[Openmcl-devel] invitation to stress test anticipatory symbol completion for CCL

[Gary Byers]

On Mon, 5 Jan 2009, Alexander Repenning wrote:

> Ralph,
>
> confirmed. Thanks for the feedback. Must be some kind of initialization 
> issue. I wonder what is different about the regular Hemlock buffer and the 
> Listener that could cause such an issue.
>
> The no crash report is still a bit of a mystery to me.

<http://clozure.com/pipermail/openmcl-devel/2008-November/008664.html>
was an attempt to explain some of this.


Has anybody figured 
> out to enable crash reports with CCL? CCL in general seems to be behave 
> pretty well but the Objective-C memory management can result in crashes that 
> are hard to track. Will have to do some digging.
>
> thanks again,  Alex
>
>
> On Jan 3, 2009, at 12:44 PM, Raffael Cavallaro wrote:
>
>> Alex,
>> 
>> I really like this facility, thanks very much for sharing it with the
>> CCL community. Not having to constantly type meta-slash is nice.
>> 
>> Unfortunately, under CCL 1.2-r11572M if I compile-load this and then
>> open a Listener, as soon as I type anything that triggers the symbol-
>> completion process CCL crashes. This is repeatable. Console message:
>> 
>> Jan  3 14:07:34 rafbookpro com.apple.launchd[159]
>> ([0x0-0x92c92c].com.clozure.Clozure CL[57877]): Exited with exit code:
>> 255
>> 
>> No .crash file in ~/Library/Logs/CrashReporter/
>> 
>> The workaround I've found is to always open some other hemlock buffer
>> after loading anticipatory-symbol-completion before attempting to open
>> a new listener which will use this facility.
>
> Prof. Alexander Repenning
>
> University of Colorado
> Computer Science Department
> Boulder, CO 80309-430
>
> vCard: http://www.cs.colorado.edu/~ralex/AlexanderRepenning.vcf
>
>