Smalltalk › Squeak › Squeak - Dev

VM crash with unhelpful debug log

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

6 messages Options

Karl Ramberg

VM crash with unhelpful debug log

I have a updated image that crash if I try to save from Monticello. Not sure why...

Saving otherwise from the image is fine

Crash log is very terse and does not show what goes wrong

Best,

Karl

crash.dmp (9K) Download Attachment

marcel.taeumel

Re: VM crash with unhelpful debug log

Hi Karl,

it seems to be a Windows machine. I have these crashes from time to time, but they are rare. Primitive 117 is an external call and my guess is that it is related to file access. In the case of a Monticello commit it could the the MCZ that is going to be stored in the local package cache.

If I had such crashes, a Windows reboot did solve it. Thus, I suspect from broken file handles or caches or whatever.

It is quite annoying because it is not reproducible.

Best,
Marcel

Levente Uzonyi

Re: VM crash with unhelpful debug log

It's very likely that the UUIDPlugin is responible for the crash[1].
It has been changed recently and is compiled as an external plugin on
Windows.
If it won't crash with 3643[2], then it's definitely the UUIDPlugin on
Windows. Someone with a Windows machine and a debugger should find the
cause and try to fix it.

If you want to work around the issue on your system, then you can either
rename/move the UUIDPlugin.dll or revert #makeUUID to its previous
version.

Levente

[1] http://forum.world.st/VM-Crash-LocaleTest-gt-gt-beginning-with-Cog-Spur-16-13-on-Windows-td4896004.html
[2] http://www.mirandabanda.org/files/Cog/VM/VM.r3643/

On Fri, 27 May 2016, marcel.taeumel wrote:

> Hi Karl,
>
> it seems to be a Windows machine. I have these crashes from time to time,
> but they are rare. Primitive 117 is an external call and my guess is that it
> is related to file access. In the case of a Monticello commit it could the
> the MCZ that is going to be stored in the local package cache.
>
> If I had such crashes, a Windows reboot did solve it. Thus, I suspect from
> broken file handles or caches or whatever.
>
> It is quite annoying because it is not reproducible.
>
> Best,
> Marcel
>
>
>
> --
> View this message in context: http://forum.world.st/VM-crash-with-unhelpful-debug-log-tp4897850p4897856.html
> Sent from the Squeak - Dev mailing list archive at Nabble.com.
>
>

Karl Ramberg

Re: VM crash with unhelpful debug log

Hi, thanks Levente.

I reverted makeUUID and it works again.

Best,

Karl

On Fri, May 27, 2016 at 5:20 PM, Levente Uzonyi <[hidden email]> wrote:

It's very likely that the UUIDPlugin is responible for the crash[1].
It has been changed recently and is compiled as an external plugin on Windows.
If it won't crash with 3643[2], then it's definitely the UUIDPlugin on Windows. Someone with a Windows machine and a debugger should find the cause and try to fix it.

If you want to work around the issue on your system, then you can either rename/move the UUIDPlugin.dll or revert #makeUUID to its previous version.

Levente

[1] http://forum.world.st/VM-Crash-LocaleTest-gt-gt-beginning-with-Cog-Spur-16-13-on-Windows-td4896004.html
[2] http://www.mirandabanda.org/files/Cog/VM/VM.r3643/

On Fri, 27 May 2016, marcel.taeumel wrote:

Hi Karl,

it seems to be a Windows machine. I have these crashes from time to time,
but they are rare. Primitive 117 is an external call and my guess is that it
is related to file access. In the case of a Monticello commit it could the
the MCZ that is going to be stored in the local package cache.

If I had such crashes, a Windows reboot did solve it. Thus, I suspect from
broken file handles or caches or whatever.

It is quite annoying because it is not reproducible.

Best,
Marcel

--
View this message in context: http://forum.world.st/VM-crash-with-unhelpful-debug-log-tp4897850p4897856.html
Sent from the Squeak - Dev mailing list archive at Nabble.com.

Tobias Pape

Re: VM crash with unhelpful debug log

In reply to this post by Karl Ramberg

Hi Karl,

On 27.05.2016, at 16:53, karl ramberg <[hidden email]> wrote:

> I have a updated image that crash if I try to save from Monticello. Not sure why...
>
> Saving otherwise from the image is fine
>
> Crash log is very terse and does not show what goes wrong

Is it reproducible? if so, could you comment out the primitive in primMakeUUID in UUID and
see whether the crash goes away?

Best regards
-Tobias

Tobias Pape

Re: VM crash with unhelpful debug log

On 27.05.2016, at 18:10, Tobias Pape <[hidden email]> wrote:

> Hi Karl,
>
> On 27.05.2016, at 16:53, karl ramberg <[hidden email]> wrote:
>
>> I have a updated image that crash if I try to save from Monticello. Not sure why...
>>
>> Saving otherwise from the image is fine
>>
>> Crash log is very terse and does not show what goes wrong
>
> Is it reproducible? if so, could you comment out the primitive in primMakeUUID in UUID and
> see whether the crash goes away?

Nevermind, my mail was slow…