[gvd-users] Oppinion on segmentation fault

[Alejandro R. Mosteo Chagoyen]

Hello,

running WinXP SP1 here, Gnat 3.15p, GVD 1.2.6

I have a fairly sized program (~5.5 MB executable) which seems to run
fine. However, when running it from GVD, I get this error after some
time of running (about half an hour - an hour):

Program received signal SIGSEGV, Segmentation fault.
[Switching to thread 3500.0x13ac]
0x72749c in <adagio__server__listPT__deleteN> (<_object>=@0x991608,
this=0xfeeefeee) at ../network/adagio-server.adb:62

My program seems to be OK except that I'm trying to determine a
strange deadlock or hanging which happens after a couple of hours of
running (experiences range from 6 to 8 hours until happening, in a
given PC. In others I have had it running for several days without
problem).

When I run it stand-alone I never get exceptions outside of expected
ones for normal operation.

The above referenced line corresponds to a call to a procedure of a
container instantiated from the Charles library. The call is made from
inside a protected object.

However, I don't know what is the meaning of the task switch note.

Any clue to interpret the "_object", this", etc. hexadecimal data?

I have had programs in the past that refused to run from GVD, failing
during elaboration with sigsevs, and they were pretty trivial ones, so
I'm not sure if GVD may have problems in some circumstances/platforms,
or really the above error could be related with my problem. Do you
think it could denote a memory corruption which slips unnoticed when
running stand-alone?

Thanks for your time,

A. Mosteo.