git commit: TS-1094: Handle buffer with just CR in it correctly for MIME parsing.

[am...@apache.org]

Updated Branches:
  refs/heads/master 498003b5e -> b55801c81


TS-1094: Handle buffer with just CR in it correctly for MIME parsing.


Project: http://git-wip-us.apache.org/repos/asf/trafficserver/repo
Commit: http://git-wip-us.apache.org/repos/asf/trafficserver/commit/b55801c8
Tree: http://git-wip-us.apache.org/repos/asf/trafficserver/tree/b55801c8
Diff: http://git-wip-us.apache.org/repos/asf/trafficserver/diff/b55801c8

Branch: refs/heads/master
Commit: b55801c814dff141b3d287f42e943e21a850a993
Parents: 498003b
Author: Alan M. Carroll <am...@network-geographics.com>
Authored: Wed Feb 1 10:51:02 2012 -0600
Committer: Alan M. Carroll <am...@network-geographics.com>
Committed: Wed Feb 1 19:19:38 2012 -0600

----------------------------------------------------------------------
 proxy/hdrs/MIME.cc |   34 ++++++++++++++++++++++++++++++----
 proxy/hdrs/MIME.h  |    1 +
 2 files changed, 31 insertions(+), 4 deletions(-)
----------------------------------------------------------------------


http://git-wip-us.apache.org/repos/asf/trafficserver/blob/b55801c8/proxy/hdrs/MIME.cc
----------------------------------------------------------------------
diff --git a/proxy/hdrs/MIME.cc b/proxy/hdrs/MIME.cc
index e2e3cbb..2fa5f66 100644
--- a/proxy/hdrs/MIME.cc
+++ b/proxy/hdrs/MIME.cc
@@ -2196,6 +2196,8 @@ mime_scanner_get(MIMEScanner *S,
 {
   const char *raw_input_c, *lf_ptr;
   MIMEParseResult zret = PARSE_CONT;
+  // Need this for handling dangling CR.
+  static char const RAW_CR = ParseRules::CHAR_CR;
 
   ink_debug_assert((raw_input_s != NULL) && (*raw_input_s != NULL));
   ink_debug_assert(raw_input_e != NULL);
@@ -2206,11 +2208,35 @@ mime_scanner_get(MIMEScanner *S,
     ptrdiff_t runway = raw_input_e - raw_input_c; // remaining input.
     switch (S->m_state) {
     case MIME_PARSE_BEFORE: // waiting to find a field.
-      // If we find leading CR LF then it's the last line of the header.
-      if (ParseRules::is_cr(*raw_input_c) && runway >= 2 && ParseRules::is_lf(raw_input_c[1])) {
-        raw_input_c += 2;
-        zret = PARSE_OK;
+      if (ParseRules::is_cr(*raw_input_c)) {
+        ++raw_input_c;
+        if (runway >= 2 && ParseRules::is_lf(*raw_input_c)) {
+          // optimize a bit - this happens >99% of the time after a CR.
+          ++raw_input_c;
+          zret = PARSE_DONE;
+        } else {
+          S->m_state = MIME_PARSE_FOUND_CR;
+        }
+      } else if (ParseRules::is_lf(*raw_input_c)) {
+	++raw_input_c;
+        zret = PARSE_DONE; // Required by regression test.
+      } else {
+        // consume this character in the next state.
+        S->m_state = MIME_PARSE_INSIDE;
+      }
+      break;
+    case MIME_PARSE_FOUND_CR:
+      // Looking for a field and found a CR, which should mean terminating
+      // the header. Note that we've left the CR in the input so we have
+      // to skip over it.
+      if (ParseRules::is_lf(*raw_input_c)) {
+        // Header terminated.
+        ++raw_input_c;
+        zret = PARSE_DONE;
       } else {
+        // This really should be an error (spec doesn't permit lone CR)
+        // but the regression tests require it.
+        mime_scanner_append(S, &RAW_CR, 1);
         S->m_state = MIME_PARSE_INSIDE;
       }
       break;

http://git-wip-us.apache.org/repos/asf/trafficserver/blob/b55801c8/proxy/hdrs/MIME.h
----------------------------------------------------------------------
diff --git a/proxy/hdrs/MIME.h b/proxy/hdrs/MIME.h
index 5af9e09..49a6b98 100644
--- a/proxy/hdrs/MIME.h
+++ b/proxy/hdrs/MIME.h
@@ -55,6 +55,7 @@ enum
 /// Parsing state.
 enum MimeParseState {
   MIME_PARSE_BEFORE, ///< Before a field.
+  MIME_PARSE_FOUND_CR, ///< Before a field, found a CR.
   MIME_PARSE_INSIDE, ///< Inside a field.
   MIME_PARSE_AFTER,  ///< After a field.
 };