != NULL) {           // n'existe pas (!)
-      adr = (char *) malloct(buflen + 1);
+      adr = (char *) malloct(len + 1);
       if (size != NULL)
         *size = len;
       if (adr != NULL) {
-        if (buflen > 0 &&
+        if (len > 0 && fread(adr, 1, len, fp) != len) { // fichier endommagé ?
-            fread(adr, 1, buflen, fp) != buflen) { // fichier endommagé ?
           freet(adr);
           adr = NULL;
         } else
-          *(adr + buflen) = '\0';
+          *(adr + len) = '\0';
       }
       fclose(fp);
     }
@@ -1273,21 +1269,19 @@ char *readfile2(const char *fil, LLint * size) {
 char *readfile_utf8(const char *fil) {
   char *adr = NULL;
   char catbuff[CATBUFF_SIZE];
-  const LLint len = fsize_utf8(fil);
+  const off_t len = fsize_utf8(fil);
-  const size_t buflen = len >= 0 ? llint_to_size_t(len) : (size_t) -1;
 
-  if (buflen != (size_t) -1) { // exists, and is addressable (see readfile2)
+  if (len >= 0) {               // exists
     FILE *const fp = FOPEN(fconv(catbuff, sizeof(catbuff), fil), "rb");
 
     if (fp != NULL) {           // n'existe pas (!)
-      adr = (char *) malloct(buflen + 1);
+      adr = (char *) malloct(len + 1);
       if (adr != NULL) {
-        if (buflen > 0 &&
+        if (len > 0 && fread(adr, 1, len, fp) != len) { // fichier endommagé ?
-            fread(adr, 1, buflen, fp) != buflen) { // fichier endommagé ?
           freet(adr);
           adr = NULL;
         } else {
-          adr[buflen] = '\0';
+          adr[len] = '\0';
         }
       }
       fclose(fp);

src/htscache_selftest.c

@@ -938,9 +938,9 @@ static void reconcile_put(httrackp *opt, const char *name, size_t size) {
 }
 
 /* Expect `name` to weigh `size` bytes, or be absent when size == -1. */
-static int reconcile_expect(httrackp *opt, const char *name, LLint size,
+static int reconcile_expect(httrackp *opt, const char *name, off_t size,
                             const char *what) {
-  const LLint got = fsize(reconcile_st_path(opt, name));
+  const off_t got = fsize(reconcile_st_path(opt, name));
 
   if (got != size) {
     fprintf(stderr, "cache-reconcile: %s: %s is %d bytes, expected %d\n", what,
@@ -954,7 +954,7 @@ int cache_reconcile_selftest(httrackp *opt, const char *dir) {
   int failures = 0;
 
   /* around the interrupted-run thresholds (new < 32768, old > 65536) */
-  static const LLint TINY = 1024, MID = 40000, SOLID = 131072;
+  static const off_t TINY = 1024, MID = 40000, SOLID = 131072;
 
   golden_setup(opt, dir);
 #ifdef _WIN32

src/htscharset.c

@@ -31,9 +31,6 @@ Please visit our Website: http://www.httrack.com
 /* ------------------------------------------------------------ */
 
 #include "htscharset.h"
-#ifdef _WIN32
-#include <shellapi.h>
-#endif
 #include "htsbase.h"
 #include "punycode.h"
 #include "htssafe.h"
@@ -385,44 +382,10 @@ char *hts_convertStringFromUTF8(const char *s, size_t size, const char *charset)
   return hts_convertStringCPFromUTF8(s, size, cp);
 }
 
-HTSEXT_API char *hts_convertStringSystemToUTF8(const char *s, size_t size) {
+char *hts_convertStringSystemToUTF8(const char *s, size_t size) {
   return hts_convertStringCPToUTF8(s, size, GetACP());
 }
 
-HTSEXT_API void hts_argv_utf8(int *pargc, char ***pargv) {
-  int wargc = 0;
-  LPWSTR *const wargv = CommandLineToArgvW(GetCommandLineW(), &wargc);
-  char **argv;
-  int i;
-
-  // On any failure keep the CRT's ANSI argv: lossy, but never half-converted.
-  if (wargv == NULL)
-    return;
-  if (wargc <= 0 ||
-      (argv = calloct((size_t) wargc + 1, sizeof(char *))) == NULL) {
-    LocalFree(wargv);
-    return;
-  }
-  for (i = 0; i < wargc; i++) {
-    const int wsize = (int) wcslen(wargv[i]);
-
-    // hts_convertUCS2StringToUTF8() returns NULL on an empty string
-    argv[i] =
-        wsize != 0 ? hts_convertUCS2StringToUTF8(wargv[i], wsize) : strdupt("");
-    if (argv[i] == NULL) {
-      while (i-- != 0)
-        freet(argv[i]);
-      freet(argv);
-      LocalFree(wargv);
-      return;
-    }
-  }
-  argv[wargc] = NULL; // callers may rely on argv[argc] == NULL
-  LocalFree(wargv);
-  *pargc = wargc;
-  *pargv = argv; // never freed: argv lives for the process
-}
-
 #else
 
 #include <errno.h>

src/htscharset.h

@@ -34,7 +34,6 @@ Please visit our Website: http://www.httrack.com
 #define HTS_CHARSET_DEFH
 
 /** Standard includes. **/
-#include "htsglobal.h"
 #include <stdlib.h>
 #include <string.h>
 #ifdef _WIN32
@@ -175,16 +174,7 @@ extern char *hts_convertUCS2StringToUTF8(LPWSTR woutput, int wsize);
  * Convert current system codepage to UTF-8.
  * This function is WIN32 specific.
  **/
-HTSEXT_API char *hts_convertStringSystemToUTF8(const char *s, size_t size);
+extern char *hts_convertStringSystemToUTF8(const char *s, size_t size);
-
-/**
- * Replace the CRT's ANSI argv by a UTF-8 one decoded from the real UTF-16
- * command line: every char* is UTF-8 on Windows (FOPEN, STAT, ... convert at
- * the syscall boundary). Keeps the CRT's argv on failure; the new array is
- * writable, NULL-terminated, and lives for the process.
- * This function is WIN32 specific.
- **/
-HTSEXT_API void hts_argv_utf8(int *pargc, char ***pargv);
 #endif
 
 #endif

src/htscore.c

@@ -855,7 +855,7 @@ int httpmirror(char *url1, httrackp * opt) {
       char *filelist_buff = NULL;
       size_t filelist_sz = 0;
       const char *filelist_err = NULL; /* failure reason, NULL on success */
-      const LLint fs = fsize(StringBuff(opt->filelist));
+      const off_t fs = fsize(StringBuff(opt->filelist));
 
       if (fs < 0) {
         /* fsize() hides the cause; redo stat() for a precise errno (#49) */
@@ -863,7 +863,7 @@ int httpmirror(char *url1, httrackp * opt) {
         filelist_err = stat(StringBuff(opt->filelist), &st) != 0
                            ? strerror(errno)
                            : "not a regular file";
-      } else if ((filelist_sz = llint_to_size_t(fs)) == (size_t) -1) {
+      } else if ((filelist_sz = off_t_to_size_t(fs)) == (size_t) -1) {
         filelist_err = "file too large";
         filelist_sz = 0;
       } else {
@@ -2089,9 +2089,10 @@ int httpmirror(char *url1, httrackp * opt) {
               (OPT_GET_BUFF(opt), OPT_GET_BUFF_SIZE(opt), StringBuff(opt->path_log),
                "hts-cache/old.lst"), "rb");
       if (old_lst) {
-        const size_t sz = llint_to_size_t(
+        const size_t sz =
-            fsize(fconcat(OPT_GET_BUFF(opt), OPT_GET_BUFF_SIZE(opt),
+          off_t_to_size_t(fsize(fconcat
-                          StringBuff(opt->path_log), "hts-cache/new.lst")));
+                (OPT_GET_BUFF(opt), OPT_GET_BUFF_SIZE(opt), StringBuff(opt->path_log),
+                 "hts-cache/new.lst")));
         new_lst =
           fopen(fconcat
                 (OPT_GET_BUFF(opt), OPT_GET_BUFF_SIZE(opt), StringBuff(opt->path_log),

src/htscoremain.c

@@ -1509,7 +1509,7 @@ static int hts_main_internal(int argc, char **argv, httrackp * opt) {
                   htsmain_free();
                   return -1;
                 } else {
-                  LLint fz;
+                  off_t fz;
 
                   na++;
                   fz = fsize(argv[na]);
@@ -2584,11 +2584,21 @@ static int hts_main_internal(int argc, char **argv, httrackp * opt) {
                 hts_get_version_info(opt), t, url);
         fprintf(opt->log, "(");
         for(i = 0; i < argc; i++) {
-          const char *arg = argv[i]; // already UTF-8 on every platform
+#ifdef _WIN32
+          char *carg =
+            hts_convertStringSystemToUTF8(argv[i], (int) strlen(argv[i]));
+          char *arg = carg != NULL ? carg : argv[i];
+#else
+          const char *arg = argv[i];
+#endif
           if (strchr(arg, ' ') == NULL || strchr(arg, '\"') != NULL)
             fprintf(opt->log, "%s ", arg);
           else                  // entre "" (si espace(s) et pas déja de ")
             fprintf(opt->log, "\"%s\" ", arg);
+#ifdef _WIN32
+          if (carg != NULL)
+            free(carg);
+#endif
         }
         fprintf(opt->log, ")" LF);
         fprintf(opt->log, LF);

src/htsfilters.c

@@ -122,11 +122,8 @@ int fa_strjoker_dual(int type, char **filters, int nfil, const char *nom1,
 }
 
 /* Real filters/URLs fit HTS_URLMAXSIZE*2; past it a hostile pattern recurses
-   O(len) deep or runs O(n^2*stars). Cap length, recursion depth and steps
+   O(len) deep or runs O(n^2*stars). Cap length (depth) and steps (work). */
-   (work): the length cap alone still allows ~2000 frames, ~900KB of stack,
-   which overflows the 1MB a Windows thread gets (#574). */
 #define STRJOKER_MAXLEN (HTS_URLMAXSIZE * 2)
-#define STRJOKER_MAXDEPTH 256u
 #define STRJOKER_MAXSTEPS 2000000u
 
 /* Failure memo for the recursive matcher: one bit per (chaine, joker) offset
@@ -136,28 +133,20 @@ typedef struct strjoker_memo {
   size_t stride;                /* strlen(joker0) + 1 */
   unsigned char *failed;        /* failed-pair bitmap; NULL: no memo */
   size_t *nsteps; /* shared work counter; NULL: unbounded (oracle) */
-  size_t maxdepth; /* deepest recursion reached */
-  hts_boolean cut; /* a branch hit the depth cap: its failures are not final */
 } strjoker_memo;
 
-static const char *strjoker_impl(strjoker_memo *memo, const char *chaine,
+static const char *strjoker_impl(const strjoker_memo *memo, const char *chaine,
-                                 const char *joker, LLint *size, int *size_flag,
+                                 const char *joker, LLint *size,
-                                 size_t depth);
+                                 int *size_flag);
 
 /* A pair that failed once fails forever (*size is only written on the success
    path), so record NULL results and cut the re-exploration. */
-static const char *strjoker_rec(strjoker_memo *memo, const char *chaine,
+static const char *strjoker_rec(const strjoker_memo *memo, const char *chaine,
-                                const char *joker, LLint *size, int *size_flag,
+                                const char *joker, LLint *size,
-                                size_t depth) {
+                                int *size_flag) {
   size_t bit = 0;
   const char *adr;
 
-  if (depth > memo->maxdepth)
-    memo->maxdepth = depth;
-  if (depth >= STRJOKER_MAXDEPTH) {
-    memo->cut = HTS_TRUE;
-    return NULL; /* nesting beyond any real filter: fail the branch safely */
-  }
   if (memo->nsteps != NULL && ++*memo->nsteps > STRJOKER_MAXSTEPS)
     return NULL; /* work budget spent: fail the match safely */
   if (memo->failed) {
@@ -166,9 +155,8 @@ static const char *strjoker_rec(strjoker_memo *memo, const char *chaine,
     if (memo->failed[bit >> 3] & (unsigned char) (1u << (bit & 7)))
       return NULL;
   }
-  adr = strjoker_impl(memo, chaine, joker, size, size_flag, depth + 1);
+  adr = strjoker_impl(memo, chaine, joker, size, size_flag);
-  /* a cut branch may fail here yet match when reached at a shallower depth */
+  if (adr == NULL && memo->failed)
-  if (adr == NULL && memo->failed && !memo->cut)
     memo->failed[bit >> 3] |= (unsigned char) (1u << (bit & 7));
   return adr;
 }
@@ -176,15 +164,13 @@ static const char *strjoker_rec(strjoker_memo *memo, const char *chaine,
 /* Match chaine against joker with a shared work budget *nsteps (a strjokerfind
    sweep passes one counter, bounding the whole scan). */
 static const char *strjoker_bounded(const char *chaine, const char *joker,
-                                    LLint *size, int *size_flag, size_t *nsteps,
+                                    LLint *size, int *size_flag,
-                                    size_t *maxdepth) {
+                                    size_t *nsteps) {
-  strjoker_memo memo = {chaine, joker, 0, NULL, nsteps, 0, HTS_FALSE};
+  strjoker_memo memo = {chaine, joker, 0, NULL, nsteps};
   unsigned char stackbits[2048];
   hts_boolean onheap = HTS_FALSE;
   const char *adr;
 
-  if (maxdepth != NULL)
-    *maxdepth = 0;
   if (chaine != NULL && joker != NULL) {
     const size_t l1 = strlen(chaine), l2 = strlen(joker);
 
@@ -203,11 +189,9 @@ static const char *strjoker_bounded(const char *chaine, const char *joker,
       }
     }
   }
-  adr = strjoker_rec(&memo, chaine, joker, size, size_flag, 0);
+  adr = strjoker_rec(&memo, chaine, joker, size, size_flag);
   if (onheap)
     freet(memo.failed);
-  if (maxdepth != NULL)
-    *maxdepth = memo.maxdepth;
   return adr;
 }
 
@@ -218,39 +202,34 @@ HTS_INLINE const char *strjoker(const char *chaine, const char *joker,
                                 LLint *size, int *size_flag) {
   size_t nsteps = 0;
 
-  return strjoker_bounded(chaine, joker, size, size_flag, &nsteps, NULL);
+  return strjoker_bounded(chaine, joker, size, size_flag, &nsteps);
 }
 
 /* Test-only oracle: the same matcher without the failure memo. */
 const char *strjoker_nomemo(const char *chaine, const char *joker, LLint *size,
                             int *size_flag) {
-  strjoker_memo memo = {chaine, joker, 0, NULL, NULL, 0, HTS_FALSE};
+  const strjoker_memo memo = {chaine, joker, 0, NULL};
 
-  return strjoker_rec(&memo, chaine, joker, size, size_flag, 0);
+  return strjoker_rec(&memo, chaine, joker, size, size_flag);
 }
 
-/* Test-only: strjoker() reporting the work and depth spent and their caps, so a
+/* Test-only: strjoker() reporting the work-budget steps spent and the cap, so a
-   self-test can prove both bound a hostile pattern. */
+   self-test can prove the budget bounds a hostile pattern's work. */
-const char *strjoker_bounds(const char *chaine, const char *joker,
+const char *strjoker_steps(const char *chaine, const char *joker,
-                            size_t *nsteps_out, size_t *maxsteps_out,
+                           size_t *nsteps_out, size_t *maxsteps_out) {
-                            size_t *depth_out, size_t *maxdepth_out) {
+  size_t nsteps = 0;
-  size_t nsteps = 0, depth = 0;
+  const char *r = strjoker_bounded(chaine, joker, NULL, NULL, &nsteps);
-  const char *r = strjoker_bounded(chaine, joker, NULL, NULL, &nsteps, &depth);
 
   if (nsteps_out != NULL)
     *nsteps_out = nsteps;
   if (maxsteps_out != NULL)
     *maxsteps_out = STRJOKER_MAXSTEPS;
-  if (depth_out != NULL)
-    *depth_out = depth;
-  if (maxdepth_out != NULL)
-    *maxdepth_out = STRJOKER_MAXDEPTH;
   return r;
 }
 
-static const char *strjoker_impl(strjoker_memo *memo, const char *chaine,
+static const char *strjoker_impl(const strjoker_memo *memo, const char *chaine,
-                                 const char *joker, LLint *size, int *size_flag,
+                                 const char *joker, LLint *size,
-                                 size_t depth) {
+                                 int *size_flag) {
   if (strnotempty(joker) == 0) {        // fin de chaine joker
     if (strnotempty(chaine) == 0)       // fin aussi pour la chaine: ok
       return chaine;
@@ -424,8 +403,7 @@ static const char *strjoker_impl(strjoker_memo *memo, const char *chaine,
 
       // tester sans le joker (pas ()+ mais ()*)
       if (!unique) {
-        if ((adr = strjoker_rec(memo, chaine, joker + jmp, size, size_flag,
+        if ((adr = strjoker_rec(memo, chaine, joker + jmp, size, size_flag))) {
-                                depth))) {
           return adr;
         }
       }
@@ -438,7 +416,7 @@ static const char *strjoker_impl(strjoker_memo *memo, const char *chaine,
       while(i < (int) max) {
         if (pass[(int) (unsigned char) chaine[i]]) {    // caractère autorisé
           if ((adr = strjoker_rec(memo, chaine + i + 1, joker + jmp, size,
-                                  size_flag, depth))) {
+                                  size_flag))) {
             return adr;
           }
           i++;
@@ -449,7 +427,7 @@ static const char *strjoker_impl(strjoker_memo *memo, const char *chaine,
       // tester chaîne vide
       if (i != max + 2)         // avant c'est ok
         if ((adr = strjoker_rec(memo, chaine + max, joker + jmp, size,
-                                size_flag, depth)))
+                                size_flag)))
           return adr;
 
       return NULL;              // perdu
@@ -471,8 +449,7 @@ static const char *strjoker_impl(strjoker_memo *memo, const char *chaine,
       // comparaison ok?
       if (ok) {
         // continuer la comparaison.
-        if (strjoker_rec(memo, chaine + jmp, joker + jmp, size, size_flag,
+        if (strjoker_rec(memo, chaine + jmp, joker + jmp, size, size_flag))
-                         depth))
           return chaine;        // retourner 1e lettre
       }
 
@@ -494,8 +471,8 @@ const char *strjokerfind(const char *chaine, const char *joker) {
   if (chaine != NULL && strlen(chaine) > STRJOKER_MAXLEN)
     return NULL;
   while(*chaine) {
-    if ((adr = strjoker_bounded(chaine, joker, NULL, NULL, &nsteps,
+    if ((adr = strjoker_bounded(chaine, joker, NULL, NULL,
-                                NULL))) { // ok trouvé
+                                &nsteps))) { // ok trouvé
       return adr;
     }
     if (nsteps > STRJOKER_MAXSTEPS) // scan budget spent: no match

src/htsfilters.h

@@ -52,12 +52,10 @@ HTS_INLINE const char *strjoker(const char *chaine, const char *joker, LLint * s
    oracle for the memoized matcher. */
 const char *strjoker_nomemo(const char *chaine, const char *joker, LLint *size,
                             int *size_flag);
-/* strjoker() reporting the work-budget steps and the recursion depth it spent,
+/* strjoker() reporting the work-budget steps it spent and the cap; test-only,
-   with their caps; test-only, lets a self-test assert both bound a hostile
+   lets a self-test assert the budget bounds a hostile pattern's work. */
-   pattern (depth bounds the stack: an uncapped one overflows Windows' 1MB). */
+const char *strjoker_steps(const char *chaine, const char *joker,
-const char *strjoker_bounds(const char *chaine, const char *joker,
+                           size_t *nsteps_out, size_t *maxsteps_out);
-                            size_t *nsteps_out, size_t *maxsteps_out,
-                            size_t *depth_out, size_t *maxdepth_out);
 const char *strjokerfind(const char *chaine, const char *joker);
 #endif
 

src/htsglobal.h

@@ -322,15 +322,16 @@ typedef int64_t TStamp;
 /* Full printf conversion, '%' included (PRId64 has none): "X: " LLintP. */
 #define LLintP "%" PRId64
 
-/* Integer type for file offsets/sizes passed to the C library; INTsysP is its
+/* Integer type for file offsets/sizes passed to the C library. Widens to
-   printf conversion. FIXME: LFS_FLAG is a configure make variable, never a C
+   LLint (with HTS_FSEEKO for fseeko/ftello) under large-file support, plain
-   macro, so this test is dead and INTsys stays int on POSIX despite large-file
+   int otherwise; INTsysP is its printf conversion. */
-   support (the real macro is HTS_LFS). Widening it there is an installed-header
+#ifdef LFS_FLAG
-   type change, so it is left alone here. */
-#if defined(LFS_FLAG) || defined(_MSC_VER)
 typedef LLint INTsys;
 
 #define INTsysP LLintP
+#ifdef __linux
+#define HTS_FSEEKO
+#endif
 #else
 typedef int INTsys;
 

src/htsindex.c

@@ -318,7 +318,7 @@ void index_finish(const char *indexpath, int mode) {
   char catbuff[CATBUFF_SIZE];
   char **tab;
   char *blk;
-  LLint size = fpsize(fp_tmpproject);
+  off_t size = fpsize(fp_tmpproject);
 
   if (size > 0) {
     if (fp_tmpproject) {

src/htslib.c

@@ -4591,31 +4591,26 @@ int fexist_utf8(const char *s) {
   return 0;
 }
 
-/* File size, -1 if absent */
+/* Taille d'un fichier, -1 si n'existe pas */
 /* Note: NOT utf-8 */
-LLint fsize(const char *s) {
+off_t fsize(const char *s) {
-  STRUCT_STAT st;
+  struct stat st;
 
-  if (!strnotempty(s)) // empty name: error
+  if (!strnotempty(s))          // nom vide: erreur
     return -1;
-  /* _stat64, not stat(): the latter is 32-bit on MSVC */
-#ifdef _WIN32
-  if (_stat64(s, &st) == 0 && S_ISREG(st.st_mode)) {
-#else
   if (stat(s, &st) == 0 && S_ISREG(st.st_mode)) {
-#endif
     return st.st_size;
   } else {
     return -1;
   }
 }
 
-/* File size, -1 if absent */
+/* Taille d'un fichier, -1 si n'existe pas */
 /* Note: utf-8 */
-LLint fsize_utf8(const char *s) {
+off_t fsize_utf8(const char *s) {
   STRUCT_STAT st;
 
-  if (!strnotempty(s)) // empty name: error
+  if (!strnotempty(s))          // nom vide: erreur
     return -1;
   if (STAT(s, &st) == 0 && S_ISREG(st.st_mode)) {
     return st.st_size;
@@ -4624,17 +4619,24 @@ LLint fsize_utf8(const char *s) {
   }
 }
 
-/* fseeko/ftello, never fseek/ftell: ftell returns long, which cannot carry an
+off_t fpsize(FILE * fp) {
-   offset past 2GB on a 32-bit platform and fails with EOVERFLOW there. */
+  off_t oldpos, size;
-LLint fpsize(FILE *fp) {
-  LLint oldpos, size;
 
   if (!fp)
     return -1;
+#ifdef HTS_FSEEKO
   oldpos = ftello(fp);
-  fseeko(fp, 0, SEEK_END);
+#else
+  oldpos = ftell(fp);
+#endif
+  fseek(fp, 0, SEEK_END);
+#ifdef HTS_FSEEKO
   size = ftello(fp);
   fseeko(fp, oldpos, SEEK_SET);
+#else
+  size = ftell(fp);
+  fseek(fp, oldpos, SEEK_SET);
+#endif
   return size;
 }
 
@@ -6398,13 +6400,13 @@ int hts_stat_utf8(const char *path, STRUCT_STAT * buf) {
   LPWSTR wpath = hts_convertUTF8StringToUCS2(path, (int) strlen(path), NULL);
 
   if (wpath != NULL) {
-    const int result = _wstat64(wpath, buf);
+    const int result = _wstat(wpath, buf);
 
     free(wpath);
     return result;
   } else {
     // Fallback on conversion error.
-    return _stat64(path, buf);
+    return _stat(path, buf);
   }
 }
 

src/htslib.h

@@ -310,11 +310,10 @@ void cut_path(char *fullpath, char *path, size_t path_size, char *pname,
 int fexist(const char *s);
 int fexist_utf8(const char *s);
 
-/* File size in bytes, -1 if absent or not a regular file. LLint, not off_t:
+/*LLint fsize(const char* s);    */
-   the latter is a 32-bit long on MSVC, truncating any file of 2GB or more. */
+off_t fpsize(FILE * fp);
-LLint fpsize(FILE *fp);
+off_t fsize(const char *s);
-LLint fsize(const char *s);
+off_t fsize_utf8(const char *s);
-LLint fsize_utf8(const char *s);
 
 // Threads
 typedef void *(*beginthread_type) (void *);
@@ -394,8 +393,7 @@ void *hts_get_callback(t_hts_htmlcheck_callbacks * callbacks,
      HTSEXT_API FILE *hts_fopen_utf8(const char *path, const char *mode);
 
 #define STAT hts_stat_utf8
-     /* _stat64: _stat's st_size is a 32-bit long, even in x64 builds */
+     typedef struct _stat STRUCT_STAT;
-     typedef struct _stat64 STRUCT_STAT;
      HTSEXT_API int hts_stat_utf8(const char *path, STRUCT_STAT * buf);
 
 #define UNLINK hts_unlink_utf8
@@ -593,9 +591,9 @@ HTS_STATIC int compare_mime(httrackp * opt, const char *mime, const char *file,
 #endif
 
 // returns (size_t) -1 upon error
-static HTS_UNUSED size_t llint_to_size_t(LLint o) {
+static HTS_UNUSED size_t off_t_to_size_t(off_t o) {
   const size_t so = (size_t) o;
-  if ((LLint) so == o) {
+  if ((off_t) so == o) {
     return so;
   } else {
     return (size_t) -1;

src/htsselftest.c

@@ -65,7 +65,6 @@ Please visit our Website: http://www.httrack.com
 #include "coucal/coucal.h"
 
 #include <ctype.h>
-#include <errno.h>
 #include <stdint.h>
 #include <stdio.h>
 #include <stdlib.h>
@@ -73,9 +72,6 @@ Please visit our Website: http://www.httrack.com
 #ifndef _WIN32
 #include <sys/socket.h>
 #include <unistd.h>
-#else
-#include <io.h>       /* _get_osfhandle, for the sparse-file hint */
-#include <winioctl.h> /* FSCTL_SET_SPARSE */
 #endif
 
 /* very minimalistic internal tests */
@@ -757,7 +753,7 @@ static int st_filterbounds(httrackp *opt, int argc, char **argv) {
   const size_t subjlen = 2048;
   char *subj = malloct(big + 1);
   char *pat = malloct(2 * stars + 2);
-  size_t steps = 0, maxsteps = 0, depth = 0, maxdepth = 0, i;
+  size_t steps = 0, maxsteps = 0, i;
 
   (void) opt;
   (void) argc;
@@ -778,22 +774,9 @@ static int st_filterbounds(httrackp *opt, int argc, char **argv) {
   subj[subjlen] = '\0';
   /* Budget must fire and hold: steps > cap (deleting the budget zeroes the
      counter that is the enforcement), steps < 10*cap (unbudgeted ~1.26e9). */
-  assertf(strjoker_bounds(subj, pat, &steps, &maxsteps, &depth, &maxdepth) ==
+  assertf(strjoker_steps(subj, pat, &steps, &maxsteps) == NULL);
-          NULL);
   assertf(steps > maxsteps && steps < 10 * maxsteps);
-  /* Depth caps the stack: uncapped this recurses 2046 frames, ~900KB (#574). */
-  assertf(depth == maxdepth);
   assertf(strjokerfind(subj, pat) == NULL);
-  /* Pin the cap from below: 32 segments must still match, so a cap set so low
-     it would break real multi-segment filters (which use far fewer) fails. */
-  for (i = 0; i < 32; i++) {
-    pat[2 * i] = '*';
-    pat[2 * i + 1] = 'a';
-  }
-  pat[64] = '\0';
-  memset(subj, 'a', 32);
-  subj[32] = '\0';
-  assertf(strjoker(subj, pat, NULL, NULL) != NULL);
   freet(pat);
   freet(subj);
   printf("filterbounds: OK\n");
@@ -830,11 +813,7 @@ static int st_mime(httrackp *opt, int argc, char **argv) {
   return 0;
 }
 
-static size_t st_decode_body(const char *arg, char *buf, size_t size);
-
 static int st_charset(httrackp *opt, int argc, char **argv) {
-  char buf[512];
-  size_t len;
   char *s;
 
   (void) opt;
@@ -842,8 +821,7 @@ static int st_charset(httrackp *opt, int argc, char **argv) {
     fprintf(stderr, "charset: needs a charset and a string\n");
     return 1;
   }
-  len = st_decode_body(argv[1], buf, sizeof(buf));
+  s = hts_convertStringToUTF8(argv[1], strlen(argv[1]), argv[0]);
-  s = hts_convertStringToUTF8(buf, len, argv[0]);
   if (s != NULL) {
     printf("%s\n", s);
     freet(s);
@@ -868,22 +846,16 @@ static int st_metacharset(httrackp *opt, int argc, char **argv) {
 }
 
 static int st_isutf8(httrackp *opt, int argc, char **argv) {
-  char buf[512];
-  size_t len;
-
   (void) opt;
   if (argc < 1) {
     fprintf(stderr, "isutf8: needs a string\n");
     return 1;
   }
-  len = st_decode_body(argv[0], buf, sizeof(buf));
+  printf("%d\n", hts_isStringUTF8(argv[0], strlen(argv[0])) ? 1 : 0);
-  printf("%d\n", hts_isStringUTF8(buf, len) ? 1 : 0);
   return 0;
 }
 
 static int st_idna_encode(httrackp *opt, int argc, char **argv) {
-  char buf[512];
-  size_t len;
   char *s;
 
   (void) opt;
@@ -891,8 +863,7 @@ static int st_idna_encode(httrackp *opt, int argc, char **argv) {
     fprintf(stderr, "idna-encode: needs a hostname\n");
     return 1;
   }
-  len = st_decode_body(argv[0], buf, sizeof(buf));
+  s = hts_convertStringUTF8ToIDNA(argv[0], strlen(argv[0]));
-  s = hts_convertStringUTF8ToIDNA(buf, len);
   if (s != NULL) {
     printf("%s\n", s);
     freet(s);
@@ -1054,7 +1025,7 @@ static int st_hashtable(httrackp *opt, int argc, char **argv) {
 
   /* produce random patterns, or read from a file */
   if (sscanf(snum, "%lu", &count) != 1) {
-    const LLint size = fsize(snum);
+    const off_t size = fsize(snum);
     FILE *fp = fopen(snum, "rb");
     if (fp != NULL) {
       buff = malloct(size);
@@ -1154,10 +1125,7 @@ static int st_strsafe(httrackp *opt, int argc, char **argv) {
        comes from argv so its length is opaque to the compiler (no static
        -Wstringop-overflow, genuine runtime check). "overflow-buff" exercises
        htsbuff. */
-    /* Not sizeof(char*): on ILP32 a char[4] equals the pointer size, and the
+    char small[4];
-       MSVC array-vs-pointer heuristic (sizeof(A) != sizeof(char*)) then reads
-       it as a pointer and silently skips the bound. */
-    char small[6];
     const char *const src = (argc >= 2) ? argv[1] : "overflowing";
 
     if (strcmp(argv[0], "overflow-buff") == 0) {
@@ -1641,8 +1609,7 @@ static int st_crange(httrackp *opt, int argc, char **argv) {
   return 0;
 }
 
-/* Decode an argument ("hex:FFD8.." or literal text) into buf. Raw non-UTF-8
+/* Decode a body argument ("hex:FFD8.." or literal text) into buf. */
-   bytes cannot survive a Windows command line, so hostile inputs go as hex. */
 static size_t st_decode_body(const char *arg, char *buf, size_t size) {
   size_t n = 0;
 
@@ -1682,94 +1649,6 @@ static int st_sniff(httrackp *opt, int argc, char **argv) {
   return 0;
 }
 
-/* fsize()/fsize_utf8()/fpsize() must report a size past 4GB: 32-bit wraps both
-   ways there (MSVC's off_t and struct _stat st_size are long, 32-bit even on
-   x64), and a size under 4GB would survive an *unsigned* 32-bit truncation. */
-static int st_fsize(httrackp *opt, int argc, char **argv) {
-  const LLint expected = 5 * 1024 * 1024 * 1024LL;
-  char BIGSTK path[HTS_URLMAXSIZE * 2];
-  char BIGSTK absent[HTS_URLMAXSIZE * 2];
-  /* both variants: only fsize() feeds the >2GB readers (file://, --list) */
-  const int width = (int) sizeof(fsize(""));
-  const int width_utf8 = (int) sizeof(fsize_utf8(""));
-  FILE *fp;
-  LLint got, got_utf8, gotp, gone;
-  int rc = 0;
-
-  (void) opt;
-  if (argc < 1) {
-    fprintf(stderr, "fsize: needs a directory\n");
-    return 1;
-  }
-  concat(path, sizeof(path), argv[0], "/sparse-5g.bin");
-  concat(absent, sizeof(absent), argv[0], "/no-such-file.bin");
-
-  /* sparse: seek past 4GB and write the last byte */
-  fp = FOPEN(path, "wb");
-  if (fp == NULL) {
-    fprintf(stderr, "fsize: cannot create '%s': %s\n", path, strerror(errno));
-    return 1;
-  }
-#ifdef _WIN32
-  {
-    /* NTFS allocates the hole unless asked not to; POSIX gives it for free.
-       Best-effort: a non-sparse file still measures the same, just costs 5GB.
-     */
-    HANDLE h = (HANDLE) _get_osfhandle(_fileno(fp));
-    DWORD ret;
-
-    if (h != INVALID_HANDLE_VALUE)
-      (void) DeviceIoControl(h, FSCTL_SET_SPARSE, NULL, 0, NULL, 0, &ret, NULL);
-  }
-#endif
-  if (fseeko(fp, expected - 1, SEEK_SET) != 0 || fputc(0, fp) == EOF ||
-      fclose(fp) != 0) {
-    fprintf(stderr, "fsize: cannot extend '%s' to " LLintP ": %s\n", path,
-            expected, strerror(errno));
-    UNLINK(path);
-    return 1;
-  }
-
-  got = fsize(path);
-  got_utf8 = fsize_utf8(path);
-  fp = FOPEN(path, "rb");
-  if (fp == NULL) {
-    fprintf(stderr, "fsize: cannot reopen '%s': %s\n", path, strerror(errno));
-    gotp = -1;
-  } else {
-    gotp = fpsize(fp);
-    fclose(fp);
-  }
-  UNLINK(path);
-  gone = fsize(absent); /* contract: -1, not 0, when absent */
-
-  printf("fsize: width=%d,%d size=" LLintP "," LLintP " psize=" LLintP
-         " absent=" LLintP "\n",
-         width, width_utf8, got, got_utf8, gotp, gone);
-  if (width != 8 || width_utf8 != 8) {
-    fprintf(stderr, "fsize: return types are %d/%d bytes, expected 8\n", width,
-            width_utf8);
-    rc = 1;
-  }
-  if (got != expected || got_utf8 != expected) {
-    fprintf(stderr,
-            "fsize: fsize/fsize_utf8 are " LLintP "/" LLintP
-            ", expected " LLintP "\n",
-            got, got_utf8, expected);
-    rc = 1;
-  }
-  if (gotp != expected) {
-    fprintf(stderr, "fsize: fpsize is " LLintP ", expected " LLintP "\n", gotp,
-            expected);
-    rc = 1;
-  }
-  if (gone != -1) {
-    fprintf(stderr, "fsize: absent file is " LLintP ", expected -1\n", gone);
-    rc = 1;
-  }
-  return rc;
-}
-
 static int st_savename(httrackp *opt, int argc, char **argv) {
   lien_adrfilsave afs;
   cache_back cache;
@@ -2916,43 +2795,11 @@ static int st_robots(httrackp *opt, int argc, char **argv) {
   return 0;
 }
 
-/* Connected stream pair over loopback; Windows has no socketpair(). */
-static int st_socketpair(T_SOC sv[2]) {
-  struct sockaddr_in sa;
-  socklen_t len = sizeof(sa);
-  T_SOC srv, cli, acc;
-
-  memset(&sa, 0, sizeof(sa));
-  sa.sin_family = AF_INET;
-  sa.sin_addr.s_addr = htonl(INADDR_LOOPBACK);
-  if ((srv = (T_SOC) socket(AF_INET, SOCK_STREAM, 0)) == INVALID_SOCKET)
-    return -1;
-  if (bind(srv, (struct sockaddr *) &sa, sizeof(sa)) != 0 ||
-      listen(srv, 1) != 0 ||
-      getsockname(srv, (struct sockaddr *) &sa, &len) != 0) {
-    deletesoc(srv);
-    return -1;
-  }
-  if ((cli = (T_SOC) socket(AF_INET, SOCK_STREAM, 0)) == INVALID_SOCKET) {
-    deletesoc(srv);
-    return -1;
-  }
-  if (connect(cli, (struct sockaddr *) &sa, sizeof(sa)) != 0 ||
-      (acc = (T_SOC) accept(srv, NULL, NULL)) == INVALID_SOCKET) {
-    deletesoc(cli);
-    deletesoc(srv);
-    return -1;
-  }
-  deletesoc(srv);
-  sv[0] = acc;
-  sv[1] = cli;
-  return 0;
-}
-
 /* get_ftp_line must bound a hostile, CRLF-less reply into its internal
    1024-byte buffer; ASan turns the pre-fix overflow into an abort here. */
+#ifndef _WIN32
 static int st_ftpline(httrackp *opt, int argc, char **argv) {
-  T_SOC sv[2];
+  int sv[2];
   char line[2048];
   char flood[4096];
 
@@ -2960,19 +2807,19 @@ static int st_ftpline(httrackp *opt, int argc, char **argv) {
   (void) argc;
   (void) argv;
   memset(flood, 'x', sizeof(flood));
-  assertf(st_socketpair(sv) == 0);
+  assertf(socketpair(AF_UNIX, SOCK_STREAM, 0, sv) == 0);
-  // the 4102-byte reply fits the loopback send buffer, so no reader is needed
+  assertf(write(sv[1], "220 ", 4) == 4); // valid 3-digit code
-  assertf(send(sv[1], "220 ", 4, 0) == 4); // valid 3-digit code
+  assertf(write(sv[1], flood, sizeof(flood)) == (ssize_t) sizeof(flood));
-  assertf(send(sv[1], flood, (int) sizeof(flood), 0) == (int) sizeof(flood));
+  assertf(write(sv[1], "\r\n", 2) == 2); // end the line so we return
-  assertf(send(sv[1], "\r\n", 2, 0) == 2); // end the line so we return
+  close(sv[1]);
-  deletesoc(sv[1]);
   line[0] = '\0';
   get_ftp_line(sv[0], line, sizeof(line), 5);
-  deletesoc(sv[0]);
+  close(sv[0]);
   printf("ftp-line self-test OK (bounded %d-byte reply)\n",
          (int) sizeof(flood));
   return 0;
 }
+#endif
 
 /* ftp_split_userpass: well-formed split, plus a hostile over-long userinfo
    that pre-fix overran user[256]/pass[256]. */
@@ -3044,11 +2891,11 @@ static const struct selftest_entry {
     {"redirect-samefile", "", "same-file redirect detection self-test (#159)",
      st_redirect_samefile},
     {"mime", "<filename>", "MIME type for a filename", st_mime},
-    {"charset", "<charset> <hex:..|string>",
+    {"charset", "<charset> <string>",
      "convert a string to UTF-8 from a charset", st_charset},
     {"metacharset", "<html>", "extract the <meta> charset from an HTML page",
      st_metacharset},
-    {"isutf8", "<hex:..|string>", "is the string valid UTF-8 (1/0)", st_isutf8},
+    {"isutf8", "<string>", "is the string valid UTF-8 (1/0)", st_isutf8},
     {"idna-encode", "<host>", "encode a hostname to IDNA/punycode",
      st_idna_encode},
     {"idna-decode", "<host>", "decode an IDNA/punycode hostname",
@@ -3085,7 +2932,6 @@ static const struct selftest_entry {
      "local save-name for a URL", st_savename},
     {"sniff", "<content-type> <hex:..|text>", "MIME magic consistency",
      st_sniff},
-    {"fsize", "<dir>", "file size past the 2GB signed-32-bit wrap", st_fsize},
     {"cache", "<dir>", "cache read/write round-trip self-test", st_cache},
     {"cacheindex", "", "cache-index (.ndx) parse must stay in bounds",
      st_cacheindex},
@@ -3116,8 +2962,10 @@ static const struct selftest_entry {
      st_contentcodings},
     {"robots", "", "robots.txt RFC 9309 Allow/Disallow precedence self-test",
      st_robots},
+#ifndef _WIN32
     {"ftp-line", "", "get_ftp_line bounds a hostile FTP reply line",
      st_ftpline},
+#endif
     {"ftp-userpass", "", "ftp_split_userpass bounds URL userinfo", st_ftpuser},
 };
 

src/httrack-library.h

@@ -743,8 +743,7 @@ HTSEXT_API hts_boolean hts_findissystem(find_handle find);
 HTSEXT_API FILE *hts_fopen_utf8(const char *path, const char *mode);
 
 #define STAT hts_stat_utf8
-/* _stat64: _stat's st_size is a 32-bit long, even in x64 builds */
+typedef struct _stat STRUCT_STAT;
-typedef struct _stat64 STRUCT_STAT;
 
 HTSEXT_API int hts_stat_utf8(const char *path, STRUCT_STAT *buf);
 

src/httrack.c

@@ -45,7 +45,6 @@ Please visit our Website: http://www.httrack.com
 #include "htsdefines.h"
 #include "httrack.h"
 #include "htslib.h"
-#include "htscharset.h" // after htslib.h: winsock2.h must precede windows.h
 
 /* Static definitions */
 static int fexist(const char *s);
@@ -207,7 +206,6 @@ int main(int argc, char **argv) {
   httrackp *opt;
 
 #ifdef _WIN32
-  hts_argv_utf8(&argc, &argv);
   {
     WORD wVersionRequested;     // requested version WinSock API
     WSADATA wsadata;            // Windows Sockets API data

src/libhttrack.vcxproj

@@ -70,7 +70,7 @@
     <Link>
       <!-- vcpkg auto-links libssl/libcrypto/zlib/brotli/zstd; only the OS import
            lib is explicit. -->
-      <AdditionalDependencies>Ws2_32.lib;Shell32.lib;%(AdditionalDependencies)</AdditionalDependencies>
+      <AdditionalDependencies>Ws2_32.lib;%(AdditionalDependencies)</AdditionalDependencies>
       <GenerateDebugInformation>true</GenerateDebugInformation>
     </Link>
   </ItemDefinitionGroup>

src/proxy/store.c

@@ -577,11 +577,9 @@ PT_Index PT_LoadCache(const char *filename) {
           chain = coucal_enum_next(&en);
         }
         if (chain != NULL) {
-          const char *scheme = link_has_authority(chain->name) ? "" : "http://";
+          if (!link_has_authority(chain->name))
-
+            strcat(index->slots.common.startUrl, "http://");
-          snprintf(index->slots.common.startUrl,
+          strcat(index->slots.common.startUrl, chain->name);
-                   sizeof(index->slots.common.startUrl), "%s%s", scheme,
-                   (const char *) chain->name);
         }
       }
     }
@@ -634,7 +632,7 @@ int PT_EnumCache(PT_Indexes indexes,
       const long int index_id = (long int) chain->value.intg;
       const char *const url = chain->name;
 
-      if (index_id >= 0 && index_id < indexes->index_size) {
+      if (index_id >= 0 && index_id <= indexes->index_size) {
         PT_Element item =
           PT_ReadCache(indexes->index[index_id], url,
                        FETCH_HEADERS | FETCH_BODY);
@@ -759,7 +757,7 @@ PT_Element PT_ReadIndex(PT_Indexes indexes, const char *url, int flags) {
     if (strncmp(url, "http://", 7) == 0)
       url += 7;
     if (coucal_read(indexes->cil, url, &index_id)) {
-      if (index_id >= 0 && index_id < indexes->index_size) {
+      if (index_id >= 0 && index_id <= indexes->index_size) {
         PT_Element item = PT_ReadCache(indexes->index[index_id], url, flags);
 
         if (item != NULL) {
@@ -782,7 +780,7 @@ int PT_LookupIndex(PT_Indexes indexes, const char *url) {
     if (strncmp(url, "http://", 7) == 0)
       url += 7;
     if (coucal_read(indexes->cil, url, &index_id)) {
-      if (index_id >= 0 && index_id < indexes->index_size) {
+      if (index_id >= 0 && index_id <= indexes->index_size) {
         return 1;
       } else {
         proxytrack_print_log(CRITICAL,
@@ -969,12 +967,10 @@ int PT_LoadCache__New(PT_Index index_, const char *filename) {
                 /* First link as starting URL */
                 if (!firstSeen) {
                   if (strstr(filenameIndex, "/robots.txt") == NULL) {
-                    const char *scheme =
-                        link_has_authority(filenameIndex) ? "" : "http://";
-
                     firstSeen = 1;
-                    snprintf(index->startUrl, sizeof(index->startUrl), "%s%s",
+                    if (!link_has_authority(filenameIndex))
-                             scheme, filenameIndex);
+                      strcat(index->startUrl, "http://");
+                    strcat(index->startUrl, filenameIndex);
                   }
                 }
               } else {
@@ -1106,8 +1102,7 @@ static PT_Element PT_ReadCache__New_u(PT_Index index_, const char *url,
                                           && (!isalpha(previous_save_[0]) || previous_save_[1] != ':')) // c:/home/foo/bar.gif
               ) {
               index->safeCache = 1;
-              snprintf(previous_save, sizeof(previous_save), "%s%s",
+              sprintf(previous_save, "%s%s", index->path, previous_save_);
-                       index->path, previous_save_);
             }
             // bogus format (includes buggy absolute path)
             else {
@@ -1132,16 +1127,15 @@ static PT_Element PT_ReadCache__New_u(PT_Index index_, const char *url,
                 int saveLen = (int) strlen(previous_save_);
 
                 if (index->fixedPath < saveLen) {
-                  snprintf(previous_save, sizeof(previous_save), "%s%s",
+                  sprintf(previous_save, "%s%s", index->path,
-                           index->path, previous_save_ + index->fixedPath);
+                          previous_save_ + index->fixedPath);
                 } else {
                   snprintf(r->msg, sizeof(r->msg), "Bogus fixePath prefix for %s (prefixLen=%d)",
                           previous_save_, (int) index->fixedPath);
                   r->statuscode = STATUSCODE_INVALID;
                 }
               } else {
-                snprintf(previous_save, sizeof(previous_save), "%s%s",
+                sprintf(previous_save, "%s%s", index->path, previous_save_);
-                         index->path, previous_save_);
               }
             }
           }
@@ -1354,71 +1348,45 @@ static int PT_SaveCache__New(PT_Indexes indexes, const char *filename) {
 /* Old HTTrack cache (dat/ndx) format                           */
 /* ------------------------------------------------------------ */
 
-static int cache_brstr(char *adr, char *s, size_t s_size) {
+static int cache_brstr(char *adr, char *s) {
   int i;
   int off;
-  char buff[256 + 4];
+  char buff[256 + 1];
 
   off = binput(adr, buff, 256);
-  /* binput stops at adr's terminating NUL; a value only follows a real line
+  adr += off;
-     terminator, so never step past end-of-buffer. */
+  sscanf(buff, "%d", &i);
-  if (adr[off - 1] == '\0') {
+  if (i > 0)
-    s[0] = '\0';
+    strncpy(s, adr, i);
-    return off - 1;
+  *(s + i) = '\0';
-  }
+  off += i;
-  /* an empty/non-numeric field leaves i unset: treat as length 0 */
-  if (sscanf(buff, "%d", &i) != 1 || i < 0 || i > 32768)
-    i = 0;
-  if (i > 0) {
-    /* a corrupt cache may declare a length past the buffer end; bound the copy
-       and the advance to the bytes actually present */
-    const size_t avail = strnlen(adr + off, (size_t) i);
-    const size_t store = avail < s_size ? avail : s_size - 1;
-
-    memcpy(s, adr + off, store);
-    s[store] = '\0';
-    off += (int) avail;
-  } else {
-    s[0] = '\0';
-  }
   return off;
 }
 
-static void cache_rstr(FILE *fp, char *s, size_t s_size) {
+static void cache_rstr(FILE * fp, char *s) {
-  INTsys i = 0;
+  INTsys i;
   char buff[256 + 4];
 
   linput(fp, buff, 256);
-  /* an unmatched sscanf leaves i untouched: length 0, not a stack value */
+  sscanf(buff, INTsysP, &i);
-  if (sscanf(buff, INTsysP, &i) != 1 || i < 0 || i > 32768)
+  if (i < 0 || i > 32768)       /* error, something nasty happened */
     i = 0;
   if (i > 0) {
-    /* store at most s_size-1 bytes, but consume all i bytes so the next field
+    if ((int) fread(s, 1, i, fp) != i) {
-       stays aligned when a tampered cache declares more than the destination
-       holds */
-    const size_t want = (size_t) i;
-    const size_t store = want < s_size ? want : s_size - 1;
-
-    if (fread(s, 1, store, fp) != store) {
       assertf(! "fread_cache_failed");
     }
-    if (want > store && fseek(fp, (long) (want - store), SEEK_CUR) != 0) {
-      assertf(!"fseek_cache_failed");
-    }
-    s[store] = '\0';
-  } else {
-    s[0] = '\0';
   }
+  *(s + i) = '\0';
 }
 
 static char *cache_rstr_addr(FILE * fp) {
-  INTsys i = 0;
+  INTsys i;
   char *addr = NULL;
   char buff[256 + 4];
 
   linput(fp, buff, 256);
-  /* an unmatched sscanf leaves i untouched: length 0, not a stack value */
+  sscanf(buff, "%d", &i);
-  if (sscanf(buff, INTsysP, &i) != 1 || i < 0 || i > 32768)
+  if (i < 0 || i > 32768)       /* error, something nasty happened */
     i = 0;
   if (i > 0) {
     addr = malloc(i + 1);
@@ -1432,31 +1400,19 @@ static char *cache_rstr_addr(FILE * fp) {
   return addr;
 }
 
-/* Allocate n+1 bytes for an n-byte cache body, rejecting a hostile/corrupt
-   length that would demand an absurd allocation or wrap n+1. NULL on refuse. */
-static char *cache_alloc_body(size_t n) {
-  if (n >= (size_t) INT32_MAX)
-    return NULL;
-  return (char *) malloc(n + 1);
-}
-
 static void cache_rint(FILE * fp, int *i) {
   char s[256];
 
-  cache_rstr(fp, s, sizeof(s));
+  cache_rstr(fp, s);
-  if (sscanf(s, "%d", i) != 1)
+  sscanf(s, "%d", i);
-    *i = 0;
 }
 
 static void cache_rLLint(FILE * fp, unsigned long *i) {
   int l;
   char s[256];
 
-  cache_rstr(fp, s, sizeof(s));
+  cache_rstr(fp, s);
-  /* reject a non-numeric or negative field: a negative length cast to unsigned
+  sscanf(s, "%d", &l);
-     would wrap a later malloc(size+1) to a tiny buffer */
-  if (sscanf(s, "%d", &l) != 1 || l < 0)
-    l = 0;
   *i = (unsigned long) l;
 }
 
@@ -1516,12 +1472,12 @@ static int PT_LoadCache__Old(PT_Index index_, const char *filename) {
         char *a = use;
 
         use[ndxSize] = '\0';
-        a += cache_brstr(a, firstline, sizeof(firstline));
+        a += cache_brstr(a, firstline);
         if (strncmp(firstline, "CACHE-", 6) == 0) {     // Nouvelle version du cache
           if (strncmp(firstline, "CACHE-1.", 8) == 0) { // Version 1.1x
             cache->version = (int) (firstline[8] - '0');        // cache 1.x
             if (cache->version <= 5) {
-              a += cache_brstr(a, firstline, sizeof(firstline));
+              a += cache_brstr(a, firstline);
               strcpy(cache->lastmodified, firstline);
             } else {
               fclose(cache->dat);
@@ -1567,12 +1523,11 @@ static int PT_LoadCache__Old(PT_Index index_, const char *filename) {
               if (!firstSeen) {
                 if (strstr(line, "/robots.txt") == NULL) {
                   PT_Index__Old index = cache;
-                  const char *scheme =
-                      link_has_authority(line) ? "" : "http://";
 
                   firstSeen = 1;
-                  snprintf(index->startUrl, sizeof(index->startUrl), "%s%s",
+                  if (!link_has_authority(line))
-                           scheme, line);
+                    strcat(index->startUrl, "http://");
+                  strcat(index->startUrl, line);
                 }
               }
 
@@ -1668,13 +1623,9 @@ static PT_Element PT_ReadCache__Old_u(PT_Index index_, const char *url,
           String urlDecoded;
 
           r->statuscode = old_r.statuscode;
-          r->size = old_r.size > 0 ? (size_t) old_r.size : 0; // taille fichier
+          r->size = old_r.size; // taille fichier
-          /* these fixed fields come straight from fread and carry no guaranteed
+          strcpy(r->msg, old_r.msg);
-             NUL; terminate them before the bounded copy */
+          strcpy(r->contenttype, old_r.contenttype);
-          old_r.msg[sizeof(old_r.msg) - 1] = '\0';
-          old_r.contenttype[sizeof(old_r.contenttype) - 1] = '\0';
-          strcpybuff(r->msg, old_r.msg);
-          strcpybuff(r->contenttype, old_r.contenttype);
 
           /* Guess the destination filename.. this sucks, because this method is not reliable.
              Yes, the old 1.0 cache format was *that* bogus. /rx */
@@ -1715,27 +1666,26 @@ static PT_Element PT_ReadCache__Old_u(PT_Index index_, const char *url,
         cache_rint(cache->dat, &r->statuscode);
         cache_rLLint(cache->dat, &size_);
         r->size = (size_t) size_;
-        cache_rstr(cache->dat, r->msg, sizeof(r->msg));
+        cache_rstr(cache->dat, r->msg);
-        cache_rstr(cache->dat, r->contenttype, sizeof(r->contenttype));
+        cache_rstr(cache->dat, r->contenttype);
         if (cache->version >= 3)
-          cache_rstr(cache->dat, r->charset, sizeof(r->charset));
+          cache_rstr(cache->dat, r->charset);
-        cache_rstr(cache->dat, r->lastmodified, sizeof(r->lastmodified));
+        cache_rstr(cache->dat, r->lastmodified);
-        cache_rstr(cache->dat, r->etag, sizeof(r->etag));
+        cache_rstr(cache->dat, r->etag);
-        cache_rstr(cache->dat, r->location, sizeof(location_default));
+        cache_rstr(cache->dat, r->location);
         if (cache->version >= 2)
-          cache_rstr(cache->dat, r->cdispo, sizeof(r->cdispo));
+          cache_rstr(cache->dat, r->cdispo);
         if (cache->version >= 4) {
-          cache_rstr(cache->dat, previous_save_, sizeof(previous_save_)); // adr
+          cache_rstr(cache->dat, previous_save_);       // adr
-          cache_rstr(cache->dat, previous_save_, sizeof(previous_save_)); // fil
+          cache_rstr(cache->dat, previous_save_);       // fil
           previous_save[0] = '\0';
-          cache_rstr(cache->dat, previous_save_,
+          cache_rstr(cache->dat, previous_save_);       // save
-                     sizeof(previous_save_)); // save
         }
         if (cache->version >= 5) {
           r->headers = cache_rstr_addr(cache->dat);
         }
         //
-        cache_rstr(cache->dat, check, sizeof(check));
+        cache_rstr(cache->dat, check);
         if (strcmp(check, "HTS") == 0) {        /* intégrité OK */
           ok = 1;
         }
@@ -1764,8 +1714,7 @@ static PT_Element PT_ReadCache__Old_u(PT_Index index_, const char *url,
                                         && (!isalpha(previous_save_[0]) || previous_save_[1] != ':'))   // c:/home/foo/bar.gif
             ) {
             index->safeCache = 1;
-            snprintf(previous_save, sizeof(previous_save), "%s%s", index->path,
+            sprintf(previous_save, "%s%s", index->path, previous_save_);
-                     previous_save_);
           }
           // bogus format (includes buggy absolute path)
           else {
@@ -1790,16 +1739,15 @@ static PT_Element PT_ReadCache__Old_u(PT_Index index_, const char *url,
               int saveLen = (int) strlen(previous_save_);
 
               if (index->fixedPath < saveLen) {
-                snprintf(previous_save, sizeof(previous_save), "%s%s",
+                sprintf(previous_save, "%s%s", index->path,
-                         index->path, previous_save_ + index->fixedPath);
+                        previous_save_ + index->fixedPath);
               } else {
                 snprintf(r->msg, sizeof(r->msg), "Bogus fixePath prefix for %s (prefixLen=%d)",
                         previous_save_, (int) index->fixedPath);
                 r->statuscode = STATUSCODE_INVALID;
               }
             } else {
-              snprintf(previous_save, sizeof(previous_save), "%s%s",
+              sprintf(previous_save, "%s%s", index->path, previous_save_);
-                       index->path, previous_save_);
             }
           }
         }
@@ -1816,7 +1764,7 @@ static PT_Element PT_ReadCache__Old_u(PT_Index index_, const char *url,
               FILE *fp = fopen(previous_save, "rb");
 
               if (fp != NULL) {
-                r->adr = cache_alloc_body(r->size);
+                r->adr = (char *) malloc(r->size + 1);
                 if (r->adr != NULL) {
                   if (r->size > 0 && fread(r->adr, 1, r->size, fp) != r->size) {
                     r->statuscode = STATUSCODE_INVALID;
@@ -1836,7 +1784,7 @@ static PT_Element PT_ReadCache__Old_u(PT_Index index_, const char *url,
           } else {
             // lire fichier (d'un coup)
             if (flags & FETCH_BODY) {
-              r->adr = cache_alloc_body(r->size);
+              r->adr = (char *) malloc(r->size + 1);
               if (r->adr != NULL) {
                 if (fread(r->adr, 1, r->size, cache->dat) != r->size) { // erreur
                   free(r->adr);

tests/01_engine-charset.test

@@ -23,7 +23,7 @@ conv 'iso-8859-1' 'café' 'café'
 
 # windows-1252 is NOT latin-1: 0x80 is the euro sign, not U+0080. This is the
 # case that actually exercises the cp1252 table (the 0x80-0x9F range).
-conv 'windows-1252' 'hex:80' '€'
+conv 'windows-1252' $'\x80' '€'
 
 # pure ASCII is charset-invariant
 conv 'us-ascii' 'hello' 'hello'
@@ -34,8 +34,8 @@ conv 'no-such-charset-xyz' 'abc' 'abc'
 test "$(httrack -O /dev/null -#test=charset 'no-such-charset-xyz' 'café' 2>/dev/null)" == "" || exit 1
 
 # malformed UTF-8 (lone continuation byte, truncated lead byte) must not crash
-runs 'utf-8' 'hex:80'
+runs 'utf-8' $'\x80'
-runs 'utf-8' 'hex:c3'
+runs 'utf-8' $'\xc3'
 
 # hts_getCharsetFromMeta: <meta> charset extraction, HTML5 and legacy forms.
 # -#test=metacharset <html> prints the charset, or "(none)".
@@ -89,11 +89,11 @@ utf8() {
 utf8 'plain' '1'
 utf8 'café' '1'
 utf8 '统计' '1'
-utf8 'hex:636166e9' '0'   # latin-1 accent
+utf8 "$(printf 'caf\xe9')" '0'              # latin-1 accent
-utf8 'hex:c3' '0'         # truncated lead byte
+utf8 "$(printf '\xc3')" '0'                 # truncated lead byte
-utf8 'hex:a9' '0'         # lone continuation byte
+utf8 "$(printf '\xa9')" '0'                 # lone continuation byte
-utf8 'hex:f09f9880' '1'   # 4-byte emoji
+utf8 "$(printf '\xf0\x9f\x98\x80')" '1'     # 4-byte emoji
-utf8 'hex:c0af' '0'       # overlong (latin-1 "À¯" must stay convertible)
+utf8 "$(printf '\xc0\xaf')" '0'             # overlong (latin-1 "À¯" must stay convertible)
-utf8 'hex:eda080' '0'     # UTF-16 surrogate
+utf8 "$(printf '\xed\xa0\x80')" '0'         # UTF-16 surrogate
-utf8 'hex:f4908080' '0'   # above U+10FFFF
+utf8 "$(printf '\xf4\x90\x80\x80')" '0'     # above U+10FFFF
-utf8 'hex:f888808080' '0' # legacy 5-byte form
+utf8 "$(printf '\xf8\x88\x80\x80\x80')" '0' # legacy 5-byte form

tests/01_engine-filelist.test

@@ -52,15 +52,14 @@ run "$tmp/dir" "$tmp/adir"
 loghas 'Could not include URL list "[^"]+": not a regular file'
 
 # unreadable regular file: the fopen() errno arm fires, distinct from the
-# directory branch. Needs an OS that enforces mode 000: root bypasses it, and
+# directory branch. Root bypasses mode 000, so skip it there.
-# Windows ignores chmod entirely, so probe rather than assume.
+if test "$(id -u)" -ne 0; then
-: >"$tmp/noperm.txt"
+    : >"$tmp/noperm.txt"
-chmod 000 "$tmp/noperm.txt"
+    chmod 000 "$tmp/noperm.txt"
-if ! cat "$tmp/noperm.txt" >/dev/null 2>&1; then
     run "$tmp/perm" "$tmp/noperm.txt"
+    chmod 644 "$tmp/noperm.txt"
     loghas 'Could not include URL list "[^"]+": .+'
     lognot 'not a regular file'
 fi
-chmod 644 "$tmp/noperm.txt"
 
 exit 0

tests/01_engine-filter.test

@@ -193,13 +193,5 @@ test "$(with_timeout httrack -O /dev/null -#test=filter "$pat" "$subj")" == "$su
 match '*[aX]X*[a]Y' 'aXaXaY'
 nomatch '*[aX]X*[a]Y' 'aXbXaY'
 
-# hostile patterns must not stack-overflow or hang: length + depth + work caps
+# hostile patterns must not stack-overflow or hang: length + work caps
 test "$(with_timeout httrack -O /dev/null -#test=filterbounds)" == "filterbounds: OK" || exit 1
-
-# and the depth cap must hold on the 1MB stack a Windows thread gets (#574):
-# uncapped, the same pattern recursed ~900KB deep. Skipped where 512K is too
-# small to even start the binary.
-if (ulimit -s 512 && httrack --version >/dev/null 2>&1); then
-    out=$(ulimit -s 512 && with_timeout httrack -O /dev/null -#test=filterbounds)
-    test "$out" == "filterbounds: OK" || exit 1
-fi

tests/01_engine-fsize.test

@@ -1,18 +0,0 @@
-#!/bin/bash
-#
-# File sizes past the 32-bit wrap ('httrack -#test=fsize <dir>'). 5GB, so a
-# signed *and* an unsigned truncation both show up; sparse, so it costs no disk.
-
-set -euo pipefail
-
-dir=$(mktemp -d)
-trap 'rm -rf "$dir"' EXIT
-
-out=$(httrack -#test=fsize "$dir")
-echo "$out"
-
-want="fsize: width=8,8 size=5368709120,5368709120 psize=5368709120 absent=-1"
-test "$out" == "$want" || {
-    echo "FAIL: unexpected size report (want '$want')"
-    exit 1
-}

tests/01_engine-idna.test

@@ -38,7 +38,4 @@ dec 'xn--!!!' '㚯'
 dec 'xn--already-encoded.com' 'ǮalǭǮreaǭǫdy.com'
 
 # second label has a truncated UTF-8 sequence: used to leak the segment buffer.
-# Fed as hex: the raw bytes must reach the encoder intact, but invalid UTF-8
+enc_bad "$(printf 'b\xc3\xbcchev.\xe4\xbe\x8b\xe5\xad\x62\xc3\xbccheple')"
-# cannot survive a Windows command line (it round-trips through UTF-16 to valid
-# U+FFFD, which then encodes fine instead of being rejected).
-enc_bad 'hex:62c3bc636865762ee4be8be5ad62c3bc636865706c65'

tests/01_engine-rcfile.test

@@ -32,13 +32,6 @@ esac
 tmp=$(mktemp -d "${TMPDIR:-/tmp}/httrack_rcfile.XXXXXX") || exit 1
 trap 'rm -rf "$tmp"' EXIT HUP INT QUIT PIPE TERM
 
-# HTS_HTTRACKRC is ".httrackrc" on POSIX but "httrackrc" on Windows: write both,
-# each platform reads the one it knows.
-write_rc() {
-    cat >"$1/.httrackrc"
-    cp "$1/.httrackrc" "$1/httrackrc"
-}
-
 # --- 1. alias token survives the bound intact -------------------------------
 d1="$tmp/intact"
 mkdir -p "$d1"
@@ -47,7 +40,7 @@ echo '<html><body>hello</body></html>' >"$d1/index.html"
 # optinclude_file() lowercases each config line, so the marker is lowercase to
 # survive the comparison verbatim.
 marker='zzz_rcfile_marker_0123456789_abcdefghijklmnopqrstuvwxyz_intact'
-printf 'user-agent=%s\n' "$marker" | write_rc "$d1"
+printf 'user-agent=%s\n' "$marker" >"$d1/.httrackrc"
 
 # Run with no -O so the working-directory .httrackrc is loaded (an -O path makes
 # the engine skip the rc files). Output lands in the temp dir. Guard the run so a
@@ -80,7 +73,7 @@ echo '<html><body>hi</body></html>' >"$d2/index.html"
 val=$(printf 'a%.0s' $(seq 1 200))
 for _ in $(seq 1 400); do
     printf 'user-agent=%s\n' "$val"
-done | write_rc "$d2"
+done >"$d2/.httrackrc"
 
 # The process aborts (httrack turns the fatal signal into exit 134 either way),
 # so the exit code does not distinguish the bounded abort from a heap overflow;

tests/01_engine-strsafe.test

@@ -32,10 +32,10 @@ case "$err" in
 esac
 
 # Same guarantee for the htsbuff builder. The source is exactly the buffer
-# capacity (6 bytes into a 6-byte buffer), so this also pins the boundary: a
+# capacity (4 bytes into a 4-byte buffer), so this also pins the boundary: a
 # '<=' off-by-one in the capacity check would let it through (and print "NOT
 # aborted"). Match the specific htsbuff abort message, not just any assert.
-err=$(httrack -#test=strsafe overflow-buff "abcdef" 2>&1) || true
+err=$(httrack -#test=strsafe overflow-buff "abcd" 2>&1) || true
 case "$err" in
 *"strsafe: NOT aborted"*)
     echo "htsbuff over-capacity write was NOT caught" >&2

tests/53_local-proxytrack-cache-corrupt.test

@@ -1,21 +0,0 @@
-#!/bin/bash
-# A corrupt proxytrack .ndx must not overflow the fixed cache-read buffers. The
-# sanitizer CI build turns any regression here into a hard stack/heap overflow.
-
-set -euo pipefail
-
-dir=$(mktemp -d)
-trap 'rm -rf "$dir"' EXIT
-
-# The first length-prefixed field declares far more than firstline[256] holds;
-# cache_brstr must clamp the copy to the destination, not the declared length.
-{
-    printf '4000\n'
-    head -c 4000 /dev/zero | tr '\0' 'A'
-} >"$dir/foo.ndx"
-: >"$dir/foo.dat" # cache_brstr runs only when the sibling .dat opens
-
-proxytrack --convert "$dir/out.arc" "$dir/foo.ndx" >/dev/null 2>&1 || {
-    echo "FAIL: proxytrack crashed/errored on a corrupt cache index"
-    exit 1
-}

tests/Makefile.am

@@ -57,7 +57,6 @@ TESTS = \
 	01_engine-pause.test \
 	01_engine-rcfile.test \
 	01_engine-reconcile.test \
-	01_engine-fsize.test \
 	01_engine-redirect.test \
 	01_engine-relative.test \
 	01_engine-robots.test \
@@ -129,7 +128,6 @@ TESTS = \
 	49_local-cookiewall.test \
 	50_local-contentcodings.test \
 	51_local-update-codec.test \
-	52_local-socks5.test \
+	52_local-socks5.test
-	53_local-proxytrack-cache-corrupt.test
 
 CLEANFILES = check-network_sh.cache