316619 - txMozillaXMLOutput::txMozillaXMLOutput(txOutputFormat* aFormat, nsIDOMDocumentFragment* aFragment)doesn't init mTreeDepth

Status: RESOLVED FIXED

(Core :: XSLT, defect)

Description

[Matt]

User-Agent:       Mozilla/5.0 (compatible; Konqueror/3.4; Linux 2.6.13-gentoo-r4; X11; x86_64) KHTML/3.4.3 (like Gecko)
Build Identifier: 

mTreeDepth in not initalized in the following ctor named txMozillaXMLOutput::txMozillaXMLOutput(txOutputFormat* aFormat, nsIDOMDocumentFragment* aFragment), but it is initialized in the other ctor. As a result, valgrind correctly reports uninitialized memory reads throughout the class.

To fix, add the mTreeDepth(0) line to the above ctor.

Reproducible: Always

Comment 1

[Matt]

patch:

Index: source/xslt/txMozillaXMLOutput.cpp
===================================================================
RCS file: /cvsroot/mozilla/extensions/transformiix/source/xslt/Attic/txMozillaXMLOutput.cpp,v
retrieving revision 1.61.8.1
diff -u -p -w -b -r1.61.8.1 txMozillaXMLOutput.cpp
--- source/xslt/txMozillaXMLOutput.cpp  14 Sep 2005 08:18:39 -0000      1.61.8.1
+++ source/xslt/txMozillaXMLOutput.cpp  16 Nov 2005 00:17:12 -0000
@@ -115,7 +115,8 @@ txMozillaXMLOutput::txMozillaXMLOutput(c

 txMozillaXMLOutput::txMozillaXMLOutput(txOutputFormat* aFormat,
                                        nsIDOMDocumentFragment* aFragment)
-    : mBadChildLevel(0),
+    : mTreeDepth(0),
+      mBadChildLevel(0),
       mTableState(NORMAL),
       mDontAddCurrent(PR_FALSE),
       mHaveTitleElement(PR_FALSE),

Comment 2

[Andrew Schultz]

Matt: can you attach your patch (Create a New Attachment) to the bug and request review (peterv, the assignee, would be a good reviewer)

(cvs diff is showing the file in Attic because the file moved to mozilla/content/xslt/src/xslt/txMozillaXMLOutput.cpp on the trunk)

Comment 3

[:Gavin Sharp [email: [e-mail]]]

Fixed in bug 226425.

*** This bug has been marked as a duplicate of 226425 ***

Comment 4

[Peter Van der Beken [:peterv]]

No, that bug's patch caused this. I'll check this in.

Comment 5

[:Gavin Sharp [email: [e-mail]]]

Huh? Hunk #1 of attachment 194860 [details] [diff] [review] is exactly the same as the patch in comment 1.

Comment 6

[:Gavin Sharp [email: [e-mail]]]

Ah, wrong decl, oops. Ignore me :).

Comment 7

[Peter Van der Beken [:peterv]]

Attachment: v1

Patch that I checked in.