Limit the final destination path lenght to 256 (Windows compatibility)

See http://code.google.com/p/httrack/issues/detail?id=9

3 files changed, 69 insertions, 2 deletions

diff --git a/src/htscharset.c b/src/htscharset.c
index f19e47c..351fa8c 100644
--- a/src/htscharset.c
+++ b/src/htscharset.c
@@ -502,6 +502,18 @@ static int is_space_or_equal_or_quote(char c) {
   return is_space_or_equal(c) || c == '"' || c == '\'';
 }
 
+size_t hts_stringLengthUTF8(const char *s) {
+  const unsigned char *const bytes = (const unsigned char*) s;
+  size_t i, len;
+  for(i = 0, len = 0 ; bytes[i] != '\0' ; i++) {
+    const unsigned char c = bytes[i];
+    if (HTS_IS_LEADING_UTF8(c)) {  // ASCII or leading byte
+      len++;
+    }
+  }
+  return len;
+}
+
 char* hts_getCharsetFromMeta(const char *html, size_t size) {
   int i;
   // <META HTTP-EQUIV="CONTENT-TYPE" CONTENT="text/html; charset=utf-8" >
diff --git a/src/htscharset.h b/src/htscharset.h
index f8f4ca6..55aa2a8 100644
--- a/src/htscharset.h
+++ b/src/htscharset.h
@@ -44,6 +44,9 @@ Please visit our Website: http://www.httrack.com
 #include <windows.h>
 #endif
 
+/** Leading character (ASCII or leading UTF-8 sequence) **/
+#define HTS_IS_LEADING_UTF8(C) ((unsigned char)(C) < 0x80 || (unsigned char)(C) >= 0xc0)
+
 /**
  * Convert the string "s" from charset "charset" to UTF-8.
  * Return NULL upon error.
@@ -78,6 +81,11 @@ extern char *hts_convertUCS2StringToUTF8(LPWSTR woutput, int wsize);
  **/
 extern char *hts_convertStringSystemToUTF8(const char *s, size_t size);
 
+/**
+ * Get an UTF-8 string length in characters.
+ **/
+extern size_t hts_stringLengthUTF8(const char *s);
+
 #endif
 
 #endif
diff --git a/src/htsname.c b/src/htsname.c
index 641e7aa..cc6c344 100644
--- a/src/htsname.c
+++ b/src/htsname.c
@@ -1380,7 +1380,55 @@ int url_savename2(char* adr_complete, char* fil_complete, char* save,
 		} else if (!IS_DELAYED_EXT(save)) {
 			strcatbuff(lastDot, "." DELAYED_EXT);
 		}
-	}
+  }
+
+  // enforce 256-character path limit before inserting destination path
+#define HTS_MAX_PATH_LEN 250
+#define MIN_LAST_SEG_RESERVE 30
+  if (hts_stringLengthUTF8(save) + hts_stringLengthUTF8(StringBuff(opt->path_html_utf8)) >= HTS_MAX_PATH_LEN) {
+    char BIGSTK tempo[HTS_URLMAXSIZE*2];
+    const size_t parentLen = hts_stringLengthUTF8(StringBuff(opt->path_html_utf8));
+    // parent path length is not insane (otherwise, ignore and pick 200 as suffix length)
+    const size_t maxLen = parentLen < HTS_MAX_PATH_LEN/2 ? HTS_MAX_PATH_LEN - parentLen : HTS_MAX_PATH_LEN;
+    size_t i, j, lastSeg, sofar;
+    // pick up last segment
+    for(i = 0, lastSeg = 0 ; save[i] != '\0' ; i++) {
+      if (save[i] == '/') {
+        lastSeg = i + 1;
+      }
+    }
+    // add as much pathes as we can
+    for(i = 0, sofar = 0 ; i < lastSeg && i + MIN_LAST_SEG_RESERVE < maxLen ; i++) {
+      tempo[i] = save[i];
+      if (save[i] == '/') {
+        // validate segment so far
+        sofar = i + 1;
+      }
+    }
+    // last segment
+#define MAX_UTF8_SEQ_CHARS 4
+    for(j = 0 ; save[j + i] != '\0' ; j++) {
+      // Stop here before next sequence
+      if (sofar + j + MAX_UTF8_SEQ_CHARS >= maxLen && HTS_IS_LEADING_UTF8(save[i + j])) {
+        break;
+      }
+      // Stop is overflowing
+      else if (sofar + j >= maxLen) {
+        break;
+      }
+      save[sofar + j] = save[i + j];
+    }
+    // terminating \0
+    save[sofar + j] = '\0';
+    // log in debug
+    if ( (opt->debug>1) && (opt->log!=NULL) ) {
+      HTS_LOG(opt,LOG_DEBUG); fprintf(opt->log,"Too long filename shortened: %s%s => %s"LF,adr_complete,fil_complete,sav);
+      test_flush;
+    }
+  }
+#undef MAX_UTF8_SEQ_CHARS
+#undef MIN_LAST_SEG_RESERVE
+#undef HTS_MAX_PATH_LEN
 
   // chemin primaire éventuel A METTRE AVANT
   if (strnotempty(StringBuff(opt->path_html_utf8))) {
@@ -1390,7 +1438,6 @@ int url_savename2(char* adr_complete, char* fil_complete, char* save,
     strcpybuff(save,tempo);
   }
 
-
   // vérifier que le nom n'est pas déja pris...
   if (liens!=NULL) { 
     int nom_ok;