Address review: keep empty-PATH parity, fold the CI script list

Review of the array refactor flagged one behaviour divergence: splitting
PATH with `IFS=: read -ra` keeps empty fields (from doubled or leading
colons) as "" elements, where the old `echo $PATH | tr : ' '` word-split
dropped them, so the search loop would probe /htsserver. Skip the empty
fields to restore exact parity.

Also reflow the CI SHELL_SCRIPTS list as a folded block scalar, one
entry per line and sorted, so it reads cleanly; the folded value is the
same space-separated string.

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

.github/workflows/ci.yml

@@ -320,6 +320,21 @@ jobs:
   lint:
     name: lint (shellcheck, shfmt)
     runs-on: ubuntu-24.04
+    # Every tracked shell script; the globs expand at run time. Kept here so the
+    # shellcheck and shfmt steps below cannot drift apart.
+    env:
+      SHELL_SCRIPTS: >-
+        .githooks/pre-commit
+        bootstrap
+        build.sh
+        html/div/search.sh
+        man/makeman.sh
+        src/htsbasiccharsets.sh
+        src/htsentities.sh
+        src/webhttrack
+        tests/*.sh
+        tests/*.test
+        tools/mkdeb.sh
     steps:
       - uses: actions/checkout@v6
 
@@ -332,12 +347,11 @@ jobs:
           sudo apt-get install -y --no-install-recommends shellcheck shfmt
           shfmt --version
 
-      # Lint the scripts we maintain; the legacy scripts are a separate cleanup.
       - name: shellcheck
-        run: shellcheck man/makeman.sh tools/mkdeb.sh .githooks/pre-commit tests/*.test tests/check-network.sh
+        run: shellcheck $SHELL_SCRIPTS
 
       - name: shfmt
-        run: shfmt -d -i 4 man/makeman.sh tools/mkdeb.sh .githooks/pre-commit
+        run: shfmt -d -i 4 $SHELL_SCRIPTS
 
   # Check clang-format on CHANGED LINES ONLY. The engine predates clang-format
   # (it was shaped by an old Visual Studio formatter) and does not round-trip,

html/div/search.sh

@@ -1,8 +1,7 @@
-
 #!/bin/sh
 
 # Simple indexing test using HTTrack
-# A "real" script/program would use advanced search, and 
+# A "real" script/program would use advanced search, and
 # use dichotomy to find the word in the index.txt file
 # This script is really basic and NOT optimized, and
 # should not be used for professional purpose :)
@@ -11,50 +10,49 @@ TESTSITE="http://localhost/"
 
 # Create an index if necessary
 if ! test -f "index.txt"; then
-	echo "Building the index .."
+    echo "Building the index .."
-	rm -rf test
+    rm -rf test
-	httrack --display "$TESTSITE"  -%I -O test
+    httrack --display "$TESTSITE" -%I -O test
-	mv test/index.txt ./
+    mv test/index.txt ./
 fi
 
 # Convert crlf to lf
-if test "`head index.txt -n 1 | tr '\r' '#' | grep -c '#'`" = "1"; then
+if test "$(head index.txt -n 1 | tr '\r' '#' | grep -c '#')" = "1"; then
-	echo "Converting index to Unix LF style (not CR/LF) .."
+    echo "Converting index to Unix LF style (not CR/LF) .."
-	mv -f index.txt index.txt.old
+    mv -f index.txt index.txt.old
-	cat index.txt.old|tr -d '\r' > index.txt
+    tr -d '\r' <index.txt.old >index.txt
 fi
 
 keyword=-
 while test -n "$keyword"; do
-	printf "Enter a keyword: "
+    printf "Enter a keyword: "
-	read keyword
+    read -r keyword
 
-	if test -n "$keyword"; then
+    if test -n "$keyword"; then
-		FOUNDK="`grep -niE \"^$keyword\" index.txt`"
+        FOUNDK="$(grep -niE "^$keyword" index.txt)"
 
-		if test -n "$FOUNDK"; then	
+        if test -n "$FOUNDK"; then
-			if ! test `echo "$FOUNDK"|wc -l` = "1"; then
+            if ! test "$(echo "$FOUNDK" | wc -l)" = "1"; then
-				# Multiple matches
+                # Multiple matches
-				printf "Found multiple keywords: "
+                printf "Found multiple keywords: "
-				echo "$FOUNDK"|cut -f2 -d':'|tr '\n' ' '
+                echo "$FOUNDK" | cut -f2 -d':' | tr '\n' ' '
-				echo ""
+                echo ""
-				echo "Use keyword$ to find only one"
+                echo "Use keyword$ to find only one"
-			else
+            else
-				# One match
+                # One match
-				N=`echo "$FOUNDK"|cut -f1 -d':'`
+                N=$(echo "$FOUNDK" | cut -f1 -d':')
-				PM=`tail +$N index.txt|grep -nE "\("|head -n 1`
+                PM=$(tail "+$N" index.txt | grep -nE "\(" | head -n 1)
-				if ! echo "$PM"|grep "ignored">/dev/null; then
+                if ! echo "$PM" | grep "ignored" >/dev/null; then
-					M=`echo $PM|cut -f1 -d':'`
+                    M=$(echo "$PM" | cut -f1 -d':')
-					echo "Found in:"
+                    echo "Found in:"
-					cat index.txt | tail "+$N" | head -n "$M" | grep -E "[0-9]* " | cut -f2 -d' '
+                    tail "+$N" index.txt | head -n "$M" | grep -E "[0-9]* " | cut -f2 -d' '
-				else
+                else
-					echo "keyword ignored (too many hits)"
+                    echo "keyword ignored (too many hits)"
-				fi
+                fi
-					fi
+            fi
-		else
+        else
-			echo "not found"
+            echo "not found"
-		fi
+        fi
 
-	fi
+    fi
 done
-

src/htsback.c

@@ -2532,8 +2532,26 @@ void back_wait(struct_back * sback, httrackp * opt, cache_back * cache,
 #if HTS_USEOPENSSL
           /* SSL mode */
           if (back[i].r.ssl) {
+            int tunnel_ok = 1;
+
+            // https via proxy: CONNECT-tunnel before TLS (#85)
+            if (back[i].r.req.proxy.active && back[i].r.ssl_con == NULL) {
+              const int timeout = back[i].timeout > 0 ? back[i].timeout : 30;
+
+              tunnel_ok =
+                  http_proxy_tunnel(opt, &back[i].r, back[i].url_adr, timeout);
+              if (!tunnel_ok) {
+                if (!strnotempty(back[i].r.msg))
+                  strcpybuff(back[i].r.msg, "proxy CONNECT failed");
+                deletehttp(&back[i].r);
+                back[i].r.soc = INVALID_SOCKET;
+                back[i].r.statuscode = STATUSCODE_NON_FATAL;
+                back[i].status = STATUS_READY;
+                back_set_finished(sback, i);
+              }
+            }
             // handshake not yet launched
-            if (!back[i].r.ssl_con) {
+            if (tunnel_ok && !back[i].r.ssl_con) {
               SSL_CTX_set_options(openssl_ctx, SSL_OP_ALL);
               // new session
               back[i].r.ssl_con = SSL_new(openssl_ctx);
@@ -2551,7 +2569,7 @@ void back_wait(struct_back * sback, httrackp * opt, cache_back * cache,
                 back[i].r.statuscode = STATUSCODE_SSL_HANDSHAKE;
             }
             /* Error */
-            if (back[i].r.statuscode == STATUSCODE_SSL_HANDSHAKE) {
+            if (tunnel_ok && back[i].r.statuscode == STATUSCODE_SSL_HANDSHAKE) {
               strcpybuff(back[i].r.msg, "bad SSL/TLS handshake");
               deletehttp(&back[i].r);
               back[i].r.soc = INVALID_SOCKET;
@@ -3838,7 +3856,7 @@ void back_wait(struct_back * sback, httrackp * opt, cache_back * cache,
         /* funny log for commandline users */
         //if (!opt->quiet) {  
         // petite animation
-        if (opt->verbosedisplay == 1) {
+        if (opt->verbosedisplay == HTS_VERBOSE_SIMPLE) {
           if (back[i].status == STATUS_READY) {
             if (back[i].r.statuscode == HTTP_OK)
               printf("* %s%s (" LLintP " bytes) - OK" VT_CLREOL "\r",

src/htsbasiccharsets.sh

@@ -3,57 +3,59 @@
 
 # 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/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/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/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/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/CP*.TXT" | lftp
-echo "mget ftp://ftp.unicode.org/Public/MAPPINGS/VENDORS/MISC/KOI8*.TXT" | lftp
+    echo "mget ftp://ftp.unicode.org/Public/MAPPINGS/VENDORS/MISC/KOI8*.TXT" | lftp
-rm -f CP932.TXT CP936.TXT CP949.TXT CP950.TXT
+    rm -f CP932.TXT CP936.TXT CP949.TXT CP950.TXT
 fi
 
 # Produce code
-printf "/** GENERATED FILE ($0), DO NOT EDIT **/\n\n"
+printf '/** GENERATED FILE (%s), DO NOT EDIT **/\n\n' "$0"
-for i in *.TXT ; do
+for i in *.TXT; do
-  echo "processing $i" >&2
+    echo "processing $i" >&2
-  grep -vE "^(#|$)" $i | grep -E "^0x" | sed -e 's/[[:space:]]/ /g' | cut -f1,2 -d' ' | \
+    grep -vE "^(#|$)" "$i" | grep -E "^0x" | sed -e 's/[[:space:]]/ /g' | cut -f1,2 -d' ' |
-  (
+        (
-    unset arr
+            unset arr
-    while read LINE ; do
+            while read -r LINE; do
-      from=$[$(echo $LINE | cut -f1 -d' ')]
+                from=$(($(echo "$LINE" | cut -f1 -d' ')))
-      if ! test -n "$from"; then
+                if ! test -n "$from"; then
-        echo "error with $i" >&2
+                    echo "error with $i" >&2
-        exit 1
+                    exit 1
-      elif test $from -ge 256; then
+                elif test $from -ge 256; then
-        echo "out-of-range ($LINE) with $i" >&2
+                    echo "out-of-range ($LINE) with $i" >&2
-        exit 1
+                    exit 1
-      fi
+                fi
-      to=$(echo $LINE | cut -f2 -d' ') 
+                to=$(echo "$LINE" | cut -f2 -d' ')
-      arr[$from]=$to
+                arr[from]=$to
-    done
+            done
-    name=$(echo $i | tr 'A-Z' 'a-z' | tr '-' '_' | sed -e 's/\.txt//' -e 's/8859/iso_8859/')
+            # shellcheck disable=SC2018,SC2019 # charset filenames are ASCII; keep C-locale A-Z/a-z
-    printf "/* Table for $i */\nstatic const hts_UCS4 table_${name}[256] = {\n  "
+            name=$(echo "$i" | tr 'A-Z' 'a-z' | tr '-' '_' | sed -e 's/\.txt//' -e 's/8859/iso_8859/')
-    i=0
+            printf '/* Table for %s */\nstatic const hts_UCS4 table_%s[256] = {\n  ' "$i" "$name"
-    while test "$i" -lt 256; do
+            idx=0
-      if test "$i" -gt 0; then
+            while test "$idx" -lt 256; do
-        printf ", "
+                if test "$idx" -gt 0; then
-        if test $[${i}%8] -eq 0; then
+                    printf ", "
-          printf "\n  "
+                    if test $((idx % 8)) -eq 0; then
-        fi
+                        printf "\n  "
-      fi
+                    fi
-      value=${arr[$i]:-0}
+                fi
-      printf "0x%04x" $value
+                value=${arr[$idx]:-0}
-      i=$[${i}+1]
+                printf "0x%04x" "$value"
-    done
+                idx=$((idx + 1))
-    printf " };\n\n"
+            done
-  )
+            printf " };\n\n"
-  echo "processed $i" >&2
+        )
+    echo "processed $i" >&2
 done
 
 # Indexes
 printf "static const struct {\n  const char *name;\n  const hts_UCS4 *table;\n} table_mappings[] = {\n"
-for i in *.TXT ; do
+for i in *.TXT; do
-  name=$(echo $i | tr 'A-Z' 'a-z' | tr '-' '_' | sed -e 's/\.txt//' -e 's/8859/iso_8859/')
+    # shellcheck disable=SC2018,SC2019 # charset filenames are ASCII; keep C-locale A-Z/a-z
-  printf "  { \"$(echo $name | tr -d '_')\", table_${name} },\n"
+    name=$(echo "$i" | tr 'A-Z' 'a-z' | tr '-' '_' | sed -e 's/\.txt//' -e 's/8859/iso_8859/')
+    printf '  { "%s", table_%s },\n' "$(echo "$name" | tr -d '_')" "$name"
 done
 printf "  { NULL, NULL }\n};\n"

src/htscatchurl.c

@@ -135,7 +135,8 @@ HTSEXT_API T_SOC catch_url_init(int *port, /* 128 bytes */ char *adr) {
 // returns 0 if error
 // url: buffer where URL must be stored - or ip:port in case of failure
 // data: 32Kb
-HTSEXT_API int catch_url(T_SOC soc, char *url, char *method, char *data) {
+HTSEXT_API hts_boolean catch_url(T_SOC soc, char *url, char *method,
+                                 char *data) {
   int retour = 0;
 
   // connexion (accept)

src/htscore.c

@@ -2585,7 +2585,7 @@ static int mkdir_compat(const char *pathname) {
 
 /* path must end with "/" or with the finename (/tmp/bar/ or /tmp/bar/foo.zip) */
 /* Note: preserve errno */
-HTSEXT_API int dir_exists(const char *path) {
+HTSEXT_API hts_boolean dir_exists(const char *path) {
   const int err = errno;
   STRUCT_STAT st;
   char BIGSTK file[HTS_URLMAXSIZE * 2];
@@ -3342,7 +3342,8 @@ int back_fill(struct_back * sback, httrackp * opt, cache_back * cache,
               int ptr, int numero_passe) {
   int n = back_pluggable_sockets(sback, opt);
 
-  if (opt->savename_delayed == 2 && !opt->delayed_cached)       /* cancel (always delayed) */
+  if (opt->savename_delayed == HTS_SAVENAME_DELAYED_HARD &&
+      !opt->delayed_cached) /* cancel (always delayed) */
     return 0;
   if (n > 0) {
     int p;
@@ -3646,7 +3647,7 @@ HTSEXT_API int hts_setpause(httrackp * opt, int p) {
 }
 
 // ask for termination
-HTSEXT_API int hts_request_stop(httrackp * opt, int force) {
+HTSEXT_API int hts_request_stop(httrackp *opt, hts_boolean force) {
   if (opt != NULL) {
     hts_log_print(opt, LOG_ERROR, "Exit requested by shell or user");
     hts_mutexlock(&opt->state.lock);
@@ -3656,7 +3657,7 @@ HTSEXT_API int hts_request_stop(httrackp * opt, int force) {
   return 0;
 }
 
-HTSEXT_API int hts_has_stopped(httrackp * opt) {
+HTSEXT_API hts_boolean hts_has_stopped(httrackp *opt) {
   int ended;
   hts_mutexlock(&opt->state.lock);
   ended = opt->state.is_ended;
@@ -3678,12 +3679,12 @@ HTSEXT_API int hts_has_stopped(httrackp * opt) {
 //}
 // ajout d'URL
 // -1 : erreur
-HTSEXT_API int hts_addurl(httrackp * opt, char **url) {
+HTSEXT_API hts_boolean hts_addurl(httrackp *opt, char **url) {
   if (url)
     opt->state._hts_addurl = url;
   return (opt->state._hts_addurl != NULL);
 }
-HTSEXT_API int hts_resetaddurl(httrackp * opt) {
+HTSEXT_API hts_boolean hts_resetaddurl(httrackp *opt) {
   opt->state._hts_addurl = NULL;
   return (opt->state._hts_addurl != NULL);
 }
@@ -3702,7 +3703,9 @@ HTSEXT_API int copy_htsopt(const httrackp * from, httrackp * to) {
   if (from->maxsoc > 0)
     to->maxsoc = from->maxsoc;
 
-  if (from->nearlink > -1)
+  /* hts_boolean/enum fields are unsigned (GCC), so a bare `> -1` unset-guard
+     is always false; cast to int to keep the -1 "unset" sentinel test. */
+  if ((int) from->nearlink > -1)
     to->nearlink = from->nearlink;
 
   if (from->timeout > -1)
@@ -3729,10 +3732,10 @@ HTSEXT_API int copy_htsopt(const httrackp * from, httrackp * to) {
   if (from->hostcontrol > -1)
     to->hostcontrol = from->hostcontrol;
 
-  if (from->errpage > -1)
+  if ((int) from->errpage > -1)
     to->errpage = from->errpage;
 
-  if (from->parseall > -1)
+  if ((int) from->parseall > -1)
     to->parseall = from->parseall;
 
   // test all: bit 8 de travel
@@ -3844,7 +3847,7 @@ int htsAddLink(htsmoduleStruct * str, char *link) {
             a = opt->savename_type;
             b = opt->savename_83;
             opt->savename_type = 0;
-            opt->savename_83 = 0;
+            opt->savename_83 = HTS_SAVENAME_83_LONG;
             // note: adr,fil peuvent être patchés
             r =
               url_savename(&afs, NULL, NULL, NULL, opt, sback, cache, hashptr, ptr, numero_passe,

src/htscoremain.c

@@ -612,12 +612,12 @@ static int hts_main_internal(int argc, char **argv, httrackp * opt) {
   /* Terminal is a tty, may ask questions and display funny information */
   if (isatty(1)) {
     opt->quiet = 0;
-    opt->verbosedisplay = 1;
+    opt->verbosedisplay = HTS_VERBOSE_SIMPLE;
   }
   /* Not a tty, no stdin input or funny output! */
   else {
     opt->quiet = 1;
-    opt->verbosedisplay = 0;
+    opt->verbosedisplay = HTS_VERBOSE_NONE;
   }
 #endif
 
@@ -953,9 +953,11 @@ static int hts_main_internal(int argc, char **argv, httrackp * opt) {
       p = buff;
       do {
         int insert_after_argc;
+        int quoted; /* "" unquotes to empty but is still a real token (#106) */
 
         // read next
         lastp = p;
+        quoted = (p != NULL && *p == '"');
         if (p) {
           p = next_token(p, 1);
           if (p) {
@@ -966,7 +968,7 @@ static int hts_main_internal(int argc, char **argv, httrackp * opt) {
 
         /* Insert parameters BUT so that they can be in the same order */
         if (lastp) {
-          if (strnotempty(lastp)) {
+          if (strnotempty(lastp) || quoted) {
             insert_after_argc = argc - insert_after;
             cmdl_ins(lastp, insert_after_argc, (argv + insert_after), x_argvblk,
                      x_argvblk_size, x_ptr);
@@ -1815,24 +1817,22 @@ static int hts_main_internal(int argc, char **argv, httrackp * opt) {
                 com++;
             }
             break;
-          case 'L':
+          case 'L': {
-            {
+            sscanf(com + 1, "%d", (int *) &opt->savename_83);
-              sscanf(com + 1, "%d", &opt->savename_83);
+            switch (opt->savename_83) {
-              switch (opt->savename_83) {
+            case 0: // 8-3 (ISO9660 L1)
-              case 0:          // 8-3 (ISO9660 L1)
+              opt->savename_83 = HTS_SAVENAME_83_DOS;
-                opt->savename_83 = 1;
+              break;
-                break;
+            case 1:
-              case 1:
+              opt->savename_83 = HTS_SAVENAME_83_LONG;
-                opt->savename_83 = 0;
+              break;
-                break;
+            default: // 2 == ISO9660 (ISO9660 L2)
-              default:         // 2 == ISO9660 (ISO9660 L2)
+              opt->savename_83 = HTS_SAVENAME_83_ISO9660;
-                opt->savename_83 = 2;
+              break;
-                break;
-              }
-              while(isdigit((unsigned char) *(com + 1)))
-                com++;
             }
-            break;
+            while (isdigit((unsigned char) *(com + 1)))
+              com++;
+          } break;
           case 's':
             if (isdigit((unsigned char) *(com + 1))) {
               sscanf(com + 1, "%d", (int *) &opt->robots);
@@ -1989,9 +1989,9 @@ static int hts_main_internal(int argc, char **argv, httrackp * opt) {
                 }
                 break;          // url hack
               case 'v':
-                opt->verbosedisplay = 2;
+                opt->verbosedisplay = HTS_VERBOSE_FULL;
                 if (isdigit((unsigned char) *(com + 1))) {
-                  sscanf(com + 1, "%d", &opt->verbosedisplay);
+                  sscanf(com + 1, "%d", (int *) &opt->verbosedisplay);
                   while(isdigit((unsigned char) *(com + 1)))
                     com++;
                 }
@@ -2004,9 +2004,9 @@ static int hts_main_internal(int argc, char **argv, httrackp * opt) {
                 }
                 break;
               case 'N':
-                opt->savename_delayed = 2;
+                opt->savename_delayed = HTS_SAVENAME_DELAYED_HARD;
                 if (isdigit((unsigned char) *(com + 1))) {
-                  sscanf(com + 1, "%d", &opt->savename_delayed);
+                  sscanf(com + 1, "%d", (int *) &opt->savename_delayed);
                   while(isdigit((unsigned char) *(com + 1)))
                     com++;
                 }
@@ -2787,6 +2787,47 @@ static int hts_main_internal(int argc, char **argv, httrackp * opt) {
                   return 0;
                 }
                 break;
+              case 'l': /* lienrelatif: relative link from curr_fil to link */
+                if (na + 2 >= argc) {
+                  HTS_PANIC_PRINTF(
+                      "Option #l needs a link and a current-file path");
+                  printf(
+                      "Example: '-#l' 'host/dir/img.gif' 'host/dir/p.html'\n");
+                  htsmain_free();
+                  return -1;
+                } else {
+                  char s[HTS_URLMAXSIZE * 2];
+
+                  if (lienrelatif(s, sizeof(s), argv[na + 1], argv[na + 2]) ==
+                      0)
+                    printf("relative=%s\n", s);
+                  else
+                    printf("relative=<ERROR>\n");
+                  htsmain_free();
+                  return 0;
+                }
+                break;
+              case 'i': /* ident_url_relatif: resolve a link -> adr/fil */
+                if (na + 3 >= argc) {
+                  HTS_PANIC_PRINTF(
+                      "Option #i needs a link, an origin address and file");
+                  printf("Example: '-#i' '../img.gif' 'www.foo.com' "
+                         "'/d/p.html'\n");
+                  htsmain_free();
+                  return -1;
+                } else {
+                  lien_adrfil af;
+                  const int r = ident_url_relatif(argv[na + 1], argv[na + 2],
+                                                  argv[na + 3], &af);
+
+                  if (r == 0)
+                    printf("adr=%s fil=%s\n", af.adr, af.fil);
+                  else
+                    printf("error=%d\n", r);
+                  htsmain_free();
+                  return 0;
+                }
+                break;
               case '2':        // mimedefs
                 if (na + 1 >= argc) {
                   HTS_PANIC_PRINTF("Option #2 needs to be followed by an URL");
@@ -3096,6 +3137,78 @@ static int hts_main_internal(int argc, char **argv, httrackp * opt) {
                 htsmain_free();
                 return 0;
                 break;
+              case '9': { // copy_htsopt selftest: httrack -#9
+                httrackp *from = hts_create_opt();
+                httrackp *to = hts_create_opt();
+                int err = 0;
+
+                /* from-values differ from both the to-values and the
+                   hts_create_opt() defaults (nearlink FALSE, errpage/parseall
+                   TRUE), so a copy that no-ops or just resets to defaults is
+                   caught too, not only the unsigned-guard bug. */
+                from->retry = 7; /* int field: positive control */
+                to->retry = 0;
+                from->nearlink = HTS_TRUE;
+                to->nearlink = HTS_FALSE;
+                from->errpage = HTS_FALSE;
+                to->errpage = HTS_TRUE;
+                from->parseall = HTS_FALSE;
+                to->parseall = HTS_TRUE;
+
+                copy_htsopt(from, to);
+
+                if (to->retry != 7)
+                  err = 1;
+                if (to->nearlink != HTS_TRUE)
+                  err = 1;
+                if (to->errpage != HTS_FALSE)
+                  err = 1;
+                if (to->parseall != HTS_FALSE)
+                  err = 1;
+
+                hts_free_opt(from);
+                hts_free_opt(to);
+                printf("copy-htsopt: %s\n", err ? "FAIL" : "OK");
+                htsmain_free();
+                return err;
+              } break;
+              case 'Q': { // cookie request-header selftest: httrack -#Q
+                static t_cookie cookie;
+                char hdr[1024];
+                /* RFC 6265: bare name=value pairs, no $Version/$Path (#151). */
+                const char *expected = "Cookie: name=value; has_js=1" H_CRLF;
+                int err = 0;
+
+                const char *dom = "www.example.com";
+                int added;
+
+                cookie.max_len = (int) sizeof(cookie.data);
+                cookie.data[0] = '\0';
+                added = cookie_add(&cookie, "name", "value", dom, "/");
+                added |= cookie_add(&cookie, "has_js", "1", dom, "/");
+                /* different domain: must be filtered out */
+                added |= cookie_add(&cookie, "junk", "x", "other.org", "/");
+                if (added) {
+                  printf("cookie-header: FAIL (cookie_add setup)\n");
+                  htsmain_free();
+                  return 1;
+                }
+
+                http_cookie_header_selftest(&cookie, dom, "/", hdr,
+                                            sizeof(hdr));
+                if (strcmp(hdr, expected) != 0)
+                  err = 1;
+                if (strstr(hdr, "$Version") != NULL ||
+                    strstr(hdr, "$Path") != NULL)
+                  err = 1;
+                if (strstr(hdr, "junk") != NULL) // wrong-domain cookie leaked
+                  err = 1;
+                printf("cookie-header: %s\n", err ? "FAIL" : "OK");
+                if (err)
+                  printf("  got: %s\n", hdr);
+                htsmain_free();
+                return err;
+              } break;
               case '!':
                 HTS_PANIC_PRINTF
                   ("Option #! is disabled for security reasons");

src/htsentities.sh

@@ -33,43 +33,43 @@ EOF
         else
             GET "${url}"
         fi
-    ) \
+    ) |
-        | grep -E '^<!ENTITY [a-zA-Z0-9_]' \
+        grep -E '^<!ENTITY [a-zA-Z0-9_]' |
-        | sed \
+        sed \
-        -e 's/<!ENTITY //' -e "s/[[:space:]][[:space:]]*/ /g" \
+            -e 's/<!ENTITY //' -e "s/[[:space:]][[:space:]]*/ /g" \
-        -e 's/-->$//' \
+            -e 's/-->$//' \
-        -e 's/\([^ ]*\) CDATA "&#\([^\"]*\);" -- \(.*\)/\1 \2 \3/'\
+            -e 's/\([^ ]*\) CDATA "&#\([^\"]*\);" -- \(.*\)/\1 \2 \3/' |
-| ( \
+        (
-        read A
+            read -r A
-        while test -n "$A"; do
+            while test -n "$A"; do
-            ent="${A%% *}"
+                ent="${A%% *}"
-            code=$(echo "$A"|cut -f2 -d' ')
+                code=$(echo "$A" | cut -f2 -d' ')
-            # compute hash
+                # compute hash
-            hash=0
+                hash=0
-            i=0
+                i=0
-            a=1664525
+                a=1664525
-            c=1013904223
+                c=1013904223
-            m="$[1 << 32]"
+                m="$((1 << 32))"
-            while test "$i" -lt ${#ent}; do
+                while test "$i" -lt ${#ent}; do
-                d="$(echo -n "${ent:${i}:1}"|hexdump -v -e '/1 "%d"')"
+                    d="$(echo -n "${ent:${i}:1}" | hexdump -v -e '/1 "%d"')"
-                hash="$[((${hash}*${a})%(${m})+${d}+${c})%(${m})]"
+                    hash="$((((hash * a) % (m) + d + c) % (m)))"
-                i=$[${i}+1]
+                    i=$((i + 1))
-            done
+                done
-            echo -e "    /* $A */"
+                echo -e "    /* $A */"
-            echo -e "  case ${hash}u:"
+                echo -e "  case ${hash}u:"
-            echo -e "    if (len == ${#ent} /* && strncmp(ent, \"${ent}\") == 0 */) {"
+                echo -e "    if (len == ${#ent} /* && strncmp(ent, \"${ent}\") == 0 */) {"
-            echo -e "      return ${code};"
+                echo -e "      return ${code};"
-            echo -e "    }"
+                echo -e "    }"
-            echo -e "    break;"
+                echo -e "    break;"
 
-            # next
+                # next
-            read A
+                read -r A
-        done
+            done
-    )
+        )
     cat <<EOF
   }
   /* unknown */
   return -1;
 }
 EOF
-) > ${dest}
+) >${dest}

src/htsglobal.h

@@ -242,6 +242,14 @@ Please visit our Website: http://www.httrack.com
 #define HTS_NOPARAM "(none)"
 #define HTS_NOPARAM2 "\"(none)\""
 
+/* Boolean flag for option fields and API yes/no returns. An enum (not C bool)
+   so it stays int-sized: option fields keep the httrackp layout/ABI, and a
+   return type stays compatible with the int it replaces. */
+#ifndef HTS_DEF_DEFSTRUCT_hts_boolean
+#define HTS_DEF_DEFSTRUCT_hts_boolean
+typedef enum hts_boolean { HTS_FALSE = 0, HTS_TRUE = 1 } hts_boolean;
+#endif
+
 /* Larger/smaller of two values. Macros: arguments are evaluated twice. */
 #define maximum(A,B) ( (A) > (B) ? (A) : (B) )
 

src/htslib.c

@@ -644,6 +644,165 @@ T_SOC http_fopen(httrackp * opt, const char *adr, const char *fil, htsblk * reto
   return http_xfopen(opt, 0, 1, 1, NULL, adr, fil, retour);
 }
 
+// Read a CRLF line from a non-blocking socket (waits up to timeout per recv).
+// Returns the line length (0 = empty), or -1 on timeout/EOF/error.
+static int proxy_getline(T_SOC soc, char *s, int max, int timeout) {
+  int j = 0;
+
+  for (;;) {
+    unsigned char ch;
+    int n;
+
+    if (!check_readinput_t(soc, timeout))
+      return -1; // timed out waiting for data
+    n = (int) recv(soc, &ch, 1, 0);
+    if (n == 1) {
+      if (ch == 13) // CR
+        continue;
+      if (ch == 10) // LF: end of line
+        break;
+      if (j >= max - 1)
+        return -1; // line too long: bound the read against a hostile proxy
+      s[j++] = (char) ch;
+    } else if (n == 0) {
+      return -1; // connection closed
+    } else {
+#ifdef _WIN32
+      if (WSAGetLastError() == WSAEWOULDBLOCK)
+        continue;
+#else
+      if (errno == EINTR || errno == EAGAIN || errno == EWOULDBLOCK)
+        continue;
+#endif
+      return -1;
+    }
+  }
+  s[j] = '\0';
+  return j;
+}
+
+int http_proxy_tunnel(httrackp *opt, htsblk *retour, const char *adr,
+                      int timeout) {
+  const T_SOC soc = retour->soc;
+  const char *const host = jump_identification_const(adr); // host[:port]
+  const char *const portsep = jump_toport_const(adr);      // ":port" or NULL
+  char BIGSTK authority[HTS_URLMAXSIZE * 2];
+  char BIGSTK req[HTS_URLMAXSIZE * 4 + 1100];
+  char line[1024];
+  int code;
+
+  if (soc == INVALID_SOCKET)
+    return 0;
+
+  // CONNECT needs an explicit host:port; default the https port
+  authority[0] = '\0';
+  if (portsep != NULL)
+    strlcatbuff(authority, host, sizeof(authority)); // already host:port
+  else
+    snprintf(authority, sizeof(authority), "%s:%d", host, 443);
+
+  // backstop: never let a stray CR/LF in the host smuggle a second line into
+  // the CONNECT request (the host is already sanitized upstream)
+  {
+    const char *c;
+
+    for (c = authority; *c != '\0'; c++) {
+      if ((unsigned char) *c < ' ') {
+        strcpybuff(retour->msg, "proxy CONNECT: invalid host");
+        return 0;
+      }
+    }
+  }
+
+  snprintf(req, sizeof(req), "CONNECT %s HTTP/1.0" H_CRLF "Host: %s" H_CRLF,
+           authority, authority);
+
+  // creds go on the CONNECT, not the tunneled origin request
+  if (link_has_authorization(retour->req.proxy.name)) {
+    const char *a = jump_identification_const(retour->req.proxy.name);
+    const char *astart = jump_protocol_const(retour->req.proxy.name);
+    char autorisation[1100];
+    char user_pass[256];
+
+    autorisation[0] = user_pass[0] = '\0';
+    strncatbuff(user_pass, astart, (int) (a - astart) - 1);
+    strcpybuff(user_pass, unescape_http(OPT_GET_BUFF(opt),
+                                        OPT_GET_BUFF_SIZE(opt), user_pass));
+    code64((unsigned char *) user_pass, (int) strlen(user_pass),
+           (unsigned char *) autorisation, 0);
+    strlcatbuff(req, "Proxy-Authorization: Basic ", sizeof(req));
+    strlcatbuff(req, autorisation, sizeof(req));
+    strlcatbuff(req, H_CRLF, sizeof(req));
+  }
+  strlcatbuff(req, H_CRLF, sizeof(req)); // end of request headers
+
+  // raw send: ssl is set, so sendc() would route to TLS
+  {
+    const char *p = req;
+    size_t remain = strlen(req);
+    int stalls = 0;
+
+    while (remain > 0) {
+      const int n = (int) send(soc, p, (int) remain, 0);
+
+      if (n > 0) {
+        p += n;
+        remain -= (size_t) n;
+        stalls = 0;
+      } else {
+#ifdef _WIN32
+        const int wouldblock = (WSAGetLastError() == WSAEWOULDBLOCK);
+#else
+        const int wouldblock =
+            (errno == EAGAIN || errno == EWOULDBLOCK || errno == EINTR);
+#endif
+        // don't spin forever on a fatal error or an unwritable socket
+        if (!wouldblock || !check_writeinput_t(soc, timeout) ||
+            ++stalls > 100) {
+          strcpybuff(retour->msg, "proxy CONNECT: write error");
+          return 0;
+        }
+      }
+    }
+  }
+
+  // proxy status line: "HTTP/1.x <code> ..."
+  if (proxy_getline(soc, line, sizeof(line), timeout) < 0) {
+    strcpybuff(retour->msg, "proxy CONNECT: no response");
+    return 0;
+  }
+  if (sscanf(line, "HTTP/%*d.%*d %d", &code) < 1)
+    code = 0;
+  if (code < 200 || code >= 300) {
+    snprintf(retour->msg, sizeof(retour->msg), "proxy CONNECT refused: %s",
+             strnotempty(line) ? line : "(no status)");
+    return 0;
+  }
+
+  // drain headers to the blank line; cap the count so a flooding proxy can't
+  // stall the crawl
+  {
+    int headers = 0;
+
+    for (;;) {
+      const int n = proxy_getline(soc, line, sizeof(line), timeout);
+
+      if (n < 0) {
+        strcpybuff(retour->msg, "proxy CONNECT: truncated response");
+        return 0;
+      }
+      if (n == 0)
+        break; // blank line: tunnel ready
+      if (++headers > 64) {
+        strcpybuff(retour->msg, "proxy CONNECT: too many response headers");
+        return 0;
+      }
+    }
+  }
+
+  return 1;
+}
+
 // ouverture d'une liaison http, envoi d'une requète
 // mode: 0 GET  1 HEAD  [2 POST]
 // treat: traiter header?
@@ -680,14 +839,14 @@ T_SOC http_xfopen(httrackp * opt, int mode, int treat, int waitconnect,
 
   /* connexion */
   if (retour) {
-    if ((!(retour->req.proxy.active))
+    /* no proxy, or proxy not usable here (local file) */
-        || ((strcmp(adr, "file://") == 0)
+    if ((!(retour->req.proxy.active)) || (strcmp(adr, "file://") == 0)) {
-            || (strncmp(adr, "https://", 8) == 0)
-        )
-      ) {                       /* pas de proxy, ou non utilisable ici */
       soc = newhttp(opt, adr, retour, -1, waitconnect);
     } else {
-      soc = newhttp(opt, retour->req.proxy.name, retour, retour->req.proxy.port, waitconnect);  // ouvrir sur le proxy à la place
+      // to the proxy; https tunnels to the origin via CONNECT in back_wait
+      // (#85)
+      soc = newhttp(opt, retour->req.proxy.name, retour, retour->req.proxy.port,
+                    waitconnect);
     }
   } else {
     soc = newhttp(opt, adr, NULL, -1, waitconnect);
@@ -874,6 +1033,50 @@ static void print_buffer(buff_struct*const str, const char *format, ...) {
   assertf(str->pos < str->capacity);
 }
 
+/* Append the request "Cookie:" header line for every stored cookie matching
+   domain/path. RFC 6265 form: bare "name=value" pairs joined by "; ", no
+   $Version/$Path attributes (those are RFC 2965 syntax that modern servers
+   reject, issue #151). Returns the number of cookies emitted. */
+static int append_cookie_header(buff_struct *bstr, t_cookie *cookie,
+                                const char *domain, const char *path) {
+  char buffer[8192];
+  char *b;
+  int cook = 0;
+  int max_cookies = 8;
+
+  if (cookie == NULL)
+    return 0;
+  b = cookie->data;
+  do {
+    b = cookie_find(b, "", domain, path); // next matching cookie
+    if (b != NULL) {
+      max_cookies--;
+      if (!cook) {
+        print_buffer(bstr, "Cookie: ");
+        cook = 1;
+      } else
+        print_buffer(bstr, "; ");
+      print_buffer(bstr, "%s", cookie_get(buffer, b, 5));
+      print_buffer(bstr, "=%s", cookie_get(buffer, b, 6));
+      b = cookie_nextfield(b);
+    }
+  } while (b != NULL && max_cookies > 0);
+  if (cook)
+    print_buffer(bstr, H_CRLF);
+  return cook;
+}
+
+/* Self-test entry for append_cookie_header(): build the request Cookie line
+   into dst (always NUL-terminated). Returns the number of cookies emitted. */
+int http_cookie_header_selftest(t_cookie *cookie, const char *domain,
+                                const char *path, char *dst, size_t dst_size) {
+  buff_struct bstr = {dst, dst_size, 0};
+
+  assertf(dst != NULL && dst_size > 0);
+  dst[0] = '\0';
+  return append_cookie_header(&bstr, cookie, domain, path);
+}
+
 // envoi d'une requète
 int http_sendhead(httrackp * opt, t_cookie * cookie, int mode,
                   const char *xsend, const char *adr, const char *fil,
@@ -999,8 +1202,8 @@ int http_sendhead(httrackp * opt, t_cookie * cookie, int mode,
     if (xsend)
       print_buffer(&bstr, "%s", xsend);  // éventuelles autres lignes
 
-    // tester proxy authentication
+    // for https, auth rides the CONNECT (the tunneled GET would leak it)
-    if (retour->req.proxy.active) {
+    if (retour->req.proxy.active && strncmp(adr, "https://", 8) != 0) {
       if (link_has_authorization(retour->req.proxy.name)) {     // et hop, authentification proxy!
         const char *a = jump_identification_const(retour->req.proxy.name);
         const char *astart = jump_protocol_const(retour->req.proxy.name);
@@ -1048,34 +1251,9 @@ int http_sendhead(httrackp * opt, t_cookie * cookie, int mode,
                          search_tag + strlen(POSTTOK) + 1))));
       }
     }
-    // gestion cookies?
+    // send stored cookies matching this host/path
     if (cookie) {
-      char buffer[8192];
+      append_cookie_header(&bstr, cookie, jump_identification_const(adr), fil);
-      char *b = cookie->data;
-      int cook = 0;
-      int max_cookies = 8;
-
-      do {
-        b = cookie_find(b, "", jump_identification_const(adr), fil);  // prochain cookie satisfaisant aux conditions
-        if (b != NULL) {
-          max_cookies--;
-          if (!cook) {
-            print_buffer(&bstr, "Cookie: $Version=1; ");
-            cook = 1;
-          } else
-            print_buffer(&bstr, "; ");
-          print_buffer(&bstr, "%s", cookie_get(buffer, b, 5));
-          print_buffer(&bstr, "=%s", cookie_get(buffer, b, 6));
-          print_buffer(&bstr, "; $Path=%s", cookie_get(buffer, b, 2));
-          b = cookie_nextfield(b);
-        }
-      } while(b != NULL && max_cookies > 0);
-      if (cook) {               // on a envoyé un (ou plusieurs) cookie?
-        print_buffer(&bstr, H_CRLF);
-#if DEBUG_COOK
-        printf("Header:\n%s\n", bstr.buffer);
-#endif
-      }
     }
     // gérer le keep-alive (garder socket)
     if (retour->req.http11 && !retour->req.nokeepalive) {
@@ -1808,6 +1986,24 @@ int check_readinput_t(T_SOC soc, int timeout) {
     return 0;
 }
 
+// wait until the socket is writable, up to timeout seconds
+int check_writeinput_t(T_SOC soc, int timeout) {
+  if (soc != INVALID_SOCKET) {
+    fd_set fds;
+    struct timeval tv;
+    const int isoc = (int) soc;
+
+    assertf(isoc == soc);
+    FD_ZERO(&fds);
+    FD_SET(isoc, &fds);
+    tv.tv_sec = timeout;
+    tv.tv_usec = 0;
+    select(isoc + 1, NULL, &fds, NULL, &tv);
+    return FD_ISSET(isoc, &fds) ? 1 : 0;
+  } else
+    return 0;
+}
+
 // idem, sauf qu'ici on peut choisir la taille max de données à recevoir
 // SI bufl==0 alors le buffer est censé être de 8kos, et on recoit par bloc de lignes
 // en éliminant les cr (ex: header), arrêt si double-lf
@@ -2409,6 +2605,8 @@ int ident_url_absolute(const char *url, lien_adrfil *adrfil) {
     for(i = 0; adrfil->fil[i] != '\0'; i++)
       if (adrfil->fil[i] == '\\')
         adrfil->fil[i] = '/';
+    // collapse ../ like the http branch above (path-traversal safety)
+    fil_simplifie(adrfil->fil);
   }
 
   // no hostname
@@ -3646,8 +3844,9 @@ HTSEXT_API char *unescape_http(char *const catbuff, const size_t size, const cha
 // DOES NOT DECODE %25 (part of CHAR_DELIM)
 // no_high & 1: decode high chars
 // no_high & 2: decode space
-HTSEXT_API char *unescape_http_unharm(char *const catbuff, const size_t size, 
+HTSEXT_API char *unescape_http_unharm(char *const catbuff, const size_t size,
-                                      const char *s, const int no_high) {
+                                      const char *s,
+                                      const hts_boolean no_high) {
   size_t i, j;
 
   RUNTIME_TIME_CHECK_SIZE(size);
@@ -3931,8 +4130,8 @@ void hts_replace(char *s, char from, char to) {
 
 // guess a local file's mime type (e.g. fil="toto.gif" -> s="image/gif")
 // returns 1 if a type was written to s, 0 otherwise
-int guess_httptype_sized(httrackp *opt, char *s, size_t ssize,
+hts_boolean guess_httptype_sized(httrackp *opt, char *s, size_t ssize,
-                         const char *fil) {
+                                 const char *fil) {
   return get_httptype_sized(opt, s, ssize, fil, 1);
 }
 
@@ -3945,8 +4144,8 @@ void guess_httptype(httrackp * opt, char *s, const char *fil) {
 // 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
-HTSEXT_API int get_httptype_sized(httrackp *opt, char *s, size_t ssize,
+HTSEXT_API hts_boolean get_httptype_sized(httrackp *opt, char *s, size_t ssize,
-                                  const char *fil, int flag) {
+                                          const char *fil, hts_boolean flag) {
   // userdef overrides get_httptype (a rule with an empty value, e.g. "--assume
   // cgi=", matches but writes nothing: report it as "no type" like the old
   // code, whose callers tested strnotempty(s))
@@ -4196,7 +4395,7 @@ HTSEXT_API int is_userknowntype(httrackp * opt, const char *fil) {
 
 // page dynamique?
 // is_dyntype(get_ext("foo.asp"))
-HTSEXT_API int is_dyntype(const char *fil) {
+HTSEXT_API hts_boolean is_dyntype(const char *fil) {
   int j = 0;
 
   if (!fil)
@@ -4214,7 +4413,7 @@ HTSEXT_API int is_dyntype(const char *fil) {
 
 // types critiques qui ne doivent pas être changés car renvoyés par des serveurs qui ne
 // connaissent pas le type
-int may_unknown(httrackp * opt, const char *st) {
+hts_boolean may_unknown(httrackp *opt, const char *st) {
   int j = 0;
 
   // types média
@@ -5236,7 +5435,8 @@ HTSEXT_API int hts_uninit_module(void) {
 }
 
 // legacy. do not use
-HTSEXT_API int hts_log(httrackp * opt, const char *prefix, const char *msg) {
+HTSEXT_API hts_boolean hts_log(httrackp *opt, const char *prefix,
+                               const char *msg) {
   if (opt->log != NULL) {
     fspc(opt, opt->log, prefix);
     fprintf(opt->log, "%s" LF, msg);
@@ -5466,9 +5666,10 @@ HTSEXT_API httrackp *hts_create_opt(void) {
              "Mozilla/4.5 (compatible; HTTrack 3.0x; Windows 98)");
   StringCopy(opt->referer, "");
   StringCopy(opt->from, "");
-  opt->savename_83 = 0;         // noms longs par défaut
+  opt->savename_83 = HTS_SAVENAME_83_LONG; // long names by default
   opt->savename_type = 0;       // avec structure originale
-  opt->savename_delayed = 2;    // hard delayed type (default)
+  opt->savename_delayed =
+      HTS_SAVENAME_DELAYED_HARD; // always delay the type check (default)
   opt->delayed_cached = HTS_TRUE;
   opt->mimehtml = HTS_FALSE;
   opt->parsejava = HTSPARSE_DEFAULT;    // parser classes
@@ -5493,7 +5694,7 @@ HTSEXT_API httrackp *hts_create_opt(void) {
   opt->parseall = HTS_TRUE;
   opt->parsedebug = HTS_FALSE;
   opt->norecatch = HTS_FALSE;
-  opt->verbosedisplay = 0;      // pas d'animation texte
+  opt->verbosedisplay = HTS_VERBOSE_NONE; // no text animation
   opt->sizehack = HTS_FALSE;
   opt->urlhack = HTS_TRUE;
   StringCopy(opt->footer, HTS_DEFAULT_FOOTER);

src/htslib.h

@@ -182,6 +182,11 @@ int http_sendhead(httrackp * opt, t_cookie * cookie, int mode, const char *xsend
                   const char *adr, const char *fil,
                   const char *referer_adr, const char *referer_fil,
                   htsblk * retour);
+/* Build the request "Cookie:" header line for stored cookies matching
+   domain/path into dst (NUL-terminated). Exposed for the -#Q self-test;
+   wraps the same logic http_sendhead() uses. Returns cookies emitted. */
+int http_cookie_header_selftest(t_cookie *cookie, const char *domain,
+                                const char *path, char *dst, size_t dst_size);
 
 //int newhttp(char* iadr,char* err=NULL);
 T_SOC newhttp(httrackp * opt, const char *iadr, htsblk * retour, int port,
@@ -193,6 +198,17 @@ HTS_INLINE void deletesoc_r(htsblk * r);
 htsblk http_test(httrackp * opt, const char *adr, const char *fil, char *loc);
 int check_readinput(htsblk * r);
 int check_readinput_t(T_SOC soc, int timeout);
+int check_writeinput_t(T_SOC soc, int timeout);
+
+/* Open an HTTP CONNECT tunnel through the active proxy for an https request:
+   `retour->soc` must already be TCP-connected to the proxy, and `adr` is the
+   origin authority (url_adr, e.g. "https://host:port"). Sends the CONNECT
+   request (with Proxy-Authorization when the proxy carries credentials) and
+   reads the proxy's status line, so the caller's TLS handshake then runs
+   end-to-end with the origin. Blocks up to `timeout` seconds. Returns 1 on a
+   2xx tunnel, 0 on failure (retour->msg/statuscode set). */
+int http_proxy_tunnel(httrackp *opt, htsblk *retour, const char *adr,
+                      int timeout);
 void treathead(t_cookie * cookie, const char *adr, const char *fil, htsblk * retour,
                char *rcvd);
 void treatfirstline(htsblk * retour, const char *rcvd);

src/htsname.c

@@ -184,10 +184,11 @@ int url_savename(lien_adrfilsave *const afs,
 
   /* 8-3 ? */
   switch (opt->savename_83) {
-  case 1:                      // 8-3
+  case HTS_SAVENAME_83_DOS: // 8-3
     max_char = 8;
     break;
-  case 2:                      // Level 2 File names may be up to 31 characters.
+  case HTS_SAVENAME_83_ISO9660: // Level 2 File names may be up to 31
+                                // characters.
     max_char = 31;
     break;
   default:
@@ -324,7 +325,7 @@ int url_savename(lien_adrfilsave *const afs,
   }
 
   /* replace shtml to html.. */
-  if (opt->savename_delayed == 2)
+  if (opt->savename_delayed == HTS_SAVENAME_DELAYED_HARD)
     is_html = -1;               /* ALWAYS delay type */
   else
     is_html = ishtml(opt, fil);
@@ -363,7 +364,9 @@ int url_savename(lien_adrfilsave *const afs,
       ) {
       // tester type avec requète HEAD si on ne connait pas le type du fichier
       if (!((opt->check_type == 1) && (fil[strlen(fil) - 1] == '/')))   // slash doit être html?
-        if (opt->savename_delayed == 2 || (ishtest = ishtml(opt, fil)) < 0) {   // on ne sait pas si c'est un html ou un fichier..
+        if (opt->savename_delayed == HTS_SAVENAME_DELAYED_HARD ||
+            (ishtest = ishtml(opt, fil)) <
+                0) { // unsure whether it's html or a file
           // lire dans le cache
           htsblk r = cache_read_including_broken(opt, cache, adr, fil); // test uniquement
 
@@ -393,11 +396,12 @@ int url_savename(lien_adrfilsave *const afs,
             }
 #endif
             //
-          } else if (opt->savename_delayed != 2 && is_userknowntype(opt, fil)) {        /* PATCH BY BRIAN SCHRÖDER. 
+          } else if (opt->savename_delayed != HTS_SAVENAME_DELAYED_HARD &&
-                                                                                           Lookup mimetype not only by extension, 
+                     is_userknowntype(opt, fil)) { /* PATCH BY BRIAN SCHRÖDER.
-                                                                                           but also by filename */
+                              Lookup mimetype not only by extension,
-            /* Note: "foo.cgi => text/html" means that foo.cgi shall have the text/html MIME file type,
+                              but also by filename */
-               that is, ".html" */
+            /* Note: "foo.cgi => text/html" means that foo.cgi shall have the
+               text/html MIME file type, that is, ".html" */
             char BIGSTK mime[1024];
 
             mime[0] = ext[0] = '\0';
@@ -408,9 +412,13 @@ int url_savename(lien_adrfilsave *const afs,
               }
             }
           }
-          // note: if savename_delayed is enabled, the naming will be temporary (and slightly invalid!)
+          // note: if savename_delayed is enabled, the naming will be temporary
-          // note: if we are about to stop (opt->state.stop), back_add() will fail later
+          // (and slightly invalid!)
-          else if (opt->savename_delayed != 0 && !opt->state.stop) {
+          //
+          // note: if we are about to stop (opt->state.stop), back_add() will
+          // fail later
+          else if (opt->savename_delayed != HTS_SAVENAME_DELAYED_NONE &&
+                   !opt->state.stop) {
             // Check if the file is ready in backing. We basically take the same logic as later.
             // FIXME: we should cleanup and factorize this unholy mess
             if (headers != NULL && headers->status >= 0 && !is_redirect) {
@@ -698,7 +706,7 @@ int url_savename(lien_adrfilsave *const afs,
             }
             // restaurer
             opt->state._hts_in_html_parsing = hihp;
-          }                     // caché?
+          } // caché?
         }
     }
   }
@@ -1190,7 +1198,8 @@ int url_savename(lien_adrfilsave *const afs,
   // Not used anymore unless non-delayed types.
   // de même en cas de manque d'extension on en place une de manière forcée..
   // cela évite les /chez/toto et les /chez/toto/index.html incompatibles
-  if (opt->savename_type != -1 && opt->savename_delayed != 2) {
+  if (opt->savename_type != -1 &&
+      opt->savename_delayed != HTS_SAVENAME_DELAYED_HARD) {
     char *a = afs->save + strlen(afs->save) - 1;
 
     while((a > afs->save) && (*a != '.') && (*a != '/'))
@@ -1236,31 +1245,21 @@ int url_savename(lien_adrfilsave *const afs,
     size_t i;
     for(i = 0 ; afs->save[i] != '\0' ; i++) {
       unsigned char c = (unsigned char) afs->save[i];
-      if (c < 32      // control
+      if (c < 32       // control
-        || c == 127   // unwise
+          || c == 127  // unwise
-        || c == '~'   // unix unwise
+          || c == '~'  // unix unwise
-        || c == '\\'  // windows separator
+          || c == '\\' // windows separator
-        || c == ':'   // windows forbidden
+          || c == ':'  // windows forbidden
-        || c == '*'   // windows forbidden
+          || c == '*'  // windows forbidden
-        || c == '?'   // windows forbidden
+          || c == '?'  // windows forbidden
-        || c == '\"'  // windows forbidden
+          || c == '\"' // windows forbidden
-        || c == '<'   // windows forbidden
+          || c == '<'  // windows forbidden
-        || c == '>'   // windows forbidden
+          || c == '>'  // windows forbidden
-        || c == '|'   // windows forbidden
+          || c == '|'  // windows forbidden
-        //|| c == '@' // ?
+          //|| c == '@' // ?
-        ||
+          || (opt->savename_83 == HTS_SAVENAME_83_ISO9660 // CDROM
-          (
+              && (c == '-' || c == '=' || c == '+'))) {
-            opt->savename_83 == 2 // CDROM
+        afs->save[i] = '_';
-            &&
-            (
-              c == '-'
-              || c == '='
-              || c == '+'
-            )
-          )
-        )
-      {
-         afs->save[i] = '_';
       }
     }
   }
@@ -1521,7 +1520,8 @@ int url_savename(lien_adrfilsave *const afs,
           char *a = afs->save + strlen(afs->save) - 1;
           char *b;
           int n = 2;
-          char collisionSeparator = ((opt->savename_83 != 2) ? '-' : '_');
+          char collisionSeparator =
+              ((opt->savename_83 != HTS_SAVENAME_83_ISO9660) ? '-' : '_');
 
           tempo[0] = '\0';
 

src/htsopt.h

@@ -342,24 +342,44 @@ typedef enum hts_seeker {
   HTS_SEEKER_UP = 1 << 1    /**< may ascend to parent directories */
 } hts_seeker;
 
-/* Link-following scope, stored in the low byte of opt->travel. */
+/* opt->travel: link-following scope in the low byte, flags OR'd in above it. */
 typedef enum hts_travel_scope {
   HTS_TRAVEL_SAME_ADDRESS = 0, /**< stay on the same address (host) */
   HTS_TRAVEL_SAME_DOMAIN = 1,  /**< stay on the same principal domain */
   HTS_TRAVEL_SAME_TLD = 2,     /**< stay on the same TLD (e.g. .com) */
-  HTS_TRAVEL_EVERYWHERE = 7    /**< follow links anywhere on the web */
+  HTS_TRAVEL_EVERYWHERE = 7,   /**< follow links anywhere on the web */
+  HTS_TRAVEL_TEST_ALL = 1 << 8 /**< also test forbidden URLs (-t) */
 } hts_travel_scope;
 
-/* Flags OR'd into opt->travel above the scope value. */
+/* Mask selecting the scope value out of opt->travel. */
-#define HTS_TRAVEL_SCOPE_MASK 0xff   /**< mask selecting the scope value */
+#define HTS_TRAVEL_SCOPE_MASK 0xff
-#define HTS_TRAVEL_TEST_ALL (1 << 8) /**< also test forbidden URLs (-t) */
 
-/* Boolean option flag. An enum (not C bool) so the option fields stay int-sized
+/* Text progress display detail (opt->verbosedisplay). */
-   and the httrackp layout/ABI is unchanged. */
+typedef enum hts_verbosedisplay {
-#ifndef HTS_DEF_DEFSTRUCT_hts_boolean
+  HTS_VERBOSE_NONE = 0,   /**< no animated progress display (default) */
-#define HTS_DEF_DEFSTRUCT_hts_boolean
+  HTS_VERBOSE_SIMPLE = 1, /**< minimal single-line progress */
-typedef enum hts_boolean { HTS_FALSE = 0, HTS_TRUE = 1 } hts_boolean;
+  HTS_VERBOSE_FULL = 2    /**< full animated progress */
-#endif
+} hts_verbosedisplay;
+
+/* Delayed file-type resolution policy (opt->savename_delayed). */
+typedef enum hts_savename_delayed {
+  HTS_SAVENAME_DELAYED_NONE = 0, /**< resolve the type immediately */
+  HTS_SAVENAME_DELAYED_SOFT = 1, /**< delay the type check when unknown */
+  HTS_SAVENAME_DELAYED_HARD = 2  /**< always delay the type check (default) */
+} hts_savename_delayed;
+
+/* Saved-name length layout (opt->savename_83). */
+typedef enum hts_savename_83 {
+  HTS_SAVENAME_83_LONG = 0,   /**< long file names (default) */
+  HTS_SAVENAME_83_DOS = 1,    /**< DOS 8.3 names (ISO9660 level 1) */
+  HTS_SAVENAME_83_ISO9660 = 2 /**< ISO9660 level 2 names (up to 31 chars) */
+} hts_savename_83;
+
+/* Host-banning triggers (opt->hostcontrol bitmask). */
+typedef enum hts_hostcontrol {
+  HTS_HOSTCONTROL_BAN_TIMEOUT = 1 << 0, /**< ban a timing-out host */
+  HTS_HOSTCONTROL_BAN_SLOW = 1 << 1     /**< ban a too-slow host */
+} hts_hostcontrol;
 
 #ifndef HTS_DEF_FWSTRUCT_lien_buffers
 #define HTS_DEF_FWSTRUCT_lien_buffers
@@ -393,7 +413,7 @@ struct httrackp {
   hts_urlmode
       urlmode; /**< saved-link rewriting style (relative, absolute, etc.) */
   hts_boolean no_type_change;   // do not change file type according to MIME
-  int debug;                    /**< debug logging level */
+  hts_log_type debug;           /**< debug logging level */
   int getmode;           /**< what to fetch (HTML, images, ...) bitmask */
   FILE *log;             /**< informational log stream; NULL mutes it */
   FILE *errlog;          /**< error log stream; NULL mutes it */
@@ -417,11 +437,12 @@ struct httrackp {
   // int aff_progress;     // progress bar
   hts_boolean shell; /**< driven by a shell over stdin/stdout pipes */
   t_proxy proxy;     /**< proxy configuration */
-  int savename_83;   /**< force 8.3 (DOS) file names */
+  hts_savename_83
+      savename_83;   /**< saved-name length layout (long/DOS/ISO9660) */
   int savename_type; /**< saved-name layout (original tree, flat, ...) */
   String
       savename_userdef; /**< user-defined name template (e.g. %h%p/%n%q.%t) */
-  int savename_delayed;         // delayed type check
+  hts_savename_delayed savename_delayed; /**< delayed type-check policy */
   hts_boolean
       delayed_cached;   // delayed type check can be cached to speedup updates
   hts_boolean mimehtml; /**< produce a single MIME/MHTML archive */
@@ -437,7 +458,7 @@ struct httrackp {
   hts_boolean makestat;  /**< maintain a transfer-statistics log */
   hts_boolean maketrack; /**< maintain an operations-statistics log */
   int parsejava;         /**< Java/JS parsing mode; see htsparsejava_flags */
-  int hostcontrol;       /**< drop hosts that are too slow, etc. */
+  int hostcontrol; /**< ban slow/timing-out hosts; see hts_hostcontrol bits */
   hts_boolean errpage;   /**< generate an error page on 404 and similar */
   hts_boolean
       check_type; /**< probe unknown-type links (cgi/asp/dir) and follow moves
@@ -462,7 +483,7 @@ struct httrackp {
       parseall; /**< parse aggressively, including unknown tags with links */
   hts_boolean parsedebug; /**< parser debug mode */
   hts_boolean norecatch;  /**< do not re-fetch files the user deleted locally */
-  int verbosedisplay; /**< animated text progress display */
+  hts_verbosedisplay verbosedisplay; /**< animated text progress display */
   String footer;      /**< footer/info line injected into pages */
   int maxcache;       /**< in-memory cache backing limit (bytes) */
   // int maxcache_anticipate; // maximum links to anticipate (upper bound)

src/htsparse.c

@@ -296,6 +296,48 @@ static const char *html_inline_safe(const char *src, char *dst, size_t size) {
   return dst;
 }
 
+/* Byte before html, or a space sentinel at the buffer start where html[-1]
+   would underflow; space reads as the word boundary the guards want there. */
+static HTS_INLINE char html_prevc(const char *html, const char *start) {
+  return html > start ? html[-1] : ' ';
+}
+
+/* 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) {
+  static const char *const methods[] = {"GET",    "POST",  "PUT",
+                                        "DELETE", "HEAD",  "OPTIONS",
+                                        "PATCH",  "TRACE", NULL};
+  int i;
+
+  for (i = 0; methods[i] != NULL; i++) {
+    if (strlen(methods[i]) == len && strfield(s, methods[i]) == (int) len)
+      return 1;
+  }
+  return 0;
+}
+
+/* Percent-encode '(' and ')' in a link emitted into an unquoted url(...) (CSS
+   or JS): a literal ')' closes the token early and the UA mis-parses the value
+   (#163). The UA decodes %28/%29 back to the saved-on-disk name. */
+static void escape_url_parens(char *const s, const size_t size) {
+  char BIGSTK buff[HTS_URLMAXSIZE * 2];
+  size_t i, j;
+
+  for (i = 0, j = 0; s[i] != '\0' && j + 3 < size && j + 3 < sizeof(buff);
+       i++) {
+    if (s[i] == '(' || s[i] == ')') {
+      buff[j++] = '%';
+      buff[j++] = '2';
+      buff[j++] = s[i] == '(' ? '8' : '9';
+    } else {
+      buff[j++] = s[i];
+    }
+  }
+  buff[j] = '\0';
+  strlcpybuff(s, buff, size);
+}
+
 /* Main parser */
 int htsparse(htsmoduleStruct * str, htsmoduleStructExtended * stre) {
   char catbuff[CATBUFF_SIZE];
@@ -556,7 +598,7 @@ int htsparse(htsmoduleStruct * str, htsmoduleStructExtended * stre) {
                   if (opt->getmode & HTS_GETMODE_HTML) {
                     p = strfield(html, "title");
                     if (p) {
-                      if (*(html - 1) == '/')
+                      if (html_prevc(html, r->adr) == '/')
                         p = 0;  // /title
                     } else {
                       if (strfield(html, "/html"))
@@ -1341,6 +1383,11 @@ int htsparse(htsmoduleStruct * str, htsmoduleStructExtended * stre) {
                     int can_avoid_quotes = 0;
                     char quotes_replacement = '\0';
                     int ensure_not_mime = 0;
+                    // .open(method,url): reject an HTTP-method first arg (#218)
+                    int ensure_not_method = 0;
+                    // @import: the quoted token is the URL; a trailing
+                    // media/supports/layer condition is not part of it
+                    int is_import = 0;
 
                     if (inscript_tag)
                       expected_end = ";\"\'";   // voir a href="javascript:doc.location='foo'"
@@ -1357,9 +1404,8 @@ int htsparse(htsmoduleStruct * str, htsmoduleStructExtended * stre) {
                       if (!nc)
                         nc = strfield(html, ":location");        // javascript:location="doc"
                       if (!nc) {        // location="doc"
-                        if ((nc = strfield(html, "location"))
+                        if ((nc = strfield(html, "location")) &&
-                            && !isspace(*(html - 1))
+                            !isspace(html_prevc(html, r->adr)))
-                          )
                           nc = 0;
                       }
                       if (!nc)
@@ -1369,6 +1415,7 @@ int htsparse(htsmoduleStruct * str, htsmoduleStructExtended * stre) {
                           expected = '(';       // parenthèse
                           expected_end = "),";  // fin: virgule ou parenthèse
                           ensure_not_mime = 1;  //* ensure the url is not a mime type */
+                          ensure_not_method = 1; // xhr.open: don't grab method
                         }
                       if (!nc)
                         if ((nc = strfield(html, ".replace"))) { // window.replace("url")
@@ -1380,7 +1427,9 @@ int htsparse(htsmoduleStruct * str, htsmoduleStructExtended * stre) {
                           expected = '(';       // parenthèse
                           expected_end = ")";   // fin: parenthèse
                         }
-                      if (!nc && (nc = strfield(html, "url")) && (!isalnum(*(html - 1))) && *(html - 1) != '_') {  // url(url)
+                      if (!nc && (nc = strfield(html, "url")) &&
+                          (!isalnum(html_prevc(html, r->adr))) &&
+                          html_prevc(html, r->adr) != '_') { // url(url)
                         expected = '('; // parenthèse
                         expected_end = ")";     // fin: parenthèse
                         can_avoid_quotes = 1;
@@ -1390,6 +1439,7 @@ int htsparse(htsmoduleStruct * str, htsmoduleStructExtended * stre) {
                         if ((nc = strfield(html, "import"))) {   // import "url"
                           if (is_space(*(html + nc))) {
                             expected = 0;       // no char expected
+                            is_import = 1;
                           } else
                             nc = 0;
                         }
@@ -1407,6 +1457,7 @@ int htsparse(htsmoduleStruct * str, htsmoduleStructExtended * stre) {
                           if ((*a == 34) || (*a == '\'') || (can_avoid_quotes)) {
                             const char *b, *c;
                             int ndelim = 1;
+                            int valid_url = 0;
 
                             if ((*a == 34) || (*a == '\''))
                               a++;
@@ -1421,12 +1472,20 @@ int htsparse(htsmoduleStruct * str, htsmoduleStructExtended * stre) {
                                 b++;
                             }
                             c = b--;
-                            c += ndelim;
+                            // no closing delimiter here (truncated input):
-                            while(*c == ' ')
+                            // Don't scan past the buffer NUL or capture it.
-                              c++;
+                            if (*c != '\0') {
-                            if ((strchr(expected_end, *c)) || (*c == '\n')
+                              c += ndelim;
-                                || (*c == '\r')) {
+                              while (*c == ' ')
-                              c -= (ndelim + 1);
+                                c++;
+                              valid_url =
+                                  (strchr(expected_end, *c)) || (*c == '\n') ||
+                                  (*c == '\r') ||
+                                  (is_import && *(b + 1 + ndelim) == ' ');
+                            }
+                            if (valid_url) {
+                              // URL end = last char (b), not the delimiter
+                              c = b;
                               if ((int) (c - a + 1)) {
                                 if (ensure_not_mime) {
                                   int i = 0;
@@ -1442,6 +1501,11 @@ int htsparse(htsmoduleStruct * str, htsmoduleStructExtended * stre) {
                                     i++;
                                   }
                                 }
+                                // XHR.open's "GET" etc. is a method, not a URL
+                                if (a != NULL && ensure_not_method &&
+                                    is_http_method(a, (size_t) (c - a + 1))) {
+                                  a = NULL;
+                                }
                                 // Check for bogus links (Vasiliy)
                                 if (a != NULL) {
                                   const size_t size = c - a + 1;
@@ -1485,7 +1549,6 @@ int htsparse(htsmoduleStruct * str, htsmoduleStructExtended * stre) {
                                 }
                               }
                             }
-
                           }
                         }
                       }
@@ -1692,6 +1755,24 @@ int htsparse(htsmoduleStruct * str, htsmoduleStructExtended * stre) {
                                                               hts_nodetect[i -
                                                                            1]);
                                               }
+                                              // xmlns / xmlns:prefix declare
+                                              // XML namespaces, not resources
+                                              // (#191)
+                                              else {
+                                                const int xl = strfield(
+                                                    intag_startattr, "xmlns");
+                                                const char xc =
+                                                    intag_startattr[xl];
+                                                if (xl &&
+                                                    (xc == ':' || xc == '=' ||
+                                                     is_space(xc))) {
+                                                  url_ok = 0;
+                                                  hts_log_print(
+                                                      opt, LOG_DEBUG,
+                                                      "dirty parsing: xmlns "
+                                                      "namespace avoided");
+                                                }
+                                              }
                                             }
                                   }
 
@@ -2967,6 +3048,10 @@ int htsparse(htsmoduleStruct * str, htsmoduleStructExtended * stre) {
                           /* Never escape high-chars (we don't know the encoding!!) */
                           inplace_escape_uri_utf(tempo, sizeof(tempo));
 
+                          // unquoted url() (CSS/JS): keep parens escaped
+                          if (ending_p == ')')
+                            escape_url_parens(tempo, sizeof(tempo));
+
                           //if (!no_esc_utf)
                           //  escape_uri(tempo);     // escape with %xx
                           //else {
@@ -3722,7 +3807,8 @@ int hts_mirror_check_moved(htsmoduleStruct * str,
         //case -1: can_retry=1; break;
       case STATUSCODE_TIMEOUT:
         if (opt->hostcontrol) { // timeout et retry épuisés
-          if ((opt->hostcontrol & 1) && (heap(ptr)->retry <= 0)) {
+          if ((opt->hostcontrol & HTS_HOSTCONTROL_BAN_TIMEOUT) &&
+              (heap(ptr)->retry <= 0)) {
             hts_log_print(opt, LOG_DEBUG, "Link banned: %s%s", urladr(), urlfil());
             host_ban(opt, ptr, sback, jump_identification_const(urladr()));
             hts_log_print(opt, LOG_DEBUG,
@@ -3735,7 +3821,7 @@ int hts_mirror_check_moved(htsmoduleStruct * str,
         break;
       case STATUSCODE_SLOW:
         if ((opt->hostcontrol) && (heap(ptr)->retry <= 0)) {   // too slow
-          if (opt->hostcontrol & 2) {
+          if (opt->hostcontrol & HTS_HOSTCONTROL_BAN_SLOW) {
             hts_log_print(opt, LOG_DEBUG, "Link banned: %s%s", urladr(), urlfil());
             host_ban(opt, ptr, sback, jump_identification_const(urladr()));
             hts_log_print(opt, LOG_DEBUG,
@@ -4261,10 +4347,10 @@ int hts_mirror_wait_for_next_file(htsmoduleStruct * str,
             char com[256];
 
             linput(stdin, com, 200);
-            if (opt->verbosedisplay == 2)
+            if (opt->verbosedisplay == HTS_VERBOSE_FULL)
-              opt->verbosedisplay = 1;
+              opt->verbosedisplay = HTS_VERBOSE_SIMPLE;
             else
-              opt->verbosedisplay = 2;
+              opt->verbosedisplay = HTS_VERBOSE_FULL;
             /* Info for wrappers */
             hts_log_print(opt, LOG_INFO, "engine: change-options");
             RUN_CALLBACK0(opt, chopt);
@@ -4374,7 +4460,7 @@ int hts_mirror_wait_for_next_file(htsmoduleStruct * str,
         printf("%c\x0d", ("/-\\|")[roll]);
         fflush(stdout);
       }
-    } else if (opt->verbosedisplay == 1) {
+    } else if (opt->verbosedisplay == HTS_VERBOSE_SIMPLE) {
       if (b >= 0) {
         if (back[b].r.statuscode == HTTP_OK)
           printf("%d/%d: %s%s (" LLintP " bytes) - OK\33[K\r", ptr, opt->lien_tot,
@@ -4465,8 +4551,8 @@ int hts_wait_delayed(htsmoduleStruct * str, lien_adrfilsave *afs,
   char in_error_msg[32];
 
   // resolve unresolved type
-  if (opt->savename_delayed != 0 && *forbidden_url == 0 && IS_DELAYED_EXT(afs->save)
+  if (opt->savename_delayed != HTS_SAVENAME_DELAYED_NONE &&
-      && !opt->state.stop) {
+      *forbidden_url == 0 && IS_DELAYED_EXT(afs->save) && !opt->state.stop) {
     int loops;
     int continue_loop;
 
@@ -4850,7 +4936,7 @@ int hts_wait_delayed(htsmoduleStruct * str, lien_adrfilsave *afs,
       }
     }
 
-  }                             // delayed type check ?
+  } // delayed type check ?
 
   ENGINE_SAVE_CONTEXT_BASE();
 

src/htstools.c

@@ -1213,7 +1213,7 @@ HTSEXT_API find_handle hts_findfirst(char *path) {
   return NULL;
 }
 
-HTSEXT_API int hts_findnext(find_handle find) {
+HTSEXT_API hts_boolean hts_findnext(find_handle find) {
   if (find) {
 #ifdef _WIN32
     if ((FindNextFileA(find->handle, &find->hdata)))
@@ -1273,7 +1273,7 @@ HTSEXT_API int hts_findgetsize(find_handle find) {
   return -1;
 }
 
-HTSEXT_API int hts_findisdir(find_handle find) {
+HTSEXT_API hts_boolean hts_findisdir(find_handle find) {
   if (find) {
     if (!hts_findissystem(find)) {
 #ifdef _WIN32
@@ -1287,7 +1287,7 @@ HTSEXT_API int hts_findisdir(find_handle find) {
   }
   return 0;
 }
-HTSEXT_API int hts_findisfile(find_handle find) {
+HTSEXT_API hts_boolean hts_findisfile(find_handle find) {
   if (find) {
     if (!hts_findissystem(find)) {
 #ifdef _WIN32
@@ -1301,7 +1301,7 @@ HTSEXT_API int hts_findisfile(find_handle find) {
   }
   return 0;
 }
-HTSEXT_API int hts_findissystem(find_handle find) {
+HTSEXT_API hts_boolean hts_findissystem(find_handle find) {
   if (find) {
 #ifdef _WIN32
     if (find->hdata.

src/htstools.h

@@ -108,15 +108,15 @@ HTSEXT_API int hts_buildtopindex(httrackp * opt, const char *path,
 // Portable directory find functions
 // Directory find functions
 HTSEXT_API find_handle hts_findfirst(char *path);
-HTSEXT_API int hts_findnext(find_handle find);
+HTSEXT_API hts_boolean hts_findnext(find_handle find);
 HTSEXT_API int hts_findclose(find_handle find);
 
 //
 HTSEXT_API char *hts_findgetname(find_handle find);
 HTSEXT_API int hts_findgetsize(find_handle find);
-HTSEXT_API int hts_findisdir(find_handle find);
+HTSEXT_API hts_boolean hts_findisdir(find_handle find);
-HTSEXT_API int hts_findisfile(find_handle find);
+HTSEXT_API hts_boolean hts_findisfile(find_handle find);
-HTSEXT_API int hts_findissystem(find_handle find);
+HTSEXT_API hts_boolean hts_findissystem(find_handle find);
 
 #endif
 

src/httrack-library.h

@@ -206,7 +206,8 @@ HTSEXT_API htsErrorCallback hts_get_error_callback(void);
 /* Logging */
 /** Legacy: write prefix then msg to opt->log. Returns 0 if written, 1 if
     opt->log is NULL. Prefer hts_log_print(). */
-HTSEXT_API int hts_log(httrackp * opt, const char *prefix, const char *msg);
+HTSEXT_API hts_boolean hts_log(httrackp *opt, const char *prefix,
+                               const char *msg);
 
 /** printf-style log at level @p type (an hts_log_type, optionally |LOG_ERRNO).
     Forwards to the registered log callback, and when the level is <= opt->debug
@@ -313,7 +314,8 @@ HTSEXT_API T_SOC catch_url_init(int *port, char *adr);
     "ip:port". The buffers are caller-allocated and not bounds-checked: @p data
     must be CATCH_URL_DATA_SIZE bytes, and @p url / @p method must fit the
     captured request line. */
-HTSEXT_API int catch_url(T_SOC soc, char *url, char *method, char *data);
+HTSEXT_API hts_boolean catch_url(T_SOC soc, char *url, char *method,
+                                 char *data);
 
 /* State */
 /** Whether the engine is parsing HTML. Returns 0 if not, otherwise the percent
@@ -334,10 +336,10 @@ HTSEXT_API int hts_is_exiting(httrackp * opt);
     caller-owned, NULL-terminated array of strings; the engine stores the
    pointer without copying, so the array and its strings must stay valid until
    the engine consumes them. @return nonzero if a list is now set. */
-HTSEXT_API int hts_addurl(httrackp * opt, char **url);
+HTSEXT_API hts_boolean hts_addurl(httrackp *opt, char **url);
 
 /** Clear any pending add-URL list set by hts_addurl(). Always returns 0. */
-HTSEXT_API int hts_resetaddurl(httrackp * opt);
+HTSEXT_API hts_boolean hts_resetaddurl(httrackp *opt);
 
 /** Apply the runtime-tunable options from @p from onto @p to, to adjust a live
     mirror. Only fields set to a non-sentinel value are copied; the rest of @p
@@ -356,7 +358,7 @@ HTSEXT_API int hts_setpause(httrackp * opt, int);
    lock, so it is safe to call from another thread). @p force is currently
    ignored.
     @return 0; no-op if @p opt is NULL. */
-HTSEXT_API int hts_request_stop(httrackp * opt, int force);
+HTSEXT_API int hts_request_stop(httrackp *opt, hts_boolean force);
 
 /** Queue a single in-progress file, by URL, to be cancelled by the engine.
     @p url is copied internally. Takes the state lock, so it is thread-safe.
@@ -373,7 +375,7 @@ HTSEXT_API void hts_cancel_parsing(httrackp * opt);
 
 /** Nonzero once the mirror has fully ended. Read under the engine state lock,
    so safe to poll from another thread. Wait for this before hts_free_opt(). */
-HTSEXT_API int hts_has_stopped(httrackp * opt);
+HTSEXT_API hts_boolean hts_has_stopped(httrackp *opt);
 
 /* Tools */
 /** Ensure the directory chain leading to @p path exists, creating missing
@@ -390,7 +392,7 @@ HTSEXT_API int structcheck_utf8(const char *path);
 /** Whether the directory containing @p path exists. The basename is stripped
     first, so passing a file path tests its parent directory. @return 1 if it is
     a directory, 0 otherwise. */
-HTSEXT_API int dir_exists(const char *path);
+HTSEXT_API hts_boolean dir_exists(const char *path);
 
 /** Write the HTTP reason phrase for @p statuscode into @p msg, a caller buffer
    of at least 64 bytes. For an unknown code a non-empty @p msg is kept,
@@ -573,14 +575,15 @@ HTSEXT_API char *unescape_http(char *const catbuff, const size_t size, const cha
     must-avoid escapes are kept encoded, and %25 is never decoded). @p no_high &
    1 also decodes high (>= 128) bytes; @p no_high & 2 also decodes an escaped
     space. Returns @p catbuff. */
-HTSEXT_API char *unescape_http_unharm(char *const catbuff, const size_t size, const char *s, const int no_high);
+HTSEXT_API char *unescape_http_unharm(char *const catbuff, const size_t size,
+                                      const char *s, const hts_boolean no_high);
 
 /** Determine the MIME type of local file name @p fil into @p s (capacity
     @p ssize): user --assume rules, then ".html", then the built-in extension
     table. @p flag != 0 forces a fallback type. @return 1 if a type was written,
     0 otherwise. */
-HTSEXT_API int get_httptype_sized(httrackp *opt, char *s, size_t ssize,
+HTSEXT_API hts_boolean get_httptype_sized(httrackp *opt, char *s, size_t ssize,
-                                  const char *fil, int flag);
+                                          const char *fil, hts_boolean flag);
 
 /** @deprecated Use get_httptype_sized(). Assumes @p s has at least
     HTS_MIMETYPE_SIZE capacity. */
@@ -600,7 +603,7 @@ HTSEXT_API int is_userknowntype(httrackp * opt, const char *fil);
 
 /** 1 if @p fil, an extension such as "asp" or "php" (not a full filename), is a
     known dynamic-page type, else 0. */
-HTSEXT_API int is_dyntype(const char *fil);
+HTSEXT_API hts_boolean is_dyntype(const char *fil);
 
 /** Extract the extension of @p fil (text after the last '.', stopping at '?')
     into caller scratch @p catbuff (capacity @p size) and return it. Returns ""
@@ -610,12 +613,12 @@ HTSEXT_API const char *get_ext(char *catbuff, size_t size, const char *fil);
 
 /** 1 if MIME type @p st must not be reclassified or renamed (hypertext types
    and a built-in keep-list of commonly mislabeled types), else 0. */
-HTSEXT_API int may_unknown(httrackp * opt, const char *st);
+HTSEXT_API hts_boolean may_unknown(httrackp *opt, const char *st);
 
 /** Guess the MIME type of local file @p fil into @p s (capacity @p ssize),
     always producing a type. @return 1 if a type was written. */
-HTSEXT_API int guess_httptype_sized(httrackp *opt, char *s, size_t ssize,
+HTSEXT_API hts_boolean guess_httptype_sized(httrackp *opt, char *s,
-                                    const char *fil);
+                                            size_t ssize, const char *fil);
 
 /** @deprecated Use guess_httptype_sized(). Assumes @p s has at least
     HTS_MIMETYPE_SIZE capacity. */
@@ -677,7 +680,7 @@ HTSEXT_API find_handle hts_findfirst(char *path);
 
 /** Advance to the next directory entry. Returns 1 if an entry is available, 0
    at end of directory. */
-HTSEXT_API int hts_findnext(find_handle find);
+HTSEXT_API hts_boolean hts_findnext(find_handle find);
 
 /** Close the iteration and free @p find. Always returns 0; NULL is accepted. */
 HTSEXT_API int hts_findclose(find_handle find);
@@ -692,16 +695,16 @@ HTSEXT_API int hts_findgetsize(find_handle find);
 
 /** 1 if the current entry is a directory, else 0 (a system/special entry, see
     hts_findissystem(), reports 0). */
-HTSEXT_API int hts_findisdir(find_handle find);
+HTSEXT_API hts_boolean hts_findisdir(find_handle find);
 
 /** 1 if the current entry is a regular file, else 0 (a system/special entry,
    see hts_findissystem(), reports 0). */
-HTSEXT_API int hts_findisfile(find_handle find);
+HTSEXT_API hts_boolean hts_findisfile(find_handle find);
 
 /** 1 if the current entry is a special/system entry to skip: "." or "..", on
     POSIX also device/fifo/socket nodes, on Windows also system, hidden or
     temporary entries. Else 0. */
-HTSEXT_API int hts_findissystem(find_handle find);
+HTSEXT_API hts_boolean hts_findissystem(find_handle find);
 
 /* UTF-8 aware FILE API */
 /* On non-Windows these macros resolve directly to the POSIX calls. On Windows

src/httrack.c

@@ -288,7 +288,7 @@ static void __cdecl htsshow_uninit(t_hts_callbackarg * carg) {
 }
 static int __cdecl htsshow_start(t_hts_callbackarg * carg, httrackp * opt) {
   use_show = 0;
-  if (opt->verbosedisplay == 2) {
+  if (opt->verbosedisplay == HTS_VERBOSE_FULL) {
     use_show = 1;
     vt_clear();
   }
@@ -852,7 +852,7 @@ static void sig_doback(int blind) {     // mettre en backing
   if (global_opt != NULL) {
     // suppress logging and asking lousy questions
     global_opt->quiet = 1;
-    global_opt->verbosedisplay = 0;
+    global_opt->verbosedisplay = HTS_VERBOSE_NONE;
   }
 
   if (!blind)

src/webhttrack

@@ -4,131 +4,140 @@
 # Initializes the htsserver GUI frontend and launch the default browser
 
 BROWSEREXE=
-SRCHBROWSEREXE="x-www-browser www-browser iceape mozilla firefox-developer-edition firefox icecat iceweasel abrowser firebird galeon konqueror midori opera google-chrome chrome chromium chromium-browser netscape firefox-developer-edition"
+SRCHBROWSEREXE=(x-www-browser www-browser iceape mozilla firefox-developer-edition firefox icecat iceweasel abrowser firebird galeon konqueror midori opera google-chrome chrome chromium chromium-browser netscape firefox-developer-edition)
+# shellcheck disable=SC2153 # BROWSER is the standard freedesktop env var, not a typo
 if test -n "${BROWSER}"; then
-# sensible-browser will f up if BROWSER is not set
+    # sensible-browser will f up if BROWSER is not set
-SRCHBROWSEREXE="xdg-open sensible-browser ${SRCHBROWSEREXE}"
+    SRCHBROWSEREXE=(xdg-open sensible-browser "${SRCHBROWSEREXE[@]}")
 fi
 # Patch for Darwin/Mac by Ross Williams
-if test "`uname -s`" == "Darwin"; then
+if test "$(uname -s)" == "Darwin"; then
-# Darwin/Mac OS X uses a system 'open' command to find
+    # Darwin/Mac OS X uses a system 'open' command to find
-# the default browser. The -W flag causes it to wait for
+    # the default browser. The -W flag causes it to wait for
-# the browser to exit
+    # the browser to exit
-BROWSEREXE="/usr/bin/open -W"
+    BROWSEREXE="/usr/bin/open -W"
 fi
-BINWD=`dirname "$0"`
+BINWD=$(dirname "$0")
-SRCHPATH="$BINWD /usr/local/bin /usr/share/bin /usr/bin /usr/lib/httrack /usr/local/lib/httrack /usr/local/share/httrack /opt/local/bin /sw/bin ${HOME}/usr/bin ${HOME}/bin"
+SRCHPATH=("$BINWD" /usr/local/bin /usr/share/bin /usr/bin /usr/lib/httrack /usr/local/lib/httrack /usr/local/share/httrack /opt/local/bin /sw/bin "${HOME}/usr/bin" "${HOME}/bin")
-SRCHPATH="$SRCHPATH "`echo $PATH | tr ":" " "`
+IFS=':' read -ra pathdirs <<<"$PATH"
-SRCHDISTPATH="$BINWD/../share $BINWD/.. /usr/share /usr/local /usr /local /usr/local/share ${HOME}/usr ${HOME}/usr/share /opt/local/share /sw ${HOME}/usr/local ${HOME}/usr/share"
+for d in "${pathdirs[@]}"; do
+    # drop empty PATH fields, matching the old echo|tr word-split
+    test -n "$d" && SRCHPATH+=("$d")
+done
+SRCHDISTPATH=("$BINWD/../share" "$BINWD/.." /usr/share /usr/local /usr /local /usr/local/share "${HOME}/usr" "${HOME}/usr/share" /opt/local/share /sw "${HOME}/usr/local" "${HOME}/usr/share")
 
 ###
 # And now some famous cuisine
 
 function log {
-echo "$0($$): $@" >&2
+    echo "$0($$): $*" >&2
-return 0
+    return 0
 }
 
 function launch_browser {
-log "Launching $1"
+    log "Launching $1"
-browser=$1
+    browser=$1
-url=$2
+    url=$2
-log "Spawning browser.."
+    log "Spawning browser.."
-${browser} "${url}"
+    ${browser} "${url}"
-# note: browser can hiddenly use the -remote feature of
+    # note: browser can hiddenly use the -remote feature of
-# mozilla and therefore return immediately
+    # mozilla and therefore return immediately
-log "Browser (or helper) exited"
+    log "Browser (or helper) exited"
 }
 
 # First ensure that we can launch the server
 BINPATH=
-for i in ${SRCHPATH}; do
+for i in "${SRCHPATH[@]}"; do
-	! test -n "${BINPATH}" && test -x ${i}/htsserver && BINPATH=${i}
+    ! test -n "${BINPATH}" && test -x "${i}/htsserver" && BINPATH="${i}"
 done
-for i in ${SRCHDISTPATH}; do
+for i in "${SRCHDISTPATH[@]}"; do
-	! test -n "${DISTPATH}" && test -f "${i}/httrack/lang.def" && DISTPATH="${i}/httrack"
+    ! test -n "${DISTPATH}" && test -f "${i}/httrack/lang.def" && DISTPATH="${i}/httrack"
 done
 test -n "${BINPATH}" || ! log "Could not find htsserver" || exit 1
 test -n "${DISTPATH}" || ! log "Could not find httrack directory" || exit 1
-test -f ${DISTPATH}/lang.def || ! log "Could not find ${DISTPATH}/lang.def" || exit 1
+test -f "${DISTPATH}/lang.def" || ! log "Could not find ${DISTPATH}/lang.def" || exit 1
-test -f ${DISTPATH}/lang.indexes || ! log "Could not find ${DISTPATH}/lang.indexes" || exit 1
+test -f "${DISTPATH}/lang.indexes" || ! log "Could not find ${DISTPATH}/lang.indexes" || exit 1
-test -d ${DISTPATH}/lang || ! log "Could not find ${DISTPATH}/lang" || exit 1
+test -d "${DISTPATH}/lang" || ! log "Could not find ${DISTPATH}/lang" || exit 1
-test -d ${DISTPATH}/html || ! log "Could not find ${DISTPATH}/html" || exit 1
+test -d "${DISTPATH}/html" || ! log "Could not find ${DISTPATH}/html" || exit 1
 
 # Locale
 HTSLANG="${LC_MESSAGES}"
 ! test -n "${HTSLANG}" && HTSLANG="${LC_ALL}"
 ! test -n "${HTSLANG}" && HTSLANG="${LANG}"
-HTSLANG="`echo $LANG | cut -f1 -d'.' | cut -f1 -d'_'`"
+HTSLANG="$(echo "$LANG" | cut -f1 -d'.' | cut -f1 -d'_')"
-LANGN=`grep -E "^${HTSLANG}:" ${DISTPATH}/lang.indexes | cut -f2 -d':'`
+LANGN=$(grep -E "^${HTSLANG}:" "${DISTPATH}/lang.indexes" | cut -f2 -d':')
 ! test -n "${LANGN}" && LANGN=1
 
 # Find the browser
 # note: not all systems have sensible-browser or www-browser alternative
 # thefeore, we have to find a bit more if sensible-browser could not be found
 
-for i in ${SRCHBROWSEREXE}; do
+for i in "${SRCHBROWSEREXE[@]}"; do
-for j in ${SRCHPATH}; do
+    for j in "${SRCHPATH[@]}"; do
-if test -x ${j}/${i}; then
+        if test -x "${j}/${i}"; then
-BROWSEREXE=${j}/${i}
+            BROWSEREXE="${j}/${i}"
-fi
+        fi
-test -n "$BROWSEREXE" && break
+        test -n "$BROWSEREXE" && break
-done
+    done
-test -n "$BROWSEREXE" && break
+    test -n "$BROWSEREXE" && break
 done
 test -n "$BROWSEREXE" || ! log "Could not find any suitable browser" || exit 1
 
 # "browse" command
 if test "$1" = "browse"; then
-if test -f "${HOME}/.httrack.ini"; then
+    if test -f "${HOME}/.httrack.ini"; then
-INDEXF=`cat ${HOME}/.httrack.ini | tr '\r' '\n' | grep -E "^path=" | cut -f2- -d'='`
+        INDEXF=$(tr '\r' '\n' <"${HOME}/.httrack.ini" | grep -E "^path=" | cut -f2- -d'=')
-if test -n "${INDEXF}" -a -d "${INDEXF}" -a -f "${INDEXF}/index.html"; then
+        if test -n "${INDEXF}" -a -d "${INDEXF}" -a -f "${INDEXF}/index.html"; then
-INDEXF="${INDEXF}/index.html"
+            INDEXF="${INDEXF}/index.html"
-else
+        else
-INDEXF=""
+            INDEXF=""
-fi
+        fi
-fi
+    fi
-if ! test -n "$INDEXF"; then 
+    if ! test -n "$INDEXF"; then
-INDEXF="${HOME}/websites/index.html"
+        INDEXF="${HOME}/websites/index.html"
-fi
+    fi
-launch_browser "${BROWSEREXE}" "file://${INDEXF}"
+    launch_browser "${BROWSEREXE}" "file://${INDEXF}"
-exit $?
+    exit $?
 fi
 
 # Create a temporary filename
-TMPSRVFILE="$(mktemp ${TMPDIR:-/tmp}/.webhttrack.XXXXXXXX)" || ! log "Could not create the temporary file ${TMPSRVFILE}" || exit 1
+TMPSRVFILE="$(mktemp "${TMPDIR:-/tmp}/.webhttrack.XXXXXXXX")" || ! log "Could not create the temporary file ${TMPSRVFILE}" || exit 1
 # Launch htsserver binary and setup the server
-(${BINPATH}/htsserver "${DISTPATH}/" --ppid "$$" path "${HOME}/websites" lang "${LANGN}" $@; echo SRVURL=error) > ${TMPSRVFILE}&
+(
+    "${BINPATH}/htsserver" "${DISTPATH}/" --ppid "$$" path "${HOME}/websites" lang "${LANGN}" "$@"
+    echo SRVURL=error
+) >"${TMPSRVFILE}" &
 # Find the generated SRVURL
 SRVURL=
 MAXCOUNT=60
 while ! test -n "$SRVURL"; do
-MAXCOUNT=$[$MAXCOUNT - 1]
+    MAXCOUNT=$((MAXCOUNT - 1))
-test $MAXCOUNT -gt 0 || exit 1
+    test $MAXCOUNT -gt 0 || exit 1
-test $MAXCOUNT -lt 50 && echo "waiting for server to reply.."
+    test $MAXCOUNT -lt 50 && echo "waiting for server to reply.."
-SRVURL=`grep -E URL= ${TMPSRVFILE} | cut -f2- -d=`
+    SRVURL=$(grep -E URL= "${TMPSRVFILE}" | cut -f2- -d=)
-test ! "$SRVURL" = "error" || ! log "Could not spawn htsserver" || exit 1
+    test ! "$SRVURL" = "error" || ! log "Could not spawn htsserver" || exit 1
-test -n "$SRVURL" || sleep 1
+    test -n "$SRVURL" || sleep 1
 done
 
 # Cleanup function
+# shellcheck disable=SC2120 # $1 is an optional "signal caught" marker; bare calls are intentional
 function cleanup {
-test -n "$1" && log "Nasty signal caught, cleaning up.."
+    test -n "$1" && log "Nasty signal caught, cleaning up.."
-# Do not kill if browser exited (chrome bug issue) ; server will die itself
+    # Do not kill if browser exited (chrome bug issue) ; server will die itself
-test -n "$1" && test -f ${TMPSRVFILE} && SRVPID=`grep -E PID= ${TMPSRVFILE} | cut -f2- -d=`
+    test -n "$1" && test -f "${TMPSRVFILE}" && SRVPID=$(grep -E PID= "${TMPSRVFILE}" | cut -f2- -d=)
-test -n "${SRVPID}" && kill -9 ${SRVPID}
+    test -n "${SRVPID}" && kill -9 "${SRVPID}"
-test -f ${TMPSRVFILE} && rm ${TMPSRVFILE}
+    test -f "${TMPSRVFILE}" && rm "${TMPSRVFILE}"
-test -n "$1" && log "..Done"
+    test -n "$1" && log "..Done"
-return 0
+    return 0
 }
 
 # Cleanup in case of emergency
-trap "cleanup now; exit" 1 2 3 4 5 6 7 8 9 11 13 14 15 16 19 24 25
+trap "cleanup now; exit" HUP INT QUIT ILL TRAP ABRT BUS FPE SEGV PIPE ALRM TERM STKFLT XCPU XFSZ
 
 # Got SRVURL, launch browser
 launch_browser "${BROWSEREXE}" "${SRVURL}"
 
 # That's all, folks!
-trap "" 1 2 3 4 5 6 7 8 9 11 13 14 15 16 19 24 25
+trap "" HUP INT QUIT ILL TRAP ABRT BUS FPE SEGV PIPE ALRM TERM STKFLT XCPU XFSZ
 cleanup
 exit 0

tests/01_engine-charset.test

@@ -6,11 +6,11 @@ set -euo pipefail
 # charset -> UTF-8 conversion (hts_convertStringToUTF8).
 # -#3 <charset> <string> prints the string re-decoded from <charset> as UTF-8.
 conv() {
-	test "$(httrack -O /dev/null -#3 "$1" "$2")" == "$3" || exit 1
+    test "$(httrack -O /dev/null -#3 "$1" "$2")" == "$3" || exit 1
 }
 # crash probe: malformed input must exit cleanly, not abort.
 runs() {
-	httrack -O /dev/null -#3 "$1" "$2" >/dev/null 2>&1 || exit 1
+    httrack -O /dev/null -#3 "$1" "$2" >/dev/null 2>&1 || exit 1
 }
 
 # the source bytes below are UTF-8 (this file is UTF-8); "café" is 0x63 61 66 C3 A9.

tests/01_engine-cookies.test

@@ -0,0 +1,15 @@
+#!/bin/bash
+#
+# Issue #151 guard: the request Cookie header must be bare RFC 6265 name=value
+# pairs, no $Version/$Path attributes. Driven by the 'httrack -#Q' selftest.
+
+set -eu
+
+# A trailing token is required; a bare '-#Q' falls through to the usage screen.
+out=$(httrack -#Q run)
+
+# Exact-match the success line so a fall-through to usage can't pass the test.
+test "$out" = "cookie-header: OK" || {
+    echo "expected 'cookie-header: OK', got: $out" >&2
+    exit 1
+}

tests/01_engine-copyopt.test

@@ -0,0 +1,17 @@
+#!/bin/bash
+#
+# Regression guard for the unsigned-enum sentinel trap: copy_htsopt's
+# `if (from->X > -1)` guard is always false for unsigned hts_boolean fields, so
+# they silently stop being copied. Driven by the in-process 'httrack -#9' test.
+# Keep POSIX-portable (harness runs it via $(BASH), a plain /bin/sh on macOS).
+
+set -eu
+
+# A trailing token is required; a bare '-#9' falls through to the usage screen.
+out=$(httrack -#9 run)
+
+# Exact-match the success line so a fall-through to usage can't pass the test.
+test "$out" = "copy-htsopt: OK" || {
+    echo "expected 'copy-htsopt: OK', got: $out" >&2
+    exit 1
+}

tests/01_engine-doitlog.test

@@ -89,4 +89,37 @@ grep -q NEWCONTENT "$(find "$out" -path '*/a.html' -print -quit)" || {
     exit 1
 }
 
+# --- 3. an empty quoted arg survives the doit.log round-trip (#106) ----------
+# -%F "" (empty footer) records an empty "" token in doit.log; -r2 follows it so
+# a "drop the empty token" bug shifts -r2 into -%F's slot (the reprise then sees
+# -%F -r2 and panics "%F needs to be followed by ..."), making the bug visible
+# rather than a harmless run off the end of argv.
+out2="$tmp/out2"
+rc=0
+"$bin" "$url" -O "$out2" --quiet -n -%v0 -%F "" -r2 >/dev/null 2>&1 || rc=$?
+test "$rc" -eq 0 || {
+    echo "FAIL: initial mirror with empty footer exited $rc"
+    exit 1
+}
+# precondition: the writer put the empty token on disk for the reader to reload.
+grep -q ' -%F "" -r2' "$out2/hts-cache/doit.log" || {
+    echo "FAIL: empty footer not recorded as -%F \"\" -r2 in doit.log"
+    grep -- '-%F' "$out2/hts-cache/doit.log" || true
+    exit 1
+}
+# no-url reprise: the reader rebuilds argv from doit.log and rewrites doit.log
+# from it. The empty token surviving in the regenerated file proves the reader
+# kept it (a drop/swallow would panic above or rewrite -%F without the "").
+rc=0
+"$bin" -O "$out2" --quiet >/dev/null 2>&1 || rc=$?
+test "$rc" -eq 0 || {
+    echo "FAIL: empty-footer reprise exited $rc (empty token dropped from doit.log?)"
+    exit 1
+}
+grep -q ' -%F "" -r2' "$out2/hts-cache/doit.log" || {
+    echo "FAIL: empty footer did not survive the doit.log reload round-trip"
+    grep -- '-%F' "$out2/hts-cache/doit.log" || true
+    exit 1
+}
+
 exit 0

tests/01_engine-entities.test

@@ -6,11 +6,11 @@ set -euo pipefail
 # HTML entity unescaping (hts_unescapeEntitiesWithCharset).
 # -#6 <string> prints the string with entities decoded (UTF-8 output).
 ent() {
-	test "$(httrack -O /dev/null -#6 "$1")" == "$2" || exit 1
+    test "$(httrack -O /dev/null -#6 "$1")" == "$2" || exit 1
 }
 # crash probe: malformed input must exit cleanly, not abort.
 runs() {
-	httrack -O /dev/null -#6 "$1" >/dev/null 2>&1 || exit 1
+    httrack -O /dev/null -#6 "$1" >/dev/null 2>&1 || exit 1
 }
 
 # named entities

tests/01_engine-filter.test

@@ -7,10 +7,10 @@ set -euo pipefail
 # -#0 <filter> <string> prints "<string> does match <filter>" or "... does NOT match ...".
 
 match() {
-	test "$(httrack -O /dev/null -#0 "$1" "$2")" == "$2 does match $1" || exit 1
+    test "$(httrack -O /dev/null -#0 "$1" "$2")" == "$2 does match $1" || exit 1
 }
 nomatch() {
-	test "$(httrack -O /dev/null -#0 "$1" "$2")" == "$2 does NOT match $1" || exit 1
+    test "$(httrack -O /dev/null -#0 "$1" "$2")" == "$2 does NOT match $1" || exit 1
 }
 
 # bare star matches everything
@@ -67,7 +67,7 @@ nomatch '*[\[]' 'a'
 # 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
+nomatch '*[\[\]]' '[' # not matched, despite the docs
-match '*[\[\]]' ']'     # only via the empty class-match + trailing ']'
+match '*[\[\]]' ']'   # only via the empty class-match + trailing ']'
-match '*[\[\]]' '[]'    # one of {'[','\'} then the trailing ']'
+match '*[\[\]]' '[]'  # one of {'[','\'} then the trailing ']'
 nomatch '*[\[\]]' '[]x'

tests/01_engine-mime.test

@@ -7,10 +7,10 @@ set -euo pipefail
 # -#2 <path> prints "<path> is '<mime>'" then "and its local type is '.<ext>'".
 
 mime() {
-	test "$(httrack -O /dev/null -#2 "$1" | head -1)" == "$1 is '$2'" || exit 1
+    test "$(httrack -O /dev/null -#2 "$1" | head -1)" == "$1 is '$2'" || exit 1
 }
 unknown() {
-	test "$(httrack -O /dev/null -#2 "$1" | head -1)" == "$1 is of an unknown MIME type" || exit 1
+    test "$(httrack -O /dev/null -#2 "$1" | head -1)" == "$1 is of an unknown MIME type" || exit 1
 }
 
 mime '/a/b.html' 'text/html'

tests/01_engine-parse.test

@@ -154,4 +154,173 @@ grep -Eq "style=\"background-image:url\('ibgs\.gif'\)\"" "$saved2" ||
 grep -q 'title="file://' "$saved2" ||
     ! echo "FAIL: a no-detect attribute (title) was wrongly rewritten" || exit 1
 
+# xmlns / xmlns:prefix decls must not be crawled (#191). Local file:// targets so a
+# regression downloads them; each is the LAST attr (heuristic only scans a value before '>').
+site3="$tmp/xmlns"
+mkdir -p "$site3"
+for f in ns og rdfs real; do gif "$site3/$f.gif"; done
+cat >"$site3/index.html" <<EOF
+<html xmlns="file://$site3/ns.gif"><body>
+<svg xmlns:og="file://$site3/og.gif"></svg>
+<div class="c" xmlns:rdfs="file://$site3/rdfs.gif"></div>
+<a href="file://$site3/real.gif">real link</a>
+</body></html>
+EOF
+out3="$tmp/xmlns-out"
+crawl "$site3/index.html" "$out3"
+
+# the real link is still captured
+found "real.gif" "$out3"
+# namespace-declaration targets must not be fetched (default + prefixed forms)
+notfound "ns.gif" "$out3"
+notfound "og.gif" "$out3"
+notfound "rdfs.gif" "$out3"
+
+# CSS @import (#94): every form's target is captured, crawling the .css directly.
+# The "cond"/"sup"/"spc" cases carry a trailing media/supports/layer condition (or
+# a space before ';'); they are the negative controls: without the parser fix the
+# URL is dropped, so a regression fails these found() checks.
+site4="$tmp/cssimport"
+mkdir -p "$site4"
+for f in nq dqu squ dqs sqs med cond sup lay spc; do printf 'body{}\n' >"$site4/$f.css"; done
+cat >"$site4/main.css" <<'EOF'
+@import url(nq.css);
+@import url("dqu.css");
+@import url('squ.css');
+@import "dqs.css";
+@import 'sqs.css';
+@import url(med.css) screen and (min-width: 400px);
+@import "cond.css" screen;
+@import "sup.css" supports(display: flex);
+@import url(lay.css) layer(base);
+@import "spc.css" ;
+EOF
+out4="$tmp/cssimport-out"
+crawl "$site4/main.css" "$out4"
+for f in nq dqu squ dqs sqs med cond sup lay spc; do found "$f.css" "$out4"; done
+
+# Over-capture guard: the trailing condition is not part of the URL, so it must
+# survive the rewrite verbatim. A regression that grabs it would mangle these.
+m4=$(find "$out4" -type f -path '*/file/*' -name main.css -print -quit)
+test -n "$m4" || ! echo "FAIL: saved main.css not found" || exit 1
+for cond in '@import "cond.css" screen;' 'supports(display: flex)' 'layer(base)'; do
+    grep -Fq "$cond" "$m4" ||
+        ! echo "FAIL #94: '$cond' altered on rewrite (condition captured as URL?)" || exit 1
+done
+
+# Malformed input: an unterminated @import quote (truncated CSS) must not crash or
+# capture a bogus link; a valid sibling import is still captured. Guards a heap
+# overflow on the URL-end scan that aborts under ASan (CI sanitizer job).
+site5="$tmp/cssimport-trunc"
+mkdir -p "$site5"
+printf 'body{}\n' >"$site5/good.css"
+printf '@import "good.css";\n@import "trunc' >"$site5/main.css"
+out5="$tmp/cssimport-trunc-out"
+crawl "$site5/main.css" "$out5"
+found "good.css" "$out5"
+notfound "trunc" "$out5"
+
+# Offset-0 underflow (#396): a token at the buffer start makes the detector's
+# word-boundary guard read *(html-1) one byte early (aborts under ASan). The
+# url() target is still captured; here it just must not underflow.
+site6="$tmp/parse-off0"
+mkdir -p "$site6"
+printf 'body{}\n' >"$site6/off0.css"
+printf 'url(off0.css)\n' >"$site6/main.css"
+out6="$tmp/parse-off0-out"
+crawl "$site6/main.css" "$out6"
+found "off0.css" "$out6"
+
+# XMLHttpRequest.open(method, url) (#218): the first argument is an HTTP method,
+# not a URL. Without the fix "GET" is captured as a link and fetched (the offline
+# fixture saves a bare file named GET; a live server mangles it to GET.html).
+# window.open(url) detection must be unaffected.
+site7="$tmp/xhropen"
+mkdir -p "$site7"
+gif "$site7/winopen.gif"
+cat >"$site7/index.html" <<EOF
+<html><body><script>
+var x = new XMLHttpRequest();
+x.open("GET", "ajax_info.txt");
+var y = new XMLHttpRequest();
+y.open("Post", "submit.cgi");
+window.open("file://$site7/winopen.gif");
+</script></body></html>
+EOF
+out7="$tmp/xhropen-out"
+crawl "$site7/index.html" "$out7"
+# negative control: without the fix a file named exactly GET is downloaded
+notfound "GET" "$out7"
+# methods are matched case-insensitively (XHR spec normalizes them): a mixed-case
+# method is rejected too, so a file named Post must not appear either
+notfound "Post" "$out7"
+# regression guard: window.open(url) is still detected, so its absolute URL is
+# rewritten to a local link. The rewrite only happens if the parser saw it, so
+# these two assertions fail if .open detection broke (not a trivial --near save).
+saved7=$(savedhtml "$out7")
+test -n "$saved7" || ! echo "FAIL: saved xhr page not found" || exit 1
+grep -Fq 'window.open("winopen.gif")' "$saved7" ||
+    ! echo "FAIL #218: window.open(url) no longer detected/rewritten" || exit 1
+! grep -Fq 'window.open("file://' "$saved7" ||
+    ! echo "FAIL #218: window.open URL left absolute (not rewritten)" || exit 1
+
+# Parens in an unquoted url(...) (#163): the source %28/%29 decode to literal
+# '(' ')' in the saved name, but a literal ')' in the rewritten url() closes the
+# token early, so they must stay encoded. Negative control: without the fix the
+# %281%29 greps fail (parens are RFC2396 "mark" chars the escaper leaves alone).
+site8="$tmp/cssparens"
+mkdir -p "$site8"
+for f in 'img (1).gif' 'a(b)c(1).gif' 'q (4).gif'; do gif "$site8/$f"; done
+cat >"$site8/style.css" <<'EOF'
+.a { background: url(img%20%281%29.gif); }
+.b { background: url(a%28b%29c%281%29.gif); }
+.c { background: url("q%20%284%29.gif"); }
+EOF
+out8="$tmp/cssparens-out"
+crawl "$site8/style.css" "$out8"
+found "img (1).gif" "$out8"
+found "a(b)c(1).gif" "$out8"
+found "q (4).gif" "$out8"
+css8=$(find "$out8" -type f -path '*/file/*' -name style.css -print -quit)
+test -n "$css8" || ! echo "FAIL: saved style.css not found" || exit 1
+grep -Fq 'url(img%20%281%29.gif)' "$css8" ||
+    ! echo "FAIL #163: parens in unquoted url() not percent-encoded on rewrite" || exit 1
+grep -Fq 'url(a%28b%29c%281%29.gif)' "$css8" ||
+    ! echo "FAIL #163: not every paren in a url() was percent-encoded" || exit 1
+grep -Fq 'url("q%20%284%29.gif")' "$css8" ||
+    ! echo "FAIL #163: quoted url() altered or parens left literal on rewrite" || exit 1
+
+# The url() detector is not CSS-specific: <script> and inline style= get the
+# same encoding, but ordinary href/src (ending_p is the quote, not ')') keep
+# literal parens -- the attribute checks guard the gate against over-firing.
+site9="$tmp/urlparens"
+mkdir -p "$site9"
+for f in 'js (1).gif' 'inl (2).gif' 'asrc (3).gif' 'ahref (4).gif'; do gif "$site9/$f"; done
+cat >"$site9/index.html" <<EOF
+<html><body>
+<script>var bg = "url(js%20%281%29.gif)";</script>
+<div style="background-image:url(inl%20%282%29.gif)"></div>
+<img src="asrc%20%283%29.gif">
+<a href="ahref%20%284%29.gif">link</a>
+</body></html>
+EOF
+out9="$tmp/urlparens-out"
+crawl "$site9/index.html" "$out9"
+saved9=$(savedhtml "$out9")
+test -n "$saved9" || ! echo "FAIL: saved urlparens page not found" || exit 1
+# rewrite-only: the JS-string asset is not queued for download
+grep -Fq 'url(js%20%281%29.gif)' "$saved9" ||
+    ! echo "FAIL #163: parens in <script> url() not percent-encoded" || exit 1
+found "inl (2).gif" "$out9"
+grep -Fq 'url(inl%20%282%29.gif)' "$saved9" ||
+    ! echo "FAIL #163: parens in inline style url() not percent-encoded" || exit 1
+found "asrc (3).gif" "$out9"
+found "ahref (4).gif" "$out9"
+grep -Fq 'src="asrc%20(3).gif"' "$saved9" ||
+    ! echo "FAIL #163: parens in a plain src attribute were wrongly encoded" || exit 1
+grep -Fq 'href="ahref%20(4).gif"' "$saved9" ||
+    ! echo "FAIL #163: parens in a plain href attribute were wrongly encoded" || exit 1
+! grep -Eq '(src|href)="[^"]*%28' "$saved9" ||
+    ! echo "FAIL #163: gate over-fired onto a non-url() attribute link" || exit 1
+
 exit 0

tests/01_engine-relative.test

@@ -0,0 +1,68 @@
+#!/bin/bash
+#
+# lienrelatif (build relative path) + ident_url_relatif (resolve a link, collapse
+# ./ and ../). Regression net for #137/#162; expected values hand-computed.
+
+set -euo pipefail
+
+# relative path from <curr>'s directory to <link>
+rel() {
+    local got
+    got=$(httrack -O /dev/null -#l "$1" "$2")
+    test "$got" == "relative=$3" ||
+        {
+            echo "FAIL rel($1, $2): got '$got' want 'relative=$3'"
+            exit 1
+        }
+}
+
+# resolve <link> against origin <adr>/<fil> -> adr=.. fil=..
+ident() {
+    local got
+    got=$(httrack -O /dev/null -#i "$1" "$2" "$3")
+    test "$got" == "$4" ||
+        {
+            echo "FAIL ident($1, $2, $3): got '$got' want '$4'"
+            exit 1
+        }
+}
+
+### lienrelatif
+
+rel 'dir/page.html' 'dir/index.html' 'page.html'
+rel 'dir/page.html' 'dir/page.html' 'page.html' # self-link
+rel 'a.html' 'dir/index.html' '../a.html'
+rel 'x.html' 'a/b/c/index.html' '../../../x.html'
+rel 'h/a/x.jpg' 'h/a/sub/page.html' '../x.jpg'
+rel 'a/b/c/x.html' 'index.html' 'a/b/c/x.html'
+rel 'h/sub/x.jpg' 'h/page.html' 'sub/x.jpg'
+rel 'h/dir2/x.jpg' 'h/dir1/page.html' '../dir2/x.jpg' # sibling dir
+rel 'h/bc/x.jpg' 'h/b/page.html' '../bc/x.jpg'        # b/bc prefix trap
+rel 'h/b/x.jpg' 'h/bc/page.html' '../b/x.jpg'
+rel 'h2/img/x.jpg' 'h1/p/page.html' '../../h2/img/x.jpg' # cross-host
+rel 'img.cdn/photo.jpg' 'www.site/articles/2020/post.html' '../../../img.cdn/photo.jpg'
+rel 'h/a/' 'h/a/sub/page.html' '../' # link is ancestor dir
+rel 'x.html' 'page.html' 'x.html'
+rel 'dir/page.html?x=1' 'dir/index.html?y=2' 'page.html' # ? stripped
+
+### ident_url_relatif
+
+ident 'img.gif' 'www.foo.com' '/dir/page.html' 'adr=www.foo.com fil=/dir/img.gif'
+ident 'sub/img.gif' 'www.foo.com' '/dir/page.html' 'adr=www.foo.com fil=/dir/sub/img.gif'
+ident '/img.gif' 'www.foo.com' '/dir/page.html' 'adr=www.foo.com fil=/img.gif'
+# embedded ../ collapses (#137)
+ident '../img.gif' 'www.foo.com' '/dir/sub/page.html' 'adr=www.foo.com fil=/dir/img.gif'
+ident 'sub/../logo.png' 'www.foo.com' '/articles/2020/post.html' 'adr=www.foo.com fil=/articles/2020/logo.png'
+ident '../../pix/sub/../logo.png' 'www.foo.com' '/articles/2020/post.html' 'adr=www.foo.com fil=/pix/logo.png'
+ident '../../../../x.gif' 'www.foo.com' '/a/b/page.html' 'adr=www.foo.com fil=/x.gif' # above-root clamp
+ident '?page=2' 'www.foo.com' '/dir/index.html?old=1' 'adr=www.foo.com fil=/dir/index.html?page=2'
+ident 'http://other.com/a/b/../c/index.html' 'www.foo.com' '/p.html' 'adr=other.com fil=/a/c/index.html'
+# file:// collapses ../ like the other schemes; traversal contained, // authority kept
+ident 'file:///var/data/pix/sub/../logo.png' 'www.foo.com' '/p.html' 'adr=file:// fil=/var/data/pix/logo.png'
+ident 'file:///a/b/c/../../d/e.gif' 'www.foo.com' '/p.html' 'adr=file:// fil=/a/d/e.gif'
+ident 'file:///a/../../b' 'www.foo.com' '/p.html' 'adr=file:// fil=/b'
+ident 'file://srv/share/../x' 'www.foo.com' '/p.html' 'adr=file:// fil=//srv/x'
+ident 'mailto:foo@bar.com' 'www.foo.com' '/p.html' 'error=-1' # unsupported scheme
+ident 'javascript:void(0)' 'www.foo.com' '/p.html' 'error=-1'
+
+echo "OK"

tests/01_engine-simplify.test

@@ -5,7 +5,7 @@ set -euo pipefail
 
 # path simplify engine (fil_simplifie): collapses ./ and ../ segments.
 simp() {
-	test "$(httrack -O /dev/null -#1 "$1")" == "simplified=$2" || exit 1
+    test "$(httrack -O /dev/null -#1 "$1")" == "simplified=$2" || exit 1
 }
 
 simp './foo/bar/' 'foo/bar/'
@@ -26,3 +26,17 @@ simp './a/../../b' 'b'
 
 # empty segments ('//') are not dot-segments and are preserved, per RFC 3986
 simp 'a//b' 'a//b'
+simp 'a//b/../c' 'a//c'
+
+# absolute paths keep the leading '/'; above-root '..' is clamped to it
+simp '/a/../b' '/b'
+simp '/a/../../b' '/b'
+simp '/../x' '/x'
+
+# collapses to nothing -> './' (relative) or '/' (absolute)
+simp '..' './'
+simp 'a/..' './'
+simp '/' '/'
+
+simp 'a/b/..' 'a/'              # trailing bare '..'
+simp 'a/../b?x=../y' 'b?x=../y' # '?' freezes simplification

tests/01_engine-strsafe.test

@@ -21,9 +21,15 @@ test "$out" == "strsafe: OK" || exit 1
 # the bounded macro aborts (non-zero exit), so don't let set -e trip on it
 err=$(httrack -#8 overflow "this string is far too long for the buffer" 2>&1) || true
 case "$err" in
-	*"strsafe: NOT aborted"*) echo "over-capacity write was NOT caught" >&2; exit 1 ;;
+*"strsafe: NOT aborted"*)
-	*"overflow while copying"*) ;;
+    echo "over-capacity write was NOT caught" >&2
-	*) echo "expected htssafe overflow abort, got: $err" >&2; exit 1 ;;
+    exit 1
+    ;;
+*"overflow while copying"*) ;;
+*)
+    echo "expected htssafe overflow abort, got: $err" >&2
+    exit 1
+    ;;
 esac
 
 # Same guarantee for the htsbuff builder. The source is exactly the buffer
@@ -32,7 +38,13 @@ esac
 # aborted"). Match the specific htsbuff abort message, not just any assert.
 err=$(httrack -#8 overflow-buff "abcd" 2>&1) || true
 case "$err" in
-	*"strsafe: NOT aborted"*) echo "htsbuff over-capacity write was NOT caught" >&2; exit 1 ;;
+*"strsafe: NOT aborted"*)
-	*"htsbuff append overflow"*) ;;
+    echo "htsbuff over-capacity write was NOT caught" >&2
-	*) echo "expected htsbuff overflow abort, got: $err" >&2; exit 1 ;;
+    exit 1
+    ;;
+*"htsbuff append overflow"*) ;;
+*)
+    echo "expected htsbuff overflow abort, got: $err" >&2
+    exit 1
+    ;;
 esac

tests/10_crawl-simple.test

@@ -3,6 +3,6 @@
 
 set -euo pipefail
 
-bash check-network.sh ||  ! echo "skipping online unit tests" || exit 77
+bash check-network.sh || ! echo "skipping online unit tests" || exit 77
 
 bash crawl-test.sh --errors 0 --files 5 httrack http://ut.httrack.com/simple/basic.html

tests/11_crawl-cookies.test

@@ -3,10 +3,10 @@
 
 set -euo pipefail
 
-bash check-network.sh ||  ! echo "skipping online unit tests" || exit 77
+bash check-network.sh || ! echo "skipping online unit tests" || exit 77
 
 bash crawl-test.sh --errors 0 --files 3 \
-	--found ut.httrack.com/cookies/third.html \
+    --found ut.httrack.com/cookies/third.html \
-	--found ut.httrack.com/cookies/second.html \
+    --found ut.httrack.com/cookies/second.html \
-	--found ut.httrack.com/cookies/entrance.html \
+    --found ut.httrack.com/cookies/entrance.html \
-	httrack http://ut.httrack.com/cookies/entrance.php
+    httrack http://ut.httrack.com/cookies/entrance.php

tests/11_crawl-idna.test

@@ -3,21 +3,21 @@
 
 set -euo pipefail
 
-bash check-network.sh ||  ! echo "skipping online unit tests" || exit 77
+bash check-network.sh || ! echo "skipping online unit tests" || exit 77
 
 # unicode tests
 bash crawl-test.sh \
-	--errors 1 --files 5 \
+    --errors 1 --files 5 \
-	--found 'café.ut.httrack.com/unicode-links/café3860.html' \
+    --found 'café.ut.httrack.com/unicode-links/café3860.html' \
-	--found 'café.ut.httrack.com/unicode-links/café30f4.html' \
+    --found 'café.ut.httrack.com/unicode-links/café30f4.html' \
-	--found 'café.ut.httrack.com/unicode-links/café5e1f.html' \
+    --found 'café.ut.httrack.com/unicode-links/café5e1f.html' \
-	--found 'café.ut.httrack.com/unicode-links/café7b30.html' \
+    --found 'café.ut.httrack.com/unicode-links/café7b30.html' \
-	httrack 'http://ut.httrack.com/unicode-links/idna.html' \
+    httrack 'http://ut.httrack.com/unicode-links/idna.html' \
-	'+*.ut.httrack.com/*' --robots=0
+    '+*.ut.httrack.com/*' --robots=0
 
 # unicode tests (bogus links)
 bash crawl-test.sh \
-	--errors 0 --files 1 \
+    --errors 0 --files 1 \
-	--found 'ut.httrack.com/unicode-links/idna_bogus.html' \
+    --found 'ut.httrack.com/unicode-links/idna_bogus.html' \
-	httrack 'http://ut.httrack.com/unicode-links/idna_bogus.html' \
+    httrack 'http://ut.httrack.com/unicode-links/idna_bogus.html' \
-	'-*' --robots=0
+    '-*' --robots=0

tests/11_crawl-international.test

@@ -3,67 +3,67 @@
 
 set -euo pipefail
 
-bash check-network.sh ||  ! echo "skipping online unit tests" || exit 77
+bash check-network.sh || ! echo "skipping online unit tests" || exit 77
 
 # unicode tests
 bash crawl-test.sh \
-	--errors 1 --files 10 \
+    --errors 1 --files 10 \
-	--found ut.httrack.com/unicode-links/caf%a91bce.html \
+    --found ut.httrack.com/unicode-links/caf%a91bce.html \
-	--found ut.httrack.com/unicode-links/café30f4.html \
+    --found ut.httrack.com/unicode-links/café30f4.html \
-	--found ut.httrack.com/unicode-links/café3860.html \
+    --found ut.httrack.com/unicode-links/café3860.html \
-	--found ut.httrack.com/unicode-links/café463e.html \
+    --found ut.httrack.com/unicode-links/café463e.html \
-	--found ut.httrack.com/unicode-links/café5e1f.html \
+    --found ut.httrack.com/unicode-links/café5e1f.html \
-	--found ut.httrack.com/unicode-links/café7b30.html \
+    --found ut.httrack.com/unicode-links/café7b30.html \
-	--found ut.httrack.com/unicode-links/café8007.html \
+    --found ut.httrack.com/unicode-links/café8007.html \
-	--found ut.httrack.com/unicode-links/café9fa8.html \
+    --found ut.httrack.com/unicode-links/café9fa8.html \
-	--found ut.httrack.com/unicode-links/caféae52.html \
+    --found ut.httrack.com/unicode-links/caféae52.html \
-	--found ut.httrack.com/unicode-links/caféc009.html \
+    --found ut.httrack.com/unicode-links/caféc009.html \
-	--found ut.httrack.com/unicode-links/utf8.html \
+    --found ut.httrack.com/unicode-links/utf8.html \
-	httrack http://ut.httrack.com/unicode-links/utf8.html
+    httrack http://ut.httrack.com/unicode-links/utf8.html
 
 bash crawl-test.sh \
-	--errors 4 --files  7 \
+    --errors 4 --files 7 \
-	--found ut.httrack.com/unicode-links/café3860.html \
+    --found ut.httrack.com/unicode-links/café3860.html \
-	--found ut.httrack.com/unicode-links/café9fa8.html \
+    --found ut.httrack.com/unicode-links/café9fa8.html \
-	--found ut.httrack.com/unicode-links/café30f4.html \
+    --found ut.httrack.com/unicode-links/café30f4.html \
-	--found ut.httrack.com/unicode-links/café5e1f.html \
+    --found ut.httrack.com/unicode-links/café5e1f.html \
-	--found ut.httrack.com/unicode-links/café7b30.html \
+    --found ut.httrack.com/unicode-links/café7b30.html \
-	--found ut.httrack.com/unicode-links/café8007.html \
+    --found ut.httrack.com/unicode-links/café8007.html \
-	--found ut.httrack.com/unicode-links/caf%e939bd.html \
+    --found ut.httrack.com/unicode-links/caf%e939bd.html \
-	--found ut.httrack.com/unicode-links/caf%e9ae52.html \
+    --found ut.httrack.com/unicode-links/caf%e9ae52.html \
-	--found ut.httrack.com/unicode-links/caféaec2.html \
+    --found ut.httrack.com/unicode-links/caféaec2.html \
-	--found ut.httrack.com/unicode-links/caféfad6.html \
+    --found ut.httrack.com/unicode-links/caféfad6.html \
-	--found ut.httrack.com/unicode-links/default.html \
+    --found ut.httrack.com/unicode-links/default.html \
-	httrack http://ut.httrack.com/unicode-links/default.html
+    httrack http://ut.httrack.com/unicode-links/default.html
 
 bash crawl-test.sh \
-	--errors 2 --files  9 \
+    --errors 2 --files 9 \
-	--found ut.httrack.com/unicode-links/caf%a9ae52.html \
+    --found ut.httrack.com/unicode-links/caf%a9ae52.html \
-	--found ut.httrack.com/unicode-links/caf%a9bf59.html \
+    --found ut.httrack.com/unicode-links/caf%a9bf59.html \
-	--found ut.httrack.com/unicode-links/café30f4.html \
+    --found ut.httrack.com/unicode-links/café30f4.html \
-	--found ut.httrack.com/unicode-links/café3860.html \
+    --found ut.httrack.com/unicode-links/café3860.html \
-	--found ut.httrack.com/unicode-links/café5e1f.html \
+    --found ut.httrack.com/unicode-links/café5e1f.html \
-	--found ut.httrack.com/unicode-links/café647f.html \
+    --found ut.httrack.com/unicode-links/café647f.html \
-	--found ut.httrack.com/unicode-links/café7b30.html \
+    --found ut.httrack.com/unicode-links/café7b30.html \
-	--found ut.httrack.com/unicode-links/café8007.html \
+    --found ut.httrack.com/unicode-links/café8007.html \
-	--found ut.httrack.com/unicode-links/caféaec2.html \
+    --found ut.httrack.com/unicode-links/caféaec2.html \
-	--found ut.httrack.com/unicode-links/caféfad6.html \
+    --found ut.httrack.com/unicode-links/caféfad6.html \
-	--found ut.httrack.com/unicode-links/iso88591.html \
+    --found ut.httrack.com/unicode-links/iso88591.html \
-	httrack http://ut.httrack.com/unicode-links/iso88591.html
+    httrack http://ut.httrack.com/unicode-links/iso88591.html
 
 bash crawl-test.sh \
-	--errors 4 --files  9 \
+    --errors 4 --files 9 \
-	--found ut.httrack.com/unicode-links/caf%a8%a6c72a.html \
+    --found ut.httrack.com/unicode-links/caf%a8%a6c72a.html \
-	--found ut.httrack.com/unicode-links/caf%a9bf59.html \
+    --found ut.httrack.com/unicode-links/caf%a9bf59.html \
-	--found ut.httrack.com/unicode-links/café8007.html \
+    --found ut.httrack.com/unicode-links/café8007.html \
-	--found ut.httrack.com/unicode-links/cafébf43.html \
+    --found ut.httrack.com/unicode-links/cafébf43.html \
-	--found ut.httrack.com/unicode-links/cafédcd8.html \
+    --found ut.httrack.com/unicode-links/cafédcd8.html \
-	--found ut.httrack.com/unicode-links/café2461.html \
+    --found ut.httrack.com/unicode-links/café2461.html \
-	--found ut.httrack.com/unicode-links/caf%a8%a61bce.html \
+    --found ut.httrack.com/unicode-links/caf%a8%a61bce.html \
-	--found ut.httrack.com/unicode-links/caf%a9ae52.html \
+    --found ut.httrack.com/unicode-links/caf%a9ae52.html \
-	--found ut.httrack.com/unicode-links/café7b30.html \
+    --found ut.httrack.com/unicode-links/café7b30.html \
-	--found ut.httrack.com/unicode-links/café30f4.html \
+    --found ut.httrack.com/unicode-links/café30f4.html \
-	--found ut.httrack.com/unicode-links/café5e1f.html \
+    --found ut.httrack.com/unicode-links/café5e1f.html \
-	--found ut.httrack.com/unicode-links/café3860.html \
+    --found ut.httrack.com/unicode-links/café3860.html \
-	--found ut.httrack.com/unicode-links/gb18030.html \
+    --found ut.httrack.com/unicode-links/gb18030.html \
-	httrack http://ut.httrack.com/unicode-links/gb18030.html
+    httrack http://ut.httrack.com/unicode-links/gb18030.html

tests/11_crawl-longurl.test

@@ -3,10 +3,10 @@
 
 set -euo pipefail
 
-bash check-network.sh ||  ! echo "skipping online unit tests" || exit 77
+bash check-network.sh || ! echo "skipping online unit tests" || exit 77
 
 # http://code.google.com/p/httrack/issues/detail?id=42&can=1
 # we expect 2 errors only because other links are too longs (to be modified if suitable)
 bash crawl-test.sh --errors 2 --files 1 \
-	--found ut.httrack.com/overflow/longquerywithaccents.html \
+    --found ut.httrack.com/overflow/longquerywithaccents.html \
-	httrack http://ut.httrack.com/overflow/longquerywithaccents.php
+    httrack http://ut.httrack.com/overflow/longquerywithaccents.php

tests/11_crawl-parsing.test

@@ -3,45 +3,45 @@
 
 set -euo pipefail
 
-bash check-network.sh ||  ! echo "skipping online unit tests" || exit 77
+bash check-network.sh || ! echo "skipping online unit tests" || exit 77
 
 # http://code.google.com/p/httrack/issues/detail?id=4&can=1
 bash crawl-test.sh --errors 0 --files 4 \
-	--found ut.httrack.com/parsing/back5e1f.gif \
+    --found ut.httrack.com/parsing/back5e1f.gif \
-	--found ut.httrack.com/parsing/events.html \
+    --found ut.httrack.com/parsing/events.html \
-	--found ut.httrack.com/parsing/fade230f4.gif \
+    --found ut.httrack.com/parsing/fade230f4.gif \
-	--found ut.httrack.com/parsing/fade3860.gif \
+    --found ut.httrack.com/parsing/fade3860.gif \
-	httrack http://ut.httrack.com/parsing/events.html
+    httrack http://ut.httrack.com/parsing/events.html
 
 # http://code.google.com/p/httrack/issues/detail?id=2&can=1
 bash crawl-test.sh --errors 0 --files 3 \
-	--found ut.httrack.com/parsing/background-image.css \
+    --found ut.httrack.com/parsing/background-image.css \
-	--found ut.httrack.com/parsing/background-image.html \
+    --found ut.httrack.com/parsing/background-image.html \
-	--found ut.httrack.com/parsing/fade.gif \
+    --found ut.httrack.com/parsing/fade.gif \
-	httrack http://ut.httrack.com/parsing/background-image.html
+    httrack http://ut.httrack.com/parsing/background-image.html
 
 # javascript parsing
 bash crawl-test.sh --errors 0 --files 3 \
-	--found ut.httrack.com/parsing/back.gif \
+    --found ut.httrack.com/parsing/back.gif \
-	--found ut.httrack.com/parsing/fade.gif \
+    --found ut.httrack.com/parsing/fade.gif \
-	--found ut.httrack.com/parsing/javascript.html \
+    --found ut.httrack.com/parsing/javascript.html \
-	httrack http://ut.httrack.com/parsing/javascript.html
+    httrack http://ut.httrack.com/parsing/javascript.html
 
 # handling of + before query string
 bash crawl-test.sh --errors 0 --files 6 \
-	--found ut.httrack.com/parsing/escaping.html \
+    --found ut.httrack.com/parsing/escaping.html \
-	--found "ut.httrack.com/parsing/foo bar30f4.html" \
+    --found "ut.httrack.com/parsing/foo bar30f4.html" \
-	--found "ut.httrack.com/parsing/foo bar5e1f.html" \
+    --found "ut.httrack.com/parsing/foo bar5e1f.html" \
-	--found "ut.httrack.com/parsing/foo+bar+plus3860.html" \
+    --found "ut.httrack.com/parsing/foo+bar+plus3860.html" \
-	--found "ut.httrack.com/parsing/foo barae52.html" \
+    --found "ut.httrack.com/parsing/foo barae52.html" \
-	--found "ut.httrack.com/parsing/foo bar7b30.html" \
+    --found "ut.httrack.com/parsing/foo bar7b30.html" \
-	httrack http://ut.httrack.com/parsing/escaping.html
+    httrack http://ut.httrack.com/parsing/escaping.html
 
 # handling of # encoded in filename
 # see http://code.google.com/p/httrack/issues/detail?id=25
 bash crawl-test.sh --errors 2 --files 4 \
-	--found "ut.httrack.com/parsing/escaping2.html" \
+    --found "ut.httrack.com/parsing/escaping2.html" \
-	--found "ut.httrack.com/parsing/++foo++bar++plus++.html" \
+    --found "ut.httrack.com/parsing/++foo++bar++plus++.html" \
-	--found "ut.httrack.com/parsing/foo#bar#.html" \
+    --found "ut.httrack.com/parsing/foo#bar#.html" \
-	--found "ut.httrack.com/parsing/foo bar.html" \
+    --found "ut.httrack.com/parsing/foo bar.html" \
-	httrack http://ut.httrack.com/parsing/escaping2.html
+    httrack http://ut.httrack.com/parsing/escaping2.html

tests/12_crawl_https.test

@@ -3,11 +3,11 @@
 
 set -euo pipefail
 
-bash check-network.sh ||  ! echo "skipping online unit tests" || exit 77
+bash check-network.sh || ! echo "skipping online unit tests" || exit 77
 
 if test "${HTTPS_SUPPORT:-}" == "no"; then
-	echo "no https support compiled, skipping"
+    echo "no https support compiled, skipping"
-	exit 77
+    exit 77
 fi
 
 bash crawl-test.sh --errors 0 --files 5 httrack https://ut.httrack.com/simple/basic.html

tests/13_crawl_proxy_https.test

@@ -0,0 +1,136 @@
+#!/bin/bash
+#
+# Issue #85: an https crawl must go through the configured proxy (CONNECT
+# tunnel), not bypass it and hit the origin directly. Fully local: a self-signed
+# TLS origin plus a logging CONNECT proxy, so no network access is needed.
+
+set -euo pipefail
+
+: "${top_srcdir:=..}"
+
+if test "${HTTPS_SUPPORT:-}" == "no"; then
+    echo "no https support compiled, skipping"
+    exit 77
+fi
+if ! command -v python3 >/dev/null 2>&1 || ! command -v openssl >/dev/null 2>&1; then
+    echo "python3/openssl missing, skipping"
+    exit 77
+fi
+
+server="$top_srcdir/tests/proxy-https-server.py"
+tmpdir=$(mktemp -d)
+pids=
+
+cleanup() {
+    for pid in $pids; do
+        kill "$pid" 2>/dev/null || true
+    done
+    rm -rf "$tmpdir"
+}
+trap cleanup EXIT
+
+# self-signed cert for the local TLS origin (httrack does not verify certs)
+openssl req -x509 -newkey rsa:2048 -keyout "$tmpdir/key.pem" \
+    -out "$tmpdir/cert.pem" -days 2 -nodes -subj "/CN=127.0.0.1" \
+    >/dev/null 2>&1
+cat "$tmpdir/key.pem" "$tmpdir/cert.pem" >"$tmpdir/both.pem"
+
+# start_server <logdir> <mode>: launches a proxy+origin pair, sets $origin_port
+# and $proxy_port from its announced ephemeral ports.
+start_server() {
+    local dir="$1" mode="$2" ports
+    mkdir -p "$dir"
+    ports="$dir/ports.txt"
+    python3 "$server" "$tmpdir/both.pem" "$dir" "$mode" \
+        >"$ports" 2>"$dir/server.err" &
+    pids="$pids $!"
+    for _ in $(seq 1 100); do
+        grep -q "^ready" "$ports" 2>/dev/null && break
+        sleep 0.1
+    done
+    grep -q "^ready" "$ports" 2>/dev/null || {
+        echo "server ($mode) did not start" >&2
+        cat "$dir/server.err" >&2
+        exit 1
+    }
+    origin_port=$(awk '/^ORIGIN/{print $2}' "$ports")
+    proxy_port=$(awk '/^PROXY/{print $2}' "$ports")
+}
+
+# Run httrack, but kill it after a deadline so a hang (e.g. a missing bound on
+# the proxy response) surfaces as the kill code $HANG_RC instead of stalling the
+# whole job. A portable stand-in for `timeout`, which macOS lacks.
+HANG_RC=137 # 128 + SIGKILL
+run_crawl() {
+    local out="$1" proxy="$2" port="$3"
+    rm -rf "$out"
+    httrack "https://127.0.0.1:${port}/" --proxy "$proxy" \
+        -O "$out" -r1 -s0 --timeout=10 >"$out.log" 2>&1 &
+    local pid=$!
+    (sleep 60 && kill -9 "$pid" 2>/dev/null) &
+    local guard=$!
+    local rc=0
+    wait "$pid" 2>/dev/null || rc=$?
+    kill "$guard" 2>/dev/null || true
+    wait "$guard" 2>/dev/null || true
+    return "$rc"
+}
+
+# --- working proxy ----------------------------------------------------------
+ok="$tmpdir/ok"
+start_server "$ok" ok
+
+# 1. page retrieved AND the proxy saw a CONNECT to the origin
+run_crawl "$ok/out" "127.0.0.1:${proxy_port}" "$origin_port"
+grep -rq "ORIGIN-PAGE-85" "$ok/out" || {
+    echo "FAIL: origin page not downloaded through proxy" >&2
+    cat "$ok/out.log" >&2
+    exit 1
+}
+grep -q "^CONNECT 127.0.0.1:${origin_port} " "$ok/proxy.log" || {
+    echo "FAIL: proxy never received a CONNECT (https bypassed the proxy)" >&2
+    cat "$ok/proxy.log" >&2
+    exit 1
+}
+echo "OK: https tunneled through proxy via CONNECT"
+
+# 2. authenticated proxy: creds ride the CONNECT, and NEVER reach the origin
+: >"$ok/proxy.log"
+: >"$ok/origin-headers.log"
+run_crawl "$ok/out2" "user:secret@127.0.0.1:${proxy_port}" "$origin_port"
+grep -rq "ORIGIN-PAGE-85" "$ok/out2" || {
+    echo "FAIL: origin page not downloaded through authenticated proxy" >&2
+    exit 1
+}
+got=$(awk '/^AUTH Basic /{print $3}' "$ok/proxy.log" | head -1)
+# base64("user:secret"); compared as a literal to stay portable (no base64 -d,
+# which differs between GNU and BSD)
+test "$got" == "dXNlcjpzZWNyZXQ=" || {
+    echo "FAIL: Proxy-Authorization not carried on CONNECT (got '$got')" >&2
+    cat "$ok/proxy.log" >&2
+    exit 1
+}
+if grep -qi "proxy-authorization" "$ok/origin-headers.log"; then
+    echo "FAIL: proxy credentials leaked to the origin through the tunnel" >&2
+    cat "$ok/origin-headers.log" >&2
+    exit 1
+fi
+echo "OK: proxy credentials carried on CONNECT, not leaked to origin"
+
+# --- hostile proxy ----------------------------------------------------------
+# A proxy that answers 200 then streams headers forever must not hang the crawl:
+# the client bounds the response. run_crawl kills a hung httrack after 60s, so a
+# missing bound surfaces as $HANG_RC here.
+flood="$tmpdir/flood"
+start_server "$flood" flood
+rc=0
+run_crawl "$flood/out" "127.0.0.1:${proxy_port}" "$origin_port" || rc=$?
+test "$rc" -ne "$HANG_RC" || {
+    echo "FAIL: crawl hung on a flooding proxy (bounded read missing)" >&2
+    exit 1
+}
+grep -rq "ORIGIN-PAGE-85" "$flood/out" 2>/dev/null && {
+    echo "FAIL: flooding proxy unexpectedly served the page" >&2
+    exit 1
+}
+echo "OK: bounded proxy response, no hang on a flooding proxy"

tests/Makefile.am

@@ -2,6 +2,7 @@
 # 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 \
+	proxy-https-server.py \
 	fixtures/cache-golden/hts-cache/new.zip
 
 TESTS_ENVIRONMENT =
@@ -24,6 +25,8 @@ TESTS = \
 	01_engine-cache-golden.test \
 	01_engine-charset.test \
 	01_engine-cmdline.test \
+	01_engine-cookies.test \
+	01_engine-copyopt.test \
 	01_engine-doitlog.test \
 	01_engine-entities.test \
 	01_engine-filter.test \
@@ -32,6 +35,7 @@ TESTS = \
 	01_engine-mime.test \
 	01_engine-parse.test \
 	01_engine-rcfile.test \
+	01_engine-relative.test \
 	01_engine-simplify.test \
 	01_engine-strsafe.test \
 	02_manpage-regen.test \
@@ -42,6 +46,7 @@ TESTS = \
 	11_crawl-international.test \
 	11_crawl-longurl.test \
 	11_crawl-parsing.test \
-	12_crawl_https.test
+	12_crawl_https.test \
+	13_crawl_proxy_https.test
 
 CLEANFILES = check-network_sh.cache

tests/check-network.sh

@@ -6,39 +6,39 @@
 
 # do not enable online tests (./configure --disable-online-unit-tests)
 if test "$ONLINE_UNIT_TESTS" == "no"; then
-echo "online tests are disabled" >&2
+    echo "online tests are disabled" >&2
-exit 1
+    exit 1
 
 # enable online tests (--enable-online-unit-tests)
 elif test "$ONLINE_UNIT_TESTS" == "yes"; then
-exit 0
+    exit 0
 
 # check if online tests are reachable
 else
 
-# test url
+    # test url
-url=http://ut.httrack.com/enabled
+    url=http://ut.httrack.com/enabled
 
-# cache file name
+    # cache file name
-cache=check-network_sh.cache
+    cache=check-network_sh.cache
 
-# cached result ?
+    # cached result ?
-if test -f $cache ; then
+    if test -f $cache; then
-	if grep -q "ok" $cache ; then
+        if grep -q "ok" $cache; then
-		exit 0
+            exit 0
-	else
+        else
-		echo "online tests are disabled (cached)" >&2
+            echo "online tests are disabled (cached)" >&2
-		exit 1
+            exit 1
-	fi
+        fi
 
-# fetch single file
+    # fetch single file
-elif bash crawl-test.sh --errors 0 --files 1 httrack --timeout=3 --max-time=3 "$url" 2>/dev/null >/dev/null ; then
+    elif bash crawl-test.sh --errors 0 --files 1 httrack --timeout=3 --max-time=3 "$url" 2>/dev/null >/dev/null; then
-	echo "ok" > $cache
+        echo "ok" >$cache
-	exit 0
+        exit 0
-else
+    else
-	echo "error" > $cache
+        echo "error" >$cache
-	echo "online tests are disabled (auto)" >&2
+        echo "online tests are disabled (auto)" >&2
-	exit 1
+        exit 1
-fi
+    fi
 
 fi

tests/crawl-test.sh

@@ -2,185 +2,184 @@
 #
 
 function warning {
-  echo "** $*" >&2
+    echo "** $*" >&2
-  return 0
+    return 0
 }
 
 function die {
-  warning "$*"
+    warning "$*"
-  exit 1
+    exit 1
 }
 
 function debug {
-  if test -n "$verbose"; then
+    if test -n "$verbose"; then
-    echo "$*" >&2
+        echo "$*" >&2
-  fi
+    fi
 }
 
 function info {
-  printf "[$*] ..\t" >&2
+    printf '[%s] ..\t' "$*" >&2
 }
 
 function result {
-  echo "$*" >&2
+    echo "$*" >&2
 }
 
 function cleanup {
-  debug "cleaning function called"
+    debug "cleaning function called"
-  if test -n "$tmpdir"; then
+    if test -n "$tmpdir"; then
-    if test -d "$tmpdir"; then
+        if test -d "$tmpdir"; then
-      if test -z "$nopurge"; then
+            if test -z "$nopurge"; then
-        debug "cleaning up $tmpdir"
+                debug "cleaning up $tmpdir"
-        rm -rf "$tmpdir"
+                rm -rf "$tmpdir"
-      fi
+            fi
+        fi
+    fi
+    if test -n "$crawlpid"; then
+        debug "killing $crawlpid"
+        kill -9 "$crawlpid"
+        crawlpid=
     fi
-  fi
-  if test -n "$crawlpid"; then
-    debug "killing $crawlpid"
-    kill -9 "$crawlpid"
-    crawlpid=
-  fi
 }
 
 function usage {
-  cat << EOF
+    cat <<EOF
 usage: $0
 EOF
 }
 
 function assert_equals {
-  info "$1"
+    info "$1"
-  if test ! "$2" == "$3"; then
+    if test ! "$2" == "$3"; then
-    result "expected '$2', got '$3'"
+        result "expected '$2', got '$3'"
-    exit 1
+        exit 1
-  else
+    else
-    result "OK ($2)"
+        result "OK ($2)"
-  fi
+    fi
 }
 
 function start-crawl {
-  # parse args
+    # parse args
-  pos=1
+    pos=1
-  while test "$#" -ge "$pos" ; do
+    while test "$#" -ge "$pos"; do
-    case "${!pos}" in
+        case "${!pos}" in
-    --debug)
+        --debug)
-      verbose=1
+            verbose=1
-      ;;
+            ;;
-    --no-purge|--summary|--print-files)
+        --no-purge | --summary | --print-files) ;;
-      ;;
+        --errors | --files | --found | --not-found | --directory)
-    --errors|--files|--found|--not-found|--directory)
+            pos=$((pos + 1))
-      pos=$[${pos}+1]
+            test "$#" -ge "$pos" || warning "missing argument" || return 1
-      test "$#" -ge "$pos" || warning "missing argument" || return 1
+            ;;
-      ;;
+        httrack)
-    httrack)
+            pos=$((pos + 1))
-      pos=$[${pos}+1]
+            break
-      break;
+            ;;
-      ;;
+        *)
-    *)
+            warning "unrecognized option ${!pos}"
-      warning "unrecognized option ${!pos}"
+            return 1
-      return 1
+            ;;
-      ;;
+        esac
-    esac
+        pos=$((pos + 1))
-    pos=$[${pos}+1]
+    done
-  done
+    debug "remaining args: ${*:pos}"
-  debug "remaining args: ${@:${pos}}"
 
-  # ut/ won't exceed 2 minutes
+    # ut/ won't exceed 2 minutes
-  moreargs="--quiet --max-time=120 --timeout=30 --connection-per-second=5"
+    moreargs=(--quiet --max-time=120 --timeout=30 --connection-per-second=5)
 
-  # proxy environment ?
+    # proxy environment ?
-  if test -n "$http_proxy"; then
+    if test -n "${http_proxy:-}"; then
-    moreargs="$moreargs --proxy $http_proxy"
+        moreargs+=(--proxy "$http_proxy")
-  fi
+    fi
 
-  test -n "$tmpdir" || ! warning "no tmpdir" || return 1
+    test -n "$tmpdir" || ! warning "no tmpdir" || return 1
-  tmp="${tmpdir}/crawl"
+    tmp="${tmpdir}/crawl"
-  rm -rf "$tmp"
-  mkdir "$tmp" || ! warning "could not create $tmp" || return 1
-
-  which httrack >/dev/null || ! warning "could not find httrack" || return 1
-  ver=$(httrack -O /dev/null --version | sed -e 's/HTTrack version //')
-  test -n "$ver" || ! warning "could not run httrack" || return 1
-
-  # start crawl
-  log="${tmp}/log"
-  debug starting httrack -O "${tmp}" ${moreargs} ${@:${pos}}
-  info "running httrack ${@:${pos}}"
-  httrack -O "${tmp}" --user-agent="httrack $ver ut ($(uname -omrs))" ${moreargs} ${@:${pos}} >"${log}" 2>&1 &
-  crawlpid="$!"
-  debug "started cralwer on pid $crawlpid"
-  wait "$crawlpid"
-  result="$?"
-  crawlpid=
-  test "$result" -eq 0 || ! result "error code $result" || return 1
-  result "OK"
-  grep -iE "^[0-9\:]*[[:space:]]Error:" "${tmp}/hts-log.txt" >&2
-
-  # now audit
-  while test "$#" -gt 0; do
-    case "$1" in
-    --no-purge)
-      nopurge=1
-      ;;
-    --summary)
-      grep -E "^HTTrack Website Copier/[^ ]* mirror complete in " "${tmp}/hts-log.txt"
-      ;;
-    --print-files)
-      find "${tmp}" -mindepth 1 -type f
-      ;;
-    --errors)
-      shift
-      assert_equals "checking errors" "$1" "$(grep -iEc "^[0-9\:]*[[:space:]]Error:" "${tmp}/hts-log.txt")"
-      ;;
-    --found)
-      shift
-      info "checking for $1"
-      if test -f "${tmp}/$1" ; then
-        result "OK"
-      else
-        result "not found"
-        exit 1
-      fi
-      ;;
-    --not-found)
-      shift
-      info "checking for $1"
-      if test -f "${tmp}/$1" ; then
-        result "OK"
-      else
-        result "not found"
-        exit 1
-      fi
-      ;;
-    --directory)
-      shift
-      info "checking for $1"
-      if test -d "${tmp}/$1" ; then
-        result "OK"
-      else
-        result "not found"
-        exit 1
-      fi
-      ;;
-    --files)
-      shift
-      nFiles=$(grep -E "^HTTrack Website Copier/[^ ]* mirror complete in " "${tmp}/hts-log.txt" \
-        | sed -e 's/.*[[:space:]]\([^ ]*\)[[:space:]]files written.*/\1/g')
-      assert_equals "checking files" "$1" "$nFiles"
-      ;;
-    httrack)
-      break;
-      ;;
-    esac
-    shift
-  done
-
-  # cleanup
-  if test -z "$nopurge"; then
     rm -rf "$tmp"
-  else
+    mkdir "$tmp" || ! warning "could not create $tmp" || return 1
-    tmpdir=
+
-  fi
+    which httrack >/dev/null || ! warning "could not find httrack" || return 1
+    ver=$(httrack -O /dev/null --version | sed -e 's/HTTrack version //')
+    test -n "$ver" || ! warning "could not run httrack" || return 1
+
+    # start crawl
+    log="${tmp}/log"
+    debug starting httrack -O "${tmp}" "${moreargs[@]}" "${@:pos}"
+    info "running httrack ${*:pos}"
+    httrack -O "${tmp}" --user-agent="httrack $ver ut ($(uname -omrs))" "${moreargs[@]}" "${@:pos}" >"${log}" 2>&1 &
+    crawlpid="$!"
+    debug "started cralwer on pid $crawlpid"
+    wait "$crawlpid"
+    result="$?"
+    crawlpid=
+    test "$result" -eq 0 || ! result "error code $result" || return 1
+    result "OK"
+    grep -iE "^[0-9\:]*[[:space:]]Error:" "${tmp}/hts-log.txt" >&2
+
+    # now audit
+    while test "$#" -gt 0; do
+        case "$1" in
+        --no-purge)
+            nopurge=1
+            ;;
+        --summary)
+            grep -E "^HTTrack Website Copier/[^ ]* mirror complete in " "${tmp}/hts-log.txt"
+            ;;
+        --print-files)
+            find "${tmp}" -mindepth 1 -type f
+            ;;
+        --errors)
+            shift
+            assert_equals "checking errors" "$1" "$(grep -iEc "^[0-9\:]*[[:space:]]Error:" "${tmp}/hts-log.txt")"
+            ;;
+        --found)
+            shift
+            info "checking for $1"
+            if test -f "${tmp}/$1"; then
+                result "OK"
+            else
+                result "not found"
+                exit 1
+            fi
+            ;;
+        --not-found)
+            shift
+            info "checking for $1"
+            if test -f "${tmp}/$1"; then
+                result "OK"
+            else
+                result "not found"
+                exit 1
+            fi
+            ;;
+        --directory)
+            shift
+            info "checking for $1"
+            if test -d "${tmp}/$1"; then
+                result "OK"
+            else
+                result "not found"
+                exit 1
+            fi
+            ;;
+        --files)
+            shift
+            nFiles=$(grep -E "^HTTrack Website Copier/[^ ]* mirror complete in " "${tmp}/hts-log.txt" |
+                sed -e 's/.*[[:space:]]\([^ ]*\)[[:space:]]files written.*/\1/g')
+            assert_equals "checking files" "$1" "$nFiles"
+            ;;
+        httrack)
+            break
+            ;;
+        esac
+        shift
+    done
+
+    # cleanup
+    if test -z "$nopurge"; then
+        rm -rf "$tmp"
+    else
+        tmpdir=
+    fi
 }
 
 # check args
@@ -195,7 +194,7 @@ tmpdir=
 crawlpid=
 nopurge=
 verbose=
-trap "cleanup" 0 1 2 3 4 5 6 7 8 9 11 13 14 15 16 19 24 25
+trap cleanup EXIT HUP INT QUIT ILL TRAP ABRT BUS FPE SEGV PIPE ALRM TERM STKFLT XCPU XFSZ
 
 # working directory
 tmpdir="${tmptopdir}/httrack_ut.$$"

tests/proxy-https-server.py

@@ -0,0 +1,151 @@
+#!/usr/bin/env python3
+"""Local CONNECT proxy + self-signed HTTPS origin for the issue #85 test.
+
+Starts a TLS origin server and an HTTP proxy that honours CONNECT, on ephemeral
+ports. Every request line the proxy receives (and any Proxy-Authorization) is
+appended to the proxy log; every header the origin receives over the tunnel is
+appended to the origin log. That lets the test assert both that an https crawl
+tunneled through the proxy and that proxy credentials never leaked to the origin.
+
+Proxy modes (argv[3], default "ok"):
+  ok    - honour CONNECT and tunnel to the origin
+  flood - answer 200 then stream headers forever with no blank line, to exercise
+          the client's bound on the proxy response (must not hang the crawl)
+
+Usage: proxy-https-server.py <cert.pem> <logdir> [mode]
+Prints "ORIGIN <port>", "PROXY <port>", then "ready" (one per line) on stdout.
+"""
+import http.server
+import os
+import socket
+import socketserver
+import ssl
+import sys
+import threading
+
+ORIGIN_BODY = b"<html><body>ORIGIN-PAGE-85</body></html>"
+PROXY_LOG = "proxy.log"
+ORIGIN_LOG = "origin-headers.log"
+
+
+def make_origin(logdir):
+    class Origin(http.server.BaseHTTPRequestHandler):
+        def do_GET(self):
+            with open(os.path.join(logdir, ORIGIN_LOG), "a") as handle:
+                for key in self.headers.keys():
+                    handle.write(key + "\n")
+            self.send_response(200)
+            self.send_header("Content-Type", "text/html")
+            self.send_header("Content-Length", str(len(ORIGIN_BODY)))
+            self.end_headers()
+            self.wfile.write(ORIGIN_BODY)
+
+        def log_message(self, *args):
+            pass
+
+    return Origin
+
+
+def start_origin(certfile, logdir):
+    httpd = socketserver.TCPServer(("127.0.0.1", 0), make_origin(logdir))
+    ctx = ssl.SSLContext(ssl.PROTOCOL_TLS_SERVER)
+    ctx.load_cert_chain(certfile)
+    httpd.socket = ctx.wrap_socket(httpd.socket, server_side=True)
+    port = httpd.socket.getsockname()[1]
+    threading.Thread(target=httpd.serve_forever, daemon=True).start()
+    return port
+
+
+def pipe(src, dst):
+    try:
+        while True:
+            data = src.recv(65536)
+            if not data:
+                break
+            dst.sendall(data)
+    except OSError:
+        pass
+    finally:
+        for sock in (src, dst):
+            try:
+                sock.shutdown(socket.SHUT_RDWR)
+            except OSError:
+                pass
+
+
+def handle_client(conn, logdir, mode):
+    rfile = conn.makefile("rb")
+    request_line = rfile.readline().decode("latin-1").strip()
+    auth = None
+    while True:
+        line = rfile.readline().decode("latin-1")
+        if line in ("\r\n", "\n", ""):
+            break
+        key, _, value = line.partition(":")
+        if key.strip().lower() == "proxy-authorization":
+            auth = value.strip()
+    with open(os.path.join(logdir, PROXY_LOG), "a") as handle:
+        handle.write(request_line + "\n")
+        if auth is not None:
+            handle.write("AUTH " + auth + "\n")
+    parts = request_line.split()
+    if not (len(parts) >= 2 and parts[0] == "CONNECT"):
+        conn.sendall(b"HTTP/1.0 501 Not Implemented\r\n\r\n")
+        conn.close()
+        return
+    if mode == "flood":
+        # 200, then an endless header stream with no terminating blank line: the
+        # client must bound this and give up, not hang.
+        try:
+            conn.sendall(b"HTTP/1.0 200 Connection established\r\n")
+            while True:
+                conn.sendall(b"X-Pad: 0123456789\r\n")
+        except OSError:
+            pass
+        conn.close()
+        return
+    host, _, port = parts[1].partition(":")
+    try:
+        upstream = socket.create_connection((host, int(port or 443)))
+    except OSError:
+        conn.sendall(b"HTTP/1.0 502 Bad Gateway\r\n\r\n")
+        conn.close()
+        return
+    conn.sendall(b"HTTP/1.0 200 Connection established\r\n\r\n")
+    threading.Thread(target=pipe, args=(conn, upstream), daemon=True).start()
+    pipe(upstream, conn)
+
+
+def start_proxy(logdir, mode):
+    srv = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
+    srv.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
+    srv.bind(("127.0.0.1", 0))
+    srv.listen(16)
+    port = srv.getsockname()[1]
+
+    def serve():
+        while True:
+            conn, _ = srv.accept()
+            threading.Thread(
+                target=handle_client, args=(conn, logdir, mode), daemon=True
+            ).start()
+
+    threading.Thread(target=serve, daemon=True).start()
+    return port
+
+
+def main():
+    certfile, logdir = sys.argv[1], sys.argv[2]
+    mode = sys.argv[3] if len(sys.argv) > 3 else "ok"
+    for name in (PROXY_LOG, ORIGIN_LOG):
+        open(os.path.join(logdir, name), "w").close()
+    origin_port = start_origin(certfile, logdir)
+    proxy_port = start_proxy(logdir, mode)
+    print("ORIGIN %d" % origin_port, flush=True)
+    print("PROXY %d" % proxy_port, flush=True)
+    print("ready", flush=True)
+    threading.Event().wait()
+
+
+if __name__ == "__main__":
+    main()

tests/run-all-tests.sh

@@ -2,19 +2,19 @@
 #
 
 error=0
-for i in *.test ; do
+for i in *.test; do
-	if bash $i ; then
+    if bash "$i"; then
-		echo "$i: passed" >&2
+        echo "$i: passed" >&2
-	else
+    else
-		echo "$i: ERROR" >&2
+        echo "$i: ERROR" >&2
-		error=$[${error}+1]
+        error=$((error + 1))
-	fi
+    fi
 done
 
 if test "$error" -eq 0; then
-	echo "all tests passed" >&2
+    echo "all tests passed" >&2
 else
-	echo "${error} test(s) failed" >&2
+    echo "${error} test(s) failed" >&2
 fi
 
 exit $error