"Steinar H. Gunderson" <[EMAIL PROTECTED]> writes:
> On Tue, Oct 31, 2006 at 01:04:40PM +0100, Goswin von Brederlow wrote:
>> when building ctsim I get lots of errors caused by
>>
>> mpiworld.cpp:28:19: error: mpi++.h: No such file or directory
>>
>> untill the build stops with
>>
>> /usr/bin/hcp: No such file or directory
>> make[3]: *** [ctsimtext-lam] Error 1
>>
>> Do you build with make -k?
>
> The package builds fine in a clean chroot. diffing the build logs with the
> ones of the buildds indicate that it finds something related to MPI on your
> machine, enables MPI support in configure, and then fails. (It also finds
> libICE, but I don't think that should be fatal.)
>
> configuring with --without-lam or build-conflicting with whatever package
> causes the detection to work (strangely, it seems to be checking for
> /usr/bin/hcp?) should do the trick, but I'm not sure if this bug is RC at
> all, given that it works on all the autobuilders.
>
> /* Steinar */
> --
> Homepage: http://www.sesse.net/
You never know when a buildd chroot will not be clean. That happens
from time to time. Also users rebuilding your package or people doing
an NMU (or you uploading yourself) are likely to not use a clean
chroot. Sad but true.
Since you already detected the problem I suggest just fixing it. No
point arguing about severities in such a trivial fix. A --without-lam
is better than a Build-Conflicts though.
MfG
Goswin
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
