Re: Status bowtie + tophat (Was: [Help] Need help for architecture specific code)
Hi,
On Fri, Aug 29, 2014 at 10:56:04AM +0200, Gianluca Della Vedova wrote:
> IMHO it makes sense to restrict all resource intensive packages to
> amd64. I am not aware of any installation on a different architecture.
Thanks for your input which is regarded helpful. Besides the general
arguments about code quality enhancements for portable software in my
opinion we should not be ignorant about current movements. We currently
see the advent of arm64 in Debian which might become quite interesting
once the first arm64 clusters might show up. Moreover even on our last
sprint a project was demonstrated which tries to port biology software
on Raspberry Pi[1]. From my point of view this means that we should
put some effort into portability.
Since I found a pretty simple solution which looks quite promising to
build bowtie[2] and bowtie2[2] on any architecture I think we should not
simply drop this support. I actually consider the upstream code
suspicious since it us using a copy of a gcc internal headerfile from
an old gcc version. I can't really believe that this is sustainable
code but I'm waiting for comments from upstream.
Alexandre, since you have maintained bowtie2 actively: Could you please
check this and perhaps run the autopkgtest code also in dh_auto_test
to make sure if we drop the architecture restriction we will realise
any non-working code at build time?
Kind regards
Andreas.
[1] http://eggg.st-andrews.ac.uk/4273pi/
[2] http://anonscm.debian.org/viewvc/debian-med/trunk/packages/bowtie/trunk/debian/patches/do_not_use_outdated_copy_of_cpuid_h.patch?view=markup
[3] http://anonscm.debian.org/cgit/debian-med/bowtie2.git/tree/debian/patches/do_not_use_outdated_copy_of_cpuid_h.patch
--
http://fam-tille.de
Reply to: