Re: [Boost-testing] bjam.exe crash, msvc-10, tag=master

Subject: Re: [Boost-testing] bjam.exe crash, msvc-10, tag=master
From: Jim Bell (Jim_at_[hidden])
Date: 2014-02-20 12:58:26

On 2014-02-19 5:43 PM, Steven Watanabe wrote:
> AMDG
>
> On 02/19/2014 03:21 PM, Jim Bell wrote:
>> bjam.exe crashed during a tag=master regression run today. The
>> regression scripts carried on.  Lots of detail below, and if anyone
>> wants more, let me know off-list.
>>
>> I built this bjam.exe using msvc-10, on 1/23/2014, from the git repo.
>>
>> <snip>
>>
>> Here's the report from the Windows Application Log:
>>
>> General:
>> Faulting application name: bjam.exe, version: 0.0.0.0, time stamp:
>> 0x52e154a5
>> Faulting module name: kernel32.dll, version: 6.1.7601.18229, time stamp:
>> 0x51fb1115
>> Exception code: 0xc00000fd
> This appears to mean stack overflow.
> I'd need a stack trace to debug it.
> I'm guessing that this is not reproducible,
> however.
>

It just ran the msvc-8 test to completion, without crashing. And the WER 
data files referenced in the report are gone. (And I wouldn't know what 
to do with them if I had them.)  Note, too, the tell-tale gap in the 
regression matrix for msvc-10.

I think you're right, that it's not reproducible. But maybe I'll attach 
a debugger on the next run anyway.

This is under Windows7x64 ("Ultimate"), by the way.

I didn't see any signs that the system was under undue stress, either.