| Re: IDL 8.1 Crashes on Windows [message #77856 is a reply to message #77855] |
Sun, 02 October 2011 01:49  |
lecacheux.alain
Messages: 325
Registered: January 2008
|
Senior Member |
|
|
On 30 sep, 21:13, David Fanning <n...@dfanning.com> wrote:
> Folks,
>
> I have received an unconfirmed report that one possible
> source of IDL 8.1 crashes on Windows computers may be
> an inadvertent leak of GDI objects. These Graphics
> Device Interface objects are responsible for much
> of the menu and graphics drawing on the computer.
> And it is apparently quite easy to make an inadvertent
> mistake when cleaning these objects up.
>
> http://msdn.microsoft.com/en-us/magazine/cc301756.aspx
>
> Windows computers apparently only allow 10,000 of these
> objects. If you have more, the machine crashes in a very
> hard way, so that the only recourse is to hit the panic
> button. You can keep track of how many of these GDI objects
> are in use at any particular time with the Task Manager.
> (Go to the View->Select Columns menu item and select the
> GDI object selection.)
>
> I am told there is patch file available that apparently
> solves the problem.
>
> Cheers,
>
> David
> --
> David Fanning, Ph.D.
> Fanning Software Consulting, Inc.
> Coyote's Guide to IDL Programming:http://www.idlcoyote.com/
> Sepore ma de ni thui. ("Perhaps thou speakest truth.")
Bravo David !
I could check that the sudden IDL crashes I described in previous
threads (with complete mess in desktop window management) seems indeed
to be due to "the number of GDI objects exceeding the 10000 limit".
This behavior is well known, for example, as occuring when using some
old, bugged versions of Outlook. I should have done the relation by
myself !
Yesterday evening, just after loading IDL 8.1 (on one of my XP 64
machines), this number was about 250 (comparable to that for IDL 6.4
and its native IDE). During the subsequent interactive work, the
number was growing regularly and never decreases. Looking at the
magnitude of increasing steps, I feel that graphical work (opening and
closing graphic windows, plotting curves, drawing images, etc...) is
not the main culprit: editing (inside Eclipse), and especially
debugging, was clearly creating a large number of "GDI objects" which
never returned to the pool. When the number was reaching 9999, (that
is about two hours after having started IDL), the crash did occur...
While impatiently waiting for a fix by ITTVIS, I would appreciate to
get the patch you told about.
Cheers,
alain.
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
