Merge pull request #393 from xroche/fix/empty-footer-doitlog-106

Keep empty quoted args when reloading doit.log (#106)

src/htsback.c

@@ -3838,7 +3838,7 @@ 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/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++;
                 }
@@ -3096,6 +3096,78 @@ 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/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

@@ -874,6 +874,50 @@ 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,
@@ -1048,34 +1092,9 @@ 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) {
@@ -3646,8 +3665,9 @@ 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 +3951,8 @@ 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 +3965,8 @@ 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 +4216,7 @@ 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 +4234,7 @@ 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 +5256,8 @@ 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 +5487,10 @@ 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 +5515,7 @@ 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,

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

@@ -3722,7 +3722,8 @@ 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 +3736,7 @@ 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 +4262,10 @@ 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 +4375,7 @@ 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 +4466,8 @@ 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 +4851,7 @@ 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)

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/Makefile.am

@@ -24,6 +24,8 @@ 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 \