Hi,
I have a fairly simple workspace that I am able to make crash while monitoring with procdump just by unbypassing a plugin.
The monitoring command was:
>procdump -mp -n 10 -e 1 Gig Performer
and the terminal output was:
[09:51:47] Exception: C0000094.INT_DIVIDE_BY_ZERO
[09:51:47] Dump 1 initiated: C:\Users\Steve\AppData\Local\CrashDumps\Performer.dmp
[09:52:16] Dump 1 complete: 3062 MB written in 28.2 seconds
[09:52:16] Exception: C0000094.INT_DIVIDE_BY_ZERO
[09:52:16] Dump 2 initiated: C:\Users\Steve\AppData\Local\CrashDumps\Performer-1.dmp
Here is a pic of the dump files generated:
A .zip including the large file was not accepted for upload here.
Here is a .zip with just the smaller files:
GigPerformer5.exe.19768 (2).zip (16.0 MB)
Here is the gig file:
MG3 Unify x3 Test 01.gig (910.2 KB)
What Plugin did you unbypass?
In this case, as shown, Unify 2.
Yes, just did it again, again with Unify 2.
Ok, what Happens when you remove the Unify 1 Plugins and unbypass unify 2?
With only Unify 2, unbypassing again created the dump and INT_DIVIDE_BY_ZERO error.
Ok, does the crash report indicate unify 2?
How do you unbypass unify 2?
Manually or via a widget?
Slightly different test, using only U-he Diva - same result.
You’re asking me?
Manually in the wiring diagram.
Does the crash report show that unify 2 is crashing or does Gig Performer Crash?
I don’t know what you mean by a crash report in this case, sorry.
What I get here are the .dmp files, which I’m hoping someone has more insight into than I do, as I have very little.
Yes, GP crashes.
Sure Gp crashes, but only because maybe the plugin crashed.
I am sure a Windows expert can analyze the crash dump
I’m hoping this is readable, or reproducible for you.
Or, alternatively, if the entire approach of monitoring with procdump has no revealing value, someone in the know about that could say so and explain.
IOW, are these particular plugins badly behaved wrt. unbypassing, or is GP itself possibly responsible for the INT_DIVIDE_BY_ZERO error, or is the entire observation/report an artifact of using procdump in this way?
Without procdump in the picture, I have no prior experience of these or other plugins crashing when unbypassed.
Maybe the INT_DIVIDE_BY_ZERO errors are coming from the plugin(s), but normally handled, successfully and silently by GP.
If this is the case, still the question is whether the plugins need to be fixed by their vendors so that we can use procdump to monitor GP without spurious errors. (IOW, to get dumps on real, traceable errors)
I merely speculate …
