[Reproducible-commits] [libxslt] 01/04: Fix generate-id() to not expose object addresses
Mattia Rizzolo
mattia at debian.org
Fri May 20 12:49:10 UTC 2016
This is an automated email from the git hooks/post-receive script.
mattia pushed a commit to annotated tag debian/1.1.26-6+squeeze1
in repository libxslt.
commit dbb14e5be43bf20c8b7a2e37bda1d8f7867dc56b
Author: Daniel Veillard <veillard at redhat.com>
Date: Tue Feb 22 10:14:23 2011 +0800
Fix generate-id() to not expose object addresses
As pointed out by Chris Evans <scarybeasts at gmail.com> it's better
security wise to not expose object addresses directly, use a diff
w.r.t. the document root own address to avoid this
* libxslt/functions.c: fix IDs generation code
---
libxslt/functions.c | 26 +++++++++++++++++++++-----
1 file changed, 21 insertions(+), 5 deletions(-)
diff --git a/libxslt/functions.c b/libxslt/functions.c
index 4720c7a..de962f4 100644
--- a/libxslt/functions.c
+++ b/libxslt/functions.c
@@ -654,8 +654,9 @@ xsltFormatNumberFunction(xmlXPathParserContextPtr ctxt, int nargs)
void
xsltGenerateIdFunction(xmlXPathParserContextPtr ctxt, int nargs){
xmlNodePtr cur = NULL;
- unsigned long val;
- xmlChar str[20];
+ long val;
+ xmlChar str[30];
+ xmlDocPtr doc;
if (nargs == 0) {
cur = ctxt->context->node;
@@ -694,9 +695,24 @@ xsltGenerateIdFunction(xmlXPathParserContextPtr ctxt, int nargs){
* Okay this is ugly but should work, use the NodePtr address
* to forge the ID
*/
- val = (unsigned long)((char *)cur - (char *)0);
- val /= sizeof(xmlNode);
- sprintf((char *)str, "id%ld", val);
+ if (cur->type != XML_NAMESPACE_DECL)
+ doc = cur->doc;
+ else {
+ xmlNsPtr ns = (xmlNsPtr) cur;
+
+ if (ns->context != NULL)
+ doc = ns->context;
+ else
+ doc = ctxt->context->doc;
+
+ }
+
+ val = (long)((char *)cur - (char *)doc);
+ if (val >= 0) {
+ sprintf((char *)str, "idp%ld", val);
+ } else {
+ sprintf((char *)str, "idm%ld", -val);
+ }
valuePush(ctxt, xmlXPathNewString(str));
}
--
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/reproducible/libxslt.git
More information about the Reproducible-commits
mailing list