Re: Segfault when running fibc benchmark in current trunk

[Andy Wingo]

Hi Juhani!

On Wed 22 Apr 2009 21:29, "Juhani Rantanen" <address@hidden> writes:

>>> I look forward to the next bug report :)
>>
>> Here are some more files that don't need the benchmark suite to run. I
>> also ran all of the tests with Gambit and there wasn't any errors.
>>
>> Testing maze under Guile-r5rs
>> Compiling...
>> Running...
>> ERROR: In procedure zero?:
>> ERROR: Wrong type argument in position 1: #f
>
> Hi,
>
> I still have some cough and flu, I hope you're luckier and have not
> catched this evil disease..

Uf, sounds quite irritating. For the moment I am free of this (knock on
wood)

> it seems that this maze_bug.scm not working is actually a portability
> issue in the testsuite, not a genuine guile bug. At least this seems
> likely. Benchmark does run with Gambit without any errors, but that is
> a different implementation with different set of primitives. So I'm
> sorry if I've sent you to a wild goose chase.

I haven't started poking at this yet, so no problem, although not for
the right reason :)

> That aside, I am interested in using Guile to extend my application

Great!

> and I
> will volunteer to test it with some other code as well before the next
> release to improve quality of the release.

Even better! :-)

> One bug might be that Guile 1.8.6 runs every benchmark (some of them
> are real-world applications, some simple synthetic benchmarks) faster
> than the current version with this new VM.

Hah! Yes that sounds like a bug to me :) Are you compiling the
benchmarks? I'd like to look into this. Let me know if you can get an
isolated test case.

Cheers,

Andy
-- 
http://wingolog.org/