From 93b8289b60234bd70a601f39a794a9ca404b4f44 Mon Sep 17 00:00:00 2001
From: David Anderson <davea@ssl.berkeley.edu>
Date: Fri, 29 Mar 2013 22:36:53 -0700
Subject: [PATCH] - XML parser: return error if string exceeds buffer size. -
 client: when parsing MD5, use 64 instead of 33 char buffer.     When the XML
 parser reads a string,     it enforces the buffer size limit BEFORE it strips
 whitespace.     So if a project put whitespaces before or after the MD5,    
 it would fail to parse.

---
 client/client_types.h | 2 +-
 lib/md5_file.h        | 4 ++++
 lib/parse.cpp         | 1 +
 lib/parse.h           | 4 +++-
 lib/parse_test.cpp    | 5 +++--
 5 files changed, 12 insertions(+), 4 deletions(-)

diff --git a/client/client_types.h b/client/client_types.h
index 5396c1821d..620f7aa60b 100644
--- a/client/client_types.h
+++ b/client/client_types.h
@@ -95,7 +95,7 @@ struct URL_LIST {
 
 struct FILE_INFO {
     char name[256];
-    char md5_cksum[33];
+    char md5_cksum[MD5_LEN];
     double max_nbytes;
     double nbytes;
     double gzipped_nbytes;  // defined if download_gzipped is true
diff --git a/lib/md5_file.h b/lib/md5_file.h
index a21b8a2c20..c9ba17be7f 100644
--- a/lib/md5_file.h
+++ b/lib/md5_file.h
@@ -21,6 +21,10 @@
 #include <string>
 
 // length of buffer to hold an MD5 hash
+// In principle need 32 + 1 for NULL,
+// but leave some room for XML whitespace
+// (since we parse before stripping whitespace)
+//
 #define MD5_LEN 64
 
 extern int md5_file(const char* path, char* output, double& nbytes);
diff --git a/lib/parse.cpp b/lib/parse.cpp
index e2b7b4b335..53d5002cf1 100644
--- a/lib/parse.cpp
+++ b/lib/parse.cpp
@@ -566,6 +566,7 @@ bool XML_PARSER::parse_str(const char* start_tag, char* buf, int len) {
     //
     int retval = get_aux(buf, len, 0, 0);
     if (retval == XML_PARSE_EOF) return false;
+    if (retval == XML_PARSE_OVERFLOW) return false;
 
     // if it's the end tag, return empty string
     //
diff --git a/lib/parse.h b/lib/parse.h
index ac1316cde0..00cd49cc20 100644
--- a/lib/parse.h
+++ b/lib/parse.h
@@ -119,7 +119,9 @@ struct XML_PARSER {
     }
 
     // Scan something, either tag or text.
-    // Strip whitespace at start and end.
+    // Strip whitespace at start and end
+    // (however, the supplied buffer must accommodate this white space).
+    // Ignore comments.
     // Return true iff reached EOF
     //
     inline int get_aux(
diff --git a/lib/parse_test.cpp b/lib/parse_test.cpp
index 925f113984..2f9d85bd34 100644
--- a/lib/parse_test.cpp
+++ b/lib/parse_test.cpp
@@ -13,6 +13,7 @@ void parse(FILE* f) {
     int val;
     double x;
 
+    strcpy(name, "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx");
     mf.init_file(f);
     if (!xp.parse_start("blah")) {
         printf("missing start tag\n");
@@ -28,8 +29,8 @@ void parse(FILE* f) {
         if (xp.match_tag("/blah")) {
             printf("success\n");
             return;
-        } else if (xp.parse_str("str", name, sizeof(name))) {
-            printf("got str: %s\n", name);
+        } else if (xp.parse_str("str", name, 33)) {
+            printf("got str: [%s]\n", name);
         } else if (xp.parse_int("int", val)) {
             printf("got int: %d\n", val);
         } else if (xp.parse_double("double", x)) {