On 5/21/07, Mattias Gaertner <[EMAIL PROTECTED]> wrote:
I'm not sure where the mem leak is.
Formerly the gtk interface did not check and simply freed the object,
risking strange errors. OTOH the win32/64 intf only returns false and
afaik no code checked the result, creating mem leaks and even worse:
non freed handles. This is common behavior for Delphi code and even
example code often forget this.
Now the gtk intf checks, does not free (winapi compatible) and gives
warnings.
Probably somewhere a brush was selected into a device context
(sgOutcomes:TStringGrid) and was not restored.
We don't do any custom drawing in the TStringGrid mentioned, so does
that point to the LCL then? We tested the same application under
Windows - it doesn't crash or freeze the application as it does under
Linux/GTK1. But I guess that could be due to the fact you mentioned -
Windows only returns false and doesn't check the result.
The screens using the TStringGrid's in our applications haven't been
changed in a while but suddenly causes the crashes. I guess this is
brought on by the fact that GTK1 is now more strict with it's resource
checks. That would explain the crashes under Linux and not under
Windows.
--
Graeme Geldenhuys
General error, hit any user to continue.
_________________________________________________________________
To unsubscribe: mail [EMAIL PROTECTED] with
"unsubscribe" as the Subject
archives at http://www.lazarus.freepascal.org/mailarchives
