Anchor the locale test's lift so it cannot silently stop testing

The lift ended at the first blank line, so a blank line grown inside the
locale block would truncate it: reinstating the old LANG-only bug below one
left the test passing 14/14 while the launcher returned English. Anchor on
the next section instead, and pin exported-empty as unset, which lang_of
cannot express.

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

lang/Finnish.txt

@@ -930,6 +930,7 @@ Server terminated
 Palvelin lopetettu
 A fatal error has occurred during this mirror
 Tällä peilillä tapahtui vakava virhe
+
 Proxy type:
 Välityspalvelimen tyyppi:
 Proxy protocol. HTTP: standard proxy. HTTP (CONNECT tunnel): sends every request through a CONNECT tunnel, for CONNECT-only proxies like Tor's HTTPTunnelPort. SOCKS5: default port 1080.

lang/Italiano.txt

@@ -374,7 +374,7 @@ Create error logging and report files
 Crea file di log per segnalare errori e informazioni
 Generate DOS 8-3 filenames ONLY
 Genera solo nomi di file in formato 8.3
-Generate ISO9660 filenames ONLY for CDROM medias
+Generate ISO 9660 filenames ONLY for CDROM medias
 Genera nomi di file in formato ISO9660 per i CDROM
 Do not create HTML error pages
 Non generare pagine d'errore HTML

lang/Portugues-Brasil.txt

@@ -918,7 +918,7 @@ normal\nextended\ndebug
 normal\nextendido\ncorrigir
 Download web site(s)\nDownload web site(s) + questions\nGet individual files\nDownload all sites in pages (multiple mirror)\nTest links in pages (bookmark test)\n* Continue interrupted download\n* Update existing download
 Copiar site(s) da Web\nCopiar site(s) interativos da Web (perguntas)\nReceber arquivos específicos\nCopiar todas as páginas do site (alternação múltipla)\nTestar links nas páginas (testar indicador)\n* Retomar download interrompido\n* Atualizar download existente
-Relative URI / Absolute URL (default)\nAbsolute URL / Absolute URL\nAbsolute URI / Absolute URL\nOriginal URL / Original URL
+Relative URL / Absolute URL (default)\nAbsolute URL / Absolute URL\nAbsolute URI / Absolute URL\nOriginal URL / Original URL
 URI Relativa / URL Absoluta (padrão)\nURL Absoluta / URL Absoluta\nURI Absoluta / URL Absoluta\nURL Original / URL Original
 Open Source offline browser
 Abrir origem offline no navegador

src/htsalias.c

@@ -567,29 +567,24 @@ int optinclude_file(const char *name, int *argc, char **argv, char *x_argvblk,
   return 0;
 }
 
-/* Get home directory, '.' if unset or empty */
+/* Get home directory, '.' if failed */
 /* example: /home/smith */
 const char *hts_gethome(void) {
   const char *home = getenv("HOME");
 
-  /* An empty $HOME would expand ~/foo into the absolute /foo */
-  return strnotempty(home) ? home : ".";
+  if (home)
+    return home;
+  else
+    return ".";
 }
 
-/* Convert ~/foo into /home/smith/foo (~user/ left alone: no getpwnam here) */
+/* Convert ~/foo into /home/smith/foo */
 void expand_home(String * str) {
-  if (StringNotEmpty(*str) && StringSub(*str, 0) == '~' &&
-      (StringLength(*str) == 1 || StringSub(*str, 1) == '/')) {
+  if (StringSub(*str, 1) == '~') {
     char BIGSTK tempo[HTS_URLMAXSIZE * 2];
-    const char *const home = hts_gethome();
-    const size_t homelen = strlen(home);
-    const size_t taillen = StringLength(*str) - 1;
 
-    /* Leave untouched rather than abort() in strcatbuff on a huge $HOME */
-    if (taillen < sizeof(tempo) && homelen < sizeof(tempo) - taillen) {
-      strcpybuff(tempo, home);
-      strcatbuff(tempo, StringBuff(*str) + 1);
-      StringCopy(*str, tempo);
-    }
+    strcpybuff(tempo, hts_gethome());
+    strcatbuff(tempo, StringBuff(*str) + 1);
+    StringCopy(*str, tempo);
   }
 }

src/htsdns_selftest.c

@@ -402,50 +402,6 @@ int dns_selftests(httrackp *opt) {
       deletesoc(s);
   }
 
-  /* A URL port outside 1..65535 must refuse the link, not fold into range and
-     connect elsewhere (#614). *addr_count discriminates: 0 only if refused
-     before the resolve, still 2 for one merely truncated or defaulted. */
-  {
-    /* an empty "dual.test:" means the default port (WHATWG, curl): keep it */
-    static const char *const good[] = {"dual.test:1",    "dual.test:80",
-                                       "dual.test:8080", "dual.test:65535",
-                                       "dual.test:080",  "dual.test:"};
-    /* 65616 and 4294967376 are load-bearing: both wrap to a plausible 80 */
-    static const char *const bad[] = {
-        "dual.test:0",          "dual.test:65536",      "dual.test:65616",
-        "dual.test:99999",      "dual.test:2147483648", "dual.test:4294967296",
-        "dual.test:4294967376", "dual.test:-1",         "dual.test:-23437",
-        "dual.test:80x",        "dual.test:+80",        "dual.test: 80",
-        "dual.test:0x50"};
-    size_t k;
-
-    for (k = 0; k < sizeof(good) / sizeof(good[0]); k++) {
-      htsblk r;
-      int count = -1;
-      T_SOC s;
-
-      hts_init_htsblk(&r);
-      s = newhttp_addr(opt, good[k], &r, -1, 0, 0, &count);
-      CHECK(count == 2); /* accepted: reached the resolve */
-      if (s != INVALID_SOCKET)
-        deletesoc(s);
-    }
-
-    for (k = 0; k < sizeof(bad) / sizeof(bad[0]); k++) {
-      htsblk r;
-      int count = -1;
-      T_SOC s;
-
-      hts_init_htsblk(&r);
-      s = newhttp_addr(opt, bad[k], &r, -1, 0, 0, &count);
-      CHECK(s == INVALID_SOCKET);
-      CHECK(count == 0); /* refused before resolving, not a failed connect */
-      CHECK(strstr(r.msg, "Invalid port") != NULL);
-      if (s != INVALID_SOCKET)
-        deletesoc(s);
-    }
-  }
-
   /* Connect-fallback decision (consumer of the multi-address list): when a
      stuck connect should abandon the current address for the next one. */
   {

src/htslib.c

@@ -2167,18 +2167,15 @@ T_SOC newhttp_addr(httrackp *opt, const char *_iadr, htsblk *retour, int port,
 #endif
 
       if (a != NULL) {
-        // folding a nonsense port into 0..65535 crawls one neither the link nor
-        // a port filter named; an empty "host:" just means the default (#614)
-        if (a[1] != '\0' && !hts_parse_url_port(a + 1, &port)) {
-          if (retour != NULL) {
-            snprintf(retour->msg, sizeof(retour->msg), "Invalid port: %s",
-                     a + 1);
-          }
-          return INVALID_SOCKET;
-        }
+        int i = -1;
 
         iadr2[0] = '\0';
-        // the address itself, without the ":port"
+        sscanf(a + 1, "%d", &i);
+        if (i != -1) {
+          port = (unsigned short int) i;
+        }
+
+        // adresse véritable (sans :xx)
         strncatbuff(iadr2, iadr, (int) (a - iadr));
         resolve_host = iadr2;
       }
@@ -3751,27 +3748,18 @@ static int proxy_default_port(const char *arg) {
   return hts_proxy_is_socks(arg) ? 1080 : 8080;
 }
 
-hts_boolean hts_parse_url_port(const char *a, int *port) {
+// port "a" of -P argument "arg": digits fitting TCP's 1..65535, else the scheme
+// default. Not sscanf("%d"): past INT_MAX it wraps to a garbage port (#602)
+static int parse_proxy_port(const char *a, const char *arg) {
   char *end;
   long p;
 
   if (!isdigit((unsigned char) *a)) // strtol would eat a sign or leading space
-    return HTS_FALSE;
+    return proxy_default_port(arg);
   p = strtol(a, &end, 10);
   if (*end != '\0' || p < 1 || p > 65535) // ERANGE lands out of range too
-    return HTS_FALSE;
-  *port = (int) p;
-  return HTS_TRUE;
-}
-
-// port "a" of -P argument "arg": digits fitting TCP's 1..65535, else the scheme
-// default. Not sscanf("%d"): past INT_MAX it wraps to a garbage port (#602)
-static int parse_proxy_port(const char *a, const char *arg) {
-  int port;
-
-  if (!hts_parse_url_port(a, &port))
     return proxy_default_port(arg);
-  return port;
+  return (int) p;
 }
 
 void hts_parse_proxy(const char *arg, char *name, size_t name_size, int *port) {

src/htslib.h

@@ -288,12 +288,6 @@ const char *strrchr_limit(const char *s, char c, const char *limit);
 char *jump_protocol(char *source);
 const char *jump_protocol_const(const char *source);
 
-/* Parse a URL's port text "a" (after the ':', up to the end of the string):
-   TRUE and *port set for a bare decimal in 1..65535, else FALSE and *port left
-   alone. Not sscanf("%d"), which range-checks nothing and wraps past INT_MAX.
- */
-hts_boolean hts_parse_url_port(const char *a, int *port);
-
 /* Split a -P proxy argument "[scheme://][user:pass@]host[:port]" into the proxy
    host string (scheme and any user:pass kept, for later stripping and auth),
    written NUL-terminated into name[name_size] (truncated to fit), and the port

src/htsproxy.c

@@ -491,11 +491,14 @@ static int socks5_handshake_stream(httrackp *opt, socks5_stream *st,
     if ((unsigned char) host[i] < ' ')
       return socks5_fail(msg, msgsize, "SOCKS5: invalid origin host");
   }
-  // the old range check ran after sscanf("%d") had wrapped a huge value into a
-  // plausible port (#614). An empty "host:" stays refused here, unlike the
-  // direct path, as it was before #614.
-  if (portsep != NULL && !hts_parse_url_port(portsep + 1, &port))
-    return socks5_fail(msg, msgsize, "SOCKS5: invalid origin port");
+  if (portsep != NULL) {
+    int p = -1;
+
+    sscanf(portsep + 1, "%d", &p);
+    if (p <= 0 || p > 65535)
+      return socks5_fail(msg, msgsize, "SOCKS5: invalid origin port");
+    port = p;
+  }
   if (link_has_authorization(proxy_name)) {
     if (!socks5_credentials(opt, proxy_name, user, sizeof(user), &userlen, pass,
                             sizeof(pass), &passlen, msg, msgsize))

src/htsselftest.c

@@ -45,7 +45,6 @@ Please visit our Website: http://www.httrack.com
 #include "htscore.h"
 #include "htsdefines.h"
 #include "htslib.h"
-#include "htsalias.h"
 #include "htsparse.h"
 #include "htscache.h"
 #include "htscache_selftest.h"
@@ -826,21 +825,6 @@ static int st_simplify(httrackp *opt, int argc, char **argv) {
   return 0;
 }
 
-static int st_expandhome(httrackp *opt, int argc, char **argv) {
-  String path = STRING_EMPTY;
-
-  (void) opt;
-  if (argc < 1) {
-    fprintf(stderr, "expandhome: needs a path\n");
-    return 1;
-  }
-  StringCopy(path, argv[0]);
-  expand_home(&path);
-  printf("expanded=%s\n", StringBuff(path));
-  StringFree(path);
-  return 0;
-}
-
 static int st_mime(httrackp *opt, int argc, char **argv) {
   char mime[256];
 
@@ -1469,25 +1453,6 @@ static int st_socks5(httrackp *opt, int argc, char **argv) {
   assertf(socks5_handshake_scripted(opt, "origin.test:8443", proxy, &io) == 1);
   assertf(memcmp(io.sent + io.sent_len - 2, "\x20\xfb", 2) == 0);
 
-  /* a bad origin port is refused before any byte goes out (#614). 4294967376 is
-     the case the old range check could not see: it overflowed the sscanf("%d")
-     into a plausible 80 and passed. 65616 would not prove anything here, since
-     it fits an int and the old check already caught it. */
-  {
-    static const char *const bad[] = {"origin.test:4294967376",
-                                      "origin.test:80x", "origin.test:+80",
-                                      "origin.test: 80", "origin.test:8.0"};
-    size_t k;
-
-    for (k = 0; k < sizeof(bad) / sizeof(bad[0]); k++) {
-      len = socks5_reply(script, 0x01, v4, sizeof(v4));
-      io.reply = script;
-      io.reply_len = len;
-      assertf(socks5_handshake_scripted(opt, bad[k], proxy, &io) == 0);
-      assertf(io.sent_len == 0);
-    }
-  }
-
   /* credentials: split on the first colon of the escaped userinfo, so %3a stays
      inside the username and a colon in the password is not a delimiter */
   {
@@ -3114,7 +3079,6 @@ static const struct selftest_entry {
     {"filterbounds", "", "matcher length/work caps reject hostile patterns",
      st_filterbounds},
     {"simplify", "<path>", "collapse ./ and ../ in a path", st_simplify},
-    {"expandhome", "<path>", "expand a leading ~/ into $HOME", st_expandhome},
     {"stripquery", "", "--strip-query pattern/key stripping self-test",
      st_stripquery},
     {"urlhack", "", "-%u url-hack sub-flag (www/slash/query) self-test",

tests/01_engine-expandhome.test

@@ -1,53 +0,0 @@
-#!/bin/bash
-#
-
-# SC2088: the literal, unexpanded '~' is the input under test.
-# shellcheck disable=SC2088
-
-set -euo pipefail
-
-# ~ expansion for the -O base path (expand_home). $HOME is pinned so the cases
-# cannot depend on the caller's, but MSYS rewrites it into a Windows path
-# before the native httrack.exe reads it: ask the engine what it saw rather
-# than hardcoding the value.
-ask() {
-    HOME=HTSHOME httrack -O /dev/null -#test=expandhome "$1"
-}
-exp() {
-    test "$(ask "$1")" == "expanded=$2" || exit 1
-}
-
-home=$(ask '~')
-home=${home#expanded=}
-
-# or every case below is vacuous: '~' means expansion never fired, '.' means it
-# fell back to the default without ever reading $HOME
-test "$home" != '~' || exit 1
-test "$home" != '.' || exit 1
-
-exp '~' "$home"
-exp '~/foo' "$home/foo"
-exp '~/foo/bar/#' "$home/foo/bar/#"
-exp '~/' "$home/"
-exp '~/../x' "$home/../x"
-
-# ~user/ needs getpwnam: left alone rather than mangled into $HOME + "user/"
-exp '~smith/foo' '~smith/foo'
-exp '~root' '~root'
-
-# only a leading '~' expands; anything else is a literal path component
-exp 'a~/x' 'a~/x'
-exp './~/x' './~/x'
-exp 'foo~bar' 'foo~bar'
-
-# an unset or empty $HOME must not turn ~/foo into the absolute /foo
-test "$(env -u HOME httrack -O /dev/null -#test=expandhome '~/foo')" == "expanded=./foo" || exit 1
-test "$(HOME='' httrack -O /dev/null -#test=expandhome '~/foo')" == "expanded=./foo" || exit 1
-
-# A $HOME past the 2 * HTS_URLMAXSIZE buffer is left alone: strcatbuff aborts
-# rather than truncates. Only meaningful where the engine sees the value we set:
-# MSYS rewrites HOME into a path of its own choosing, long or not.
-if test "$home" = HTSHOME; then
-    long=$(printf '%04096d' 0 | tr '0' 'A')
-    test "$(HOME="$long" httrack -O /dev/null -#test=expandhome '~/foo')" == "expanded=~/foo" || exit 1
-fi

tests/52_local-socks5.test

@@ -165,13 +165,6 @@ auth="$tmpdir/auth"
 start_server "$auth" ok
 run_crawl "$auth/out" "http://${host}:${http_port}/" \
     "socks5://user:secret@127.0.0.1:${socks_port}"
-# name a rejected version byte here: the crawl check below only knows the
-# handshake failed, not why
-grep -q "^AUTHVER-BAD" "$auth/socks.log" && {
-    echo "FAIL: sub-negotiation version was not RFC 1929's 0x01" >&2
-    cat "$auth/socks.log" >&2
-    exit 1
-}
 grep -rq "ORIGIN-PAGE-563" "$auth/out" || {
     echo "FAIL: origin not downloaded through the authenticated socks proxy" >&2
     cat "$auth/out.log" "$auth/socks.log" >&2

tests/62_lang-integrity.test

@@ -1,90 +0,0 @@
-#!/bin/bash
-#
-# lang/*.txt are line pairs: an English msgid, then its translation. A stray
-# blank or a drifted msgid silently unhooks every translation that follows,
-# rather than failing, so assert the pairing and the join against English.txt.
-# Pairs are physical lines here; a msgid ending in \ continues onto the next one
-# for the engine (linput_cpp) but not for us. None do today.
-
-set -euo pipefail
-
-langdir="${top_srcdir:-..}/lang"
-def="${top_srcdir:-..}/lang.def"
-eng="$langdir/English.txt"
-for f in "$eng" "$def"; do
-    [ -f "$f" ] || {
-        echo "cannot find $f"
-        exit 1
-    }
-done
-
-tmp=$(mktemp -d)
-trap 'rm -rf "$tmp"' EXIT
-keys="$tmp/english.keys"
-
-# Byte-wise: each file is in its own declared legacy charset, lang.def is CRLF.
-LC_ALL=C awk 'NR%2==1 { sub(/\r$/, ""); print }' "$eng" >"$keys"
-
-# An empty parse would make every check below pass vacuously.
-nkeys=$(wc -l <"$keys")
-if [ "$nkeys" -lt 400 ]; then
-    echo "only $nkeys msgids parsed from $eng; the parse is broken"
-    exit 1
-fi
-
-nfiles=0
-fail=0
-for f in "$langdir"/*.txt; do
-    nfiles=$((nfiles + 1))
-    LC_ALL=C awk -v keys="$keys" -v name="${f##*/}" '
-        BEGIN { while ((getline k < keys) > 0) known[k] = 1 }
-        { sub(/\r$/, "") }
-        NR % 2 == 1 {
-            if ($0 == "")
-                printf "%s:%d: blank line where an English msgid belongs\n", name, NR
-            else if (!($0 in known))
-                printf "%s:%d: msgid absent from English.txt: %s\n", name, NR, $0
-            else
-                next
-            bad++
-        }
-        END {
-            if (NR % 2) {
-                printf "%s: odd line count (%d): msgid/translation pairing is broken\n", name, NR
-                bad++
-            }
-            exit(bad > 0)
-        }
-    ' "$f" || fail=1
-done
-
-if [ "$nfiles" -lt 25 ]; then
-    echo "only $nfiles language files found in $langdir"
-    exit 1
-fi
-
-# lang.def maps each LANG_* macro to the English string the GUI compiles in,
-# which is the lookup key into every lang/*.txt: no msgid, no translations.
-LC_ALL=C awk -v keys="$keys" '
-    BEGIN { while ((getline k < keys) > 0) known[k] = 1 }
-    { sub(/\r$/, "") }
-    NR % 2 == 1 { macro = $0; next }
-    macro ~ /^(LANG_|LISTDEF_)/ {
-        n++
-        if (!($0 in known)) {
-            printf "lang.def:%d: %s maps to \"%s\", which is not a msgid in English.txt\n", NR, macro, $0
-            bad++
-        }
-    }
-    END {
-        if (n < 400) {
-            printf "lang.def: only %d LANG_ entries parsed; the parse is broken\n", n
-            bad++
-        }
-        exit(bad > 0)
-    }
-' "$def" || fail=1
-
-[ "$fail" -eq 0 ] || exit 1
-
-echo "$nfiles language files consistent with $nkeys English msgids"

tests/Makefile.am

@@ -59,7 +59,6 @@ TESTS = \
 	01_engine-pause.test \
 	01_engine-rcfile.test \
 	01_engine-reconcile.test \
-	01_engine-expandhome.test \
 	01_engine-fsize.test \
 	01_engine-redirect.test \
 	01_engine-relative.test \
@@ -141,7 +140,6 @@ TESTS = \
 	58_watchdog.test \
 	59_local-tls-stall.test \
 	60_crawl-log-salvage.test \
-	61_webhttrack-locale.test \
-	62_lang-integrity.test
+	61_webhttrack-locale.test
 
 CLEANFILES = check-network_sh.cache

tests/socks5-server.py

@@ -24,7 +24,6 @@ from proxytestlib import bind_ephemeral, pipe  # noqa: E402
 # The one name the proxy answers for; a .invalid TLD never resolves (RFC 6761),
 # so a locally-resolving client could not reach us -- success proves remote DNS.
 REMOTE_HOST = b"socks-origin.invalid"
-AUTH_VERSION = 0x01  # RFC 1929 sub-negotiation version
 # index links the subpages so an -r3 crawl reuses one keep-alive socket
 LINKS = "".join('<a href="/p%d.html">%d</a>' % (i, i) for i in range(1, 6))
 ORIGIN_BODY = ("<html><body>ORIGIN-PAGE-563 " + LINKS + "</body></html>").encode()
@@ -105,12 +104,6 @@ def negotiate_auth(conn, logdir):
     if 0x02 in methods:  # prefer user/pass so the auth test exercises RFC 1929
         conn.sendall(b"\x05\x02")
         (subver,) = recvn(conn, 1)
-        # RFC 1929's version byte is 0x01, not SOCKS5's 0x05: a real proxy
-        # rejects a mismatch instead of tunnelling anyway
-        if subver != AUTH_VERSION:
-            log(logdir, "AUTHVER-BAD %d" % subver)
-            conn.sendall(bytes([AUTH_VERSION, 0x01]))  # sub-negotiation failure
-            return False
         (ulen,) = recvn(conn, 1)
         uname = recvn(conn, ulen)
         (plen,) = recvn(conn, 1)