Re: RFS: failmalloc - Memory allocation failure crash-test tool (2nd try)
Alessandro Ghedini <al3xbio@gmail.com> writes:
>> > Is anyone interested in uploading this package?
>> >
>> I'm interested; it's a potentially very useful piece of software. I did
>> a quick look at your package, and found the following issues:
>>
>> - You seem to have repacked the upstream source:
>>
>> % cmp failmalloc_1.0.orig.tar.gz failmalloc-1.0.tar.gz
>> failmalloc_1.0.orig.tar.gz failmalloc-1.0.tar.gz differ: char 5, line 1
>>
>> This is for both the .tar.gz as obtained from pristine-tar as well as
>> the one from dget (they are identical, but different from upstream).
>> As there's no debian/README.source, I assume this is accidential.
>
> Yes, it was. I fixed it (now cmp is silent).
>
>> - debian/copyright is in an old DEP-5 format; the current one ([0],
>> which is now approaching stability, IIUC) can be converted to and
>> checked with the tools from the libconfig-model-perl package as
>> described in [1]. For example:
>
> Ok, I updated to the latest version of the format. Now also config-edit
> is silent.
>
>> I've not actually tested failmalloc yet, but assuming it's working, and
>> you fix the above issues (the first one in particuliar), I'll sponsor
>> it.
>
> Thank you very much.
>
Uploaded -- you're welcome!
One thing I noticed is that the "failmalloc" script hangs if you don't
supply the "--" separator; this should be fixed. Also, most people
expect "--" to be optional, so making the script behave that way would
be another improvement (i.e. stop processing failmalloc options at the
first argument not starting with a dash).
Regards, Rotty
--
Andreas Rottmann -- <http://rotty.yi.org/>
Reply to: