Re: ksw2 library - nobody else at it, right?
Hi Steffen,
On Fri, May 15, 2020 at 02:39:51AM +0200, Steffen Möller wrote:
> I am happy that you asked - it is here
> https://salsa.debian.org/med-team/ksw2
>
> Sadly not a problem of that library which features the same files that
> are also distributed with minimap2, just as an older version in the
> separate ksw2 repository if I am not erroneous.
Yes, file ksw2_exts2_sse.c seems to be older than in the code copy in
minimap2.
> We need to find the
> problem within the minimap2 package, I am afraid. Was looking for a
> while now - thought multiple times that I had it - but ... sigh.
>
> That ksw2 library needs some love by upstream and I suggest to avoid
> packaging it for now. Also, it does not ship with a license and comes
> with the sources of a journal publication - er - have a look.
Besides that I agree that it should be sorted out with upstream I
think the proper action would be:
1. Sort out the code update with upstream
2. Ask upstream for a proper release
3. Fix the build system I've invented to get a proper library
Regarding the last item: I have added extremely rough automake stuff to
the packaging. The idea was that I want a proper library with static and
dynamic library. However, my automake skills are limited. Here are the
flaws:
1. Proper check for zlib is needed (see the dirty hack in d/rules)
2. There is no propagation of HAVE_KALLOC definition (no idea why)
3. Check and use parasail
4. May be package gaba (whatever this might be and check and link
against this)
Finally I took over the simde patch from minimap2. Unfortunately I have
no idea how to add this to the linker. This has also be done inside
my weak attempt of automake. (If you comment the simde patch the package
at least builds.)
Kind regards
Andreas.
--
http://fam-tille.de
Reply to: