[mapserver-dev] Mapserver memory leaks

[Ian Walberg]

Alan,

Good point on the timing.

We seem to have 2 leaks we need to fix.

First is in FriBidi, rebuilding mapserver without FriBidi cured the leak
but I assume that the right to left text support has been lost. I do not
have the valgrind trace to hand but we will capture this tomorrow.

Second is :-

==15926== 45,440 bytes in 1,136 blocks are definitely lost in loss
record 8,107 of 8,126
==15926==    at 0x4005447: calloc (vg_replace_malloc.c:467)
==15926==    by 0x46B91B3: msSmallCalloc (in
/navlibs/lib/php_mapscript.so)
==15926==    by 0x46DF85C: msDrawRasterLayer (in
/navlibs/lib/php_mapscript.so)
==15926==    by 0x46E65DB: msDrawLayer (in
/navlibs/lib/php_mapscript.so)
==15926==    by 0x46E74A7: msDrawMap (in /navlibs/lib/php_mapscript.so)
==15926==    by 0x467EE72: mapObj_draw (in
/navlibs/lib/php_mapscript.so)
==15926==    by 0x4675621: zim_mapObj_draw (in
/navlibs/lib/php_mapscript.so)
==15926==    by 0x836BFED: zend_do_fcall_common_helper_SPEC
(/data/php-5.3.6/Zend/zend_vm_execute.h:316)
==15926==    by 0x836B525: execute
(/data/php-5.3.6/Zend/zend_vm_execute.h:107)
==15926==    by 0x834AAA8: zend_execute_scripts
(/data/php-5.3.6/Zend/zend.c:1194)
==15926==    by 0x82F997B: php_execute_script
(/data/php-5.3.6/main/main.c:2268)
==15926==    by 0x83CB1EA: main
(/data/php-5.3.6/sapi/cli/php_cli.c:1193)

We will rebuild with debug and symbols and grab the details tomorrow.

Regards

Ian


-----Original Message-----
From: mapserver-dev-bounces at lists.osgeo.org
[mailto:mapserver-dev-bounces at lists.osgeo.org] On Behalf Of Alan
Boudreault
Sent: Wednesday, February 01, 2012 1:40 PM
To: mapserver-dev at lists.osgeo.org
Subject: Re: [mapserver-dev] Mapserver memory leaks

Hi Ian,

It's a good time to fix them for the next MS 6.2 release. Please create
a ticket with test cases and the valgrind output. Do you know if the bug
is related to the data source used?

Alan

On 12-02-01 03:44 PM, Ian Walberg wrote:
> Folks,
>
> We have a mapscript (6.x) application which runs as a process so the 
> php script does not exit.
>
> What we appear to be seeing (using valgrind) is a surprisingly large 
> number of memory leaks.
>
> Our next step is to look into each one and try and determine the cause

> but they appear to be in multiple places.
>
> Does anyone have any experience of similar conditions?
>
> Many thanks
>
> Ian
> _______________________________________________
> mapserver-dev mailing list
> mapserver-dev at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/mapserver-dev


--
Alan Boudreault
http://www.mapgears.com/
_______________________________________________
mapserver-dev mailing list
mapserver-dev at lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/mapserver-dev