From 5b119c8e7fb67ef78335c1542cedb33339613555 Mon Sep 17 00:00:00 2001
From: Steve Pinkham <steve.pinkham@gmail.com>
Date: Fri, 10 Sep 2010 12:59:06 -0400
Subject: [PATCH] 1.65b: dictionary & CSS MIME sniffing improvements

  - Relaxed MIME matching on claimed CSS/JS that fails MIME sniffing
logic.
  - Proper detection of @media in CSS.
---
 ChangeLog                |  7 +++++++
 Makefile                 |  2 +-
 analysis.c               | 11 ++++++++---
 config.h                 |  1 +
 dictionaries/complete.wl |  2 ++
 dictionaries/default.wl  |  2 ++
 dictionaries/minimal.wl  |  2 ++
 7 files changed, 23 insertions(+), 4 deletions(-)

diff --git a/ChangeLog b/ChangeLog
index ac4bcd5..c3dbbff 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,10 @@
+Version 1.65b:
+--------------
+
+  - Relaxed MIME matching on claimed CSS/JS that fails MIME sniffing logic.
+
+  - Proper detection of @media in CSS.
+
 Version 1.64b:
 --------------
 
diff --git a/Makefile b/Makefile
index 4fd80a4..53fc643 100644
--- a/Makefile
+++ b/Makefile
@@ -20,7 +20,7 @@
 #
 
 PROGNAME   = skipfish
-VERSION    = 1.64b
+VERSION    = 1.65b
 
 OBJFILES   = http_client.c database.c crawler.c analysis.c report.c
 INCFILES   = alloc-inl.h string-inl.h debug.h types.h http_client.h \
diff --git a/analysis.c b/analysis.c
index 8d46701..60b0798 100644
--- a/analysis.c
+++ b/analysis.c
@@ -1066,9 +1066,10 @@ static u8 is_css(struct http_response* res) {
 
     }
 
-    /* @import or @charset is a clear indicator of CSS. */
+    /* @import, @media, or @charset is a clear indicator of CSS. */
 
     if (*text == '@' && (!strncasecmp((char*)text + 1, "import", 6) ||
+        !strncasecmp((char*)text + 1, "media", 5) ||
         !strncasecmp((char*)text + 1, "charset", 7))) {
       res->css_type = 2;
       return 1;
@@ -1562,7 +1563,7 @@ void content_checks(struct http_request* req, struct http_response* res) {
 
     if (*tmp == '<') {
       u8* tag_name;
-      u32 len = strcspn((char*)++tmp, "> \t\r\n"), space_len;
+      u32 len = strcspn((char*)++tmp, ">= \t\r\n"), space_len;
       u8  remote_script = 0;
 
       /* Skip comments where possible. */
@@ -1809,9 +1810,13 @@ binary_checks:
      value... and also failed to sniff.
 
      Mismatch between MIME_ASC_HTML and MIME_XML_XHTML is not worth
-     complaining about, too. */
+     complaining about; the same about JS or CSS responses being
+     sniffed as "unknown ASCII". */
 
   if (res->sniff_mime_id != res->decl_mime_id &&
+      !((res->decl_mime_id == MIME_ASC_JAVASCRIPT ||
+         res->decl_mime_id == MIME_ASC_CSS) && 
+        res->sniff_mime_id == MIME_ASC_GENERIC) &&
       !(res->decl_mime_id == MIME_ASC_HTML && 
         res->sniff_mime_id == MIME_XML_XHTML) &&
       !(res->decl_mime_id == MIME_XML_XHTML && 
diff --git a/config.h b/config.h
index e0899a8..09ebca3 100644
--- a/config.h
+++ b/config.h
@@ -231,6 +231,7 @@ static const char* form_suggestion[][2] = {
   { "user"     , "skipfish" },
   { "nick"     , "skipfish" },
   { "pass"     , "skipfish" },
+  { "pwd"      , "skipfish" },
   { "year"     , "2010" },
   { "card"     , "4111111111111111" }, /* Reserved */
   { "code"     , "000" },
diff --git a/dictionaries/complete.wl b/dictionaries/complete.wl
index 239f567..7f3e3f2 100644
--- a/dictionaries/complete.wl
+++ b/dictionaries/complete.wl
@@ -1221,6 +1221,7 @@ w 1 1 1 online
 w 1 1 1 op
 w 1 1 1 open
 w 1 1 1 openbsd
+w 1 1 1 opencart
 w 1 1 1 opendir
 w 1 1 1 openejb
 w 1 1 1 openjpa
@@ -1986,6 +1987,7 @@ w 1 1 1 zboard
 w 1 1 1 zencart
 w 1 1 1 zend
 w 1 1 1 zero
+w 1 1 1 zimbra
 w 1 1 1 zipfiles
 w 1 1 1 zips
 w 1 1 1 zoom
diff --git a/dictionaries/default.wl b/dictionaries/default.wl
index eebdfbf..28d2444 100644
--- a/dictionaries/default.wl
+++ b/dictionaries/default.wl
@@ -1209,6 +1209,7 @@ w 1 1 1 online
 w 1 1 1 op
 w 1 1 1 open
 w 1 1 1 openbsd
+w 1 1 1 opencart
 w 1 1 1 opendir
 w 1 1 1 openejb
 w 1 1 1 openjpa
@@ -1985,6 +1986,7 @@ w 1 1 1 zboard
 w 1 1 1 zencart
 w 1 1 1 zend
 w 1 1 1 zero
+w 1 1 1 zimbra
 w 1 1 1 zipfiles
 w 1 1 1 zips
 w 1 1 1 zoom
diff --git a/dictionaries/minimal.wl b/dictionaries/minimal.wl
index 7912a2a..76dcaf7 100644
--- a/dictionaries/minimal.wl
+++ b/dictionaries/minimal.wl
@@ -1192,6 +1192,7 @@ w 1 1 1 online
 w 1 1 1 op
 w 1 1 1 open
 w 1 1 1 openbsd
+w 1 1 1 opencart
 w 1 1 1 opendir
 w 1 1 1 openejb
 w 1 1 1 openjpa
@@ -1982,6 +1983,7 @@ w 1 1 1 zboard
 w 1 1 1 zencart
 w 1 1 1 zend
 w 1 1 1 zero
+w 1 1 1 zimbra
 w 1 1 1 zipfiles
 w 1 1 1 zips
 w 1 1 1 zoom