[AWS] memory leak in AWS-based application

[Petr Holub]

> Which seems to be exactly the problem I was talking about (GNAT bug on
> aggregate with controlled objects). One option is to upgrade to a more
> recent AWS version (containing the workaround) or live with this memory
> leak until a new compiler or official AWS package is available.

Thanks a lot Pascal, this really helped. I've migrated to SVN version
of AWS and the leak is gone. Two comments associated with the
SVN version:

1) it also requires SVN version of XMLAda, otherwise the compilation
   fails with the following error (but that's sort of expectable):

gcc -c -gnat05 -gnatwcfijkmruv -gnaty3abcefhiklmnoprstx -Wall -O2 -gnatn -I- -gnatA
/home/hopet/gnat-2008/dist/AWS-SVN/soap/soap-message-reader.adb
soap-message-reader.adb:63:45: expected type "Standard.String"
soap-message-reader.adb:63:45: found type "Byte_Sequence_Access" defined at unicode-ces.ads:49
soap-message-reader.adb:161:12: "Set_Node_List_Growth_Factor" not declared in "Core"
gnatmake: "/home/hopet/gnat-2008/dist/AWS-SVN/soap/soap-message-reader.adb" compilation error
make[1]: *** [build] Error 4
make: *** [src_build] Error 2

2) there is a mess regarding installation of hotplug_password

$ make install
cp: cannot stat `../.build/release/static/tools/hotplug_password': No such file or directory
cp: cannot stat `aws.html': No such file or directory
cp: cannot stat `aws.txt': No such file or directory
cp: cannot stat `*.info*': No such file or directory
cp: cannot stat `templates_parser/docs/templates_parser.html': No such file or directory
cp: cannot stat `templates_parser/docs/templates_parser.info*': No such file or directory

readme.txt:  hotplug_password tool has been renamed to aws_password.

   Thus correcting tools/makefile solves the problem.


Btw, do you have a clue who's responsible for XMLAda? I've realized
that there are a few (sort of) errors in the docs - there are
missing Close () calls and Free () calls in the examples given
in the docs in order to creat leak-free code... (or is it again
the problem with GNAT bug?)

Last but not least, I'd like to support Ludovic regarding his
question about the GNAT bug - is it already corrected in the
development version of the GNAT? And even if yes, it may be worth
noting it somewhere since the recent stable version of GNAT still
contains it... Btw, is there any estimate for the new GNAT release
that would sort all of this out?

Thanks a lot once again,
Petr
 
================================================================
                           Petr Holub
CESNET z.s.p.o.                       Supercomputing Center Brno
Zikova 4                             Institute of Compt. Science
162 00 Praha 6, CZ                            Masaryk University
Czech Republic                     Botanicka 68a, 60200 Brno, CZ 
e-mail: Petr.Holub at cesnet.cz               phone: +420-549493944
                                             fax: +420-541212747
                                       e-mail: hopet at ics.muni.cz