[PolyORB-users] Memory leaks in polyorb

[Xavier Grave]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Tony,

Le 17/08/2012 18:36, Tony Gair a écrit :
> 
> 
> I'm getting memory leaks using a remote call interface. I am also
> passing variant records across the interface as a result of a function.
> Using Valgrind they look like

As far as I can read from the log and in my humble opinion, 480 bytes
can't be considered like an memory leak. In fact, at initialisation
during elaboration packages can allocate memory once for all the life
time of the process relying on the system for freeing it when the
process end.

Can we have access to the code that leaks from your point of view ?

xavier

> ==2733== 480 bytes in 5 blocks are possibly lost in loss record 2,577 of
> 2,646
> ==2733==    at 0x4C244E8: malloc (vg_replace_malloc.c:236)
> ==2733==    by 0xC45CED7: __gnat_malloc (in /usr/lib/libgnat-4.4.so.1)
> ==2733==    by 0xBCBFD9D: polyorb__any__initialize (in
> /usr/lib/libpolyorb.so.2)
> ==2733==    by 0xBCA629D: polyorb__any__to_any_instances__to_any__17Xn
> (in /usr/lib/libpolyorb.so.2)
> ==2733==    by 0xBCC0EDB: polyorb__any__to_any__15 (in
> /usr/lib/libpolyorb.so.2)
> ==2733==    by 0x4AD69A:
> main_store_pkg__building_controller_check_in___typecode12173___typecode12217___typecode12219.42841
> (in /home/target/dsaworking/bc_partition)
> ==2733==    by 0x4AD287:
> main_store_pkg__building_controller_check_in___typecode12173___typecode12217.41996
> (in /home/target/dsaworking/bc_partition)
> ==2733==    by 0x4A645A:
> main_store_pkg__building_controller_check_in___typecode12173.41923 (in
> /home/target/dsaworking/bc_partition)
> ==2733==    by 0x4A2E6C: main_store_pkg__building_controller_check_in
> (in /home/target/dsaworking/bc_partition)
> ==2733==    by 0x4E6DA3: dsa_comms_pkg__deal_with_remote_orders
> (dsa_comms_pkg.adb:22)
> ==2733==    by 0x4E80AB: dsa_comms_pkg__dsa_comms_task_typeB
> (dsa_comms_pkg.adb:234)
> ==2733==    by 0xBFC8802: ??? (in /usr/lib/libgnarl-4.4.so.1)
> 
> I am executing the calls from inside a task. Do I have to stop using
> variant records across the interface?
> 
> other information :
> I am using debian squeeze and the debian packages
> Gnat is version 4.4.5
> and polyorb version 2.6.0
> 
> _______________________________________________
> PolyORB-users mailing list
> PolyORB-users at lists.adacore.com
> /no-more-mailman.html

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
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=brLN
-----END PGP SIGNATURE-----