[racket] crash possibly related to sqlite

From: Dmitry Pavlov (dpavlov at ipa.nw.ru)
Date: Mon Jul 14 05:08:38 EDT 2014

[After some private communication by email]

Matthew, thank you very much for addressing this issue in Racket's VM!

I just finished checking the nightly build from 20140712---it does not
crash anymore.

Best regards,

Dmitry

On Wed, Jul 9, 2014 at 2:43 PM, Matthew Flatt <mflatt at cs.utah.edu> wrote:
> Sometimes you can stress the VM by throwing in calls to
> `collect-garbage`, possibly in anything that acts like a callback, or
> possibly in a thread that forces a collection every few seconds.
>
> But, crashes after a lot of numerical work with SQLite work much
> earlier... I'm not so sure. I guess you can't get it to crash by
> interrupting the numerical part early?
>
> I would try running myself even if it takes hours. Sometimes, I can get
> information out of a crash that lets me create a different program that
> crashes quickly.
>
> At Wed, 9 Jul 2014 14:32:13 +0400, Dmitry Pavlov wrote:
>> > Would it be possible to send me a program and instructions so that I
>> > can replicate the crash?
>>
>> See, the problem is that the crash happens after hours of numerical
>> integration (although it happens every time, but that is where the good
>> news end).
>>
>> I would love to make a "minimal" crashing version of the program
>> for you to debug, but with every crash-or-not-crash check taking
>> a few hours, that would take quite a long time.
>>
>> Are there ways to "stress" the Racket VM so it crashes as
>> soon as it possibly can? :) Because, really, the numerical
>> integration routines are just repeating themselves all the way
>> with different numbers, and all the SQLiIte stuff takes place
>> in the very beginning of the program.
>>
>>
>> Best regards,
>>
>> Dmitry
>

Posted on the users mailing list.