Re: [tcljava-dev] bug 2866640, tclblend - potential memory leak

[Mo D. <mo...@mo...>]

Hi Anthony

Wow, that was a tricky bug to figure out the proper fix for. Please try 
out Patch 2982514, it should address the leak you were seeing, but 
verification of the fix is needed. I will append it to this email since 
it is very small.

Mo DeJong


RCS file: /cvsroot/tcljava/tcljava/src/native/javaObj.c,v
retrieving revision 1.17
diff -u -r1.17 javaObj.c
--- src/native/javaObj.c        31 Dec 2002 20:16:27 -0000      1.17
+++ src/native/javaObj.c        6 Apr 2010 08:00:36 -0000
@@ -233,6 +233,13 @@
     jobject exception;

     /*
+     * If object was freed already, or if the ptr2 field was
+     * explicitly set to NULL, then do nothing.
+     */
+    if (object == NULL)
+         return;
+
+    /*
      * Clear pending Java exception.
      */

@@ -969,8 +976,8 @@
        if (objPtr->bytes == NULL) {
            UpdateTclObject(objPtr);
        }
-       objPtr->typePtr = NULL;
        ptr2 = objPtr->internalRep.twoPtrValue.ptr2;
+       objPtr->internalRep.twoPtrValue.ptr2 = NULL;
        result = (oldCmdType.setFromAnyProc)(interp, objPtr);
        objPtr->internalRep.twoPtrValue.ptr2 = ptr2;
     } else {


Anthony Borthwick wrote:
> Hi,
>
> I added some test code to the bug report which demonstrates this issue.
> Is there any news?
>
> Tony
>
> -----Original Message-----
> From: Mo DeJong [mailto:mo...@mo...] 
> Sent: 04 February 2010 19:33
> To: discuss implementaition and patches here
> Subject: Re: [tcljava-dev] bug 2866640, tclblend - potential memory leak
>
> Anthony Borthwick wrote:
>   
>> Hi,
>>
>>  
>>
>> Just wondering if anything is being done with this issue
>>
>>  
>>
>>
>>     
> http://sourceforge.net/tracker/?func=detail&aid=2866640&group_id=13005&a
> tid=113005 
>   
> <http://sourceforge.net/tracker/?func=detail&aid=2866640&group_id=13005&
> atid=113005>
>   
>>  
>>
>> Tony
>>
>>  
>>     
> Memory leak? Well, I am willing to take a look at it today. But, could 
> you provide a small script example that demonstrates how this bug can be
>
> reproduced? Often, it is very difficult to figure out how to reproduce 
> the problem without an example of how to generate the bug. You could 
> post it here or append it to the bug report.
>
> Mo
>