afflib fails to build on some architectures

Simson Garfinkel simsong at acm.org
Sun Dec 13 00:15:36 UTC 2009


Thanks for the email. 

I have had a persistent problem getting the tests to work because the binary and test files are in different locations when I am do a "make check" vs. a "make distcheck". The directory where the binaries are is set to be read-only in some cases but not in other cases. There is a file that is distributed which ends up I-dont-know-where.

If you can show me a list of all of the environment variables at my disposal for the test scripts and how they are supposed to be used, I can put out a bug fix within the next 24 hours.

Simson


On Dec 12, 2009, at 3:57 PM, Michael Prokop wrote:

> [Please CC me on replies as I'm not subscribed to the aff-discuss list, thanks.]
> 
> Hi,
> 
> I'm member of the Debian forensics team and wanted to inform you
> that we noticed a problem with building afflib on some
> architectures:
> 
>  https://buildd.debian.org/status/package.php?suite=unstable&p=afflib
> 
> As you can see the tests fail at/with:
> 
>  af_open(blank.aff): No such file or directory
>  ENCRYPTED SEGMENTS IN BLANKE.AFF --- STOP
>  FAIL: test_passphrase.sh
> 
> The bugs are reported in Debian's BTS at:
> 
>  http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=552049
>  http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=549832
> 
> As this is a showstopper for afflib being shipped with the upcoming
> stable release of Debian I'd appreciate any help so we can resolve
> this issue.
> 
> Do you have any ideas how we could fix this issue (without having to
> disable the tests of course :))?
> 
> BTW: We already have a patch which addresses a bashism issue in the
> tools/test_*.sh scripts. You're welcome to apply it in your upstream
> tree:
> 
>  http://patch-tracker.debian.org/patch/series/dl/afflib/3.5.2+dfsg-1/03-bashism.patch
> 
> PS: The mailaddress referenced in the source tree where bugreports
> should be sent to is broken (mails to bugs at afflib.org just bounce),
> Cc-ing Simson directly therefore.
> 
> Thanks.
> 
> regards,
> -mika-




More information about the forensics-devel mailing list