Re: igblast accepted - how can it be used to test igdiscover
Hi Aaron,
Am Tue, Aug 02, 2022 at 05:44:05PM -0400 schrieb Aaron M. Ucko:
> > Am 02.08.2022 um 10:48 schrieb Andreas Tille:
> [...]
> >> Do you have any other hint how to create the main binaries reliably?
>
> Ah, sorry, a closer look indicates that the build system needs
> additional hints to find system VDB:
>
> --with-vdb=/usr VDB_INCLUDE=-I/usr/include/ncbi-vdb
Hmmm, this was running into an error which I tried to fix with
--with-vdb VDB_LIBS=-lncbi-vdb VDB_INCLUDE=-I/usr/include/ncbi-vdb
but this results in
checking for libsqlite3... no
configure: trying to build the NCBI SRA/VDB Toolkit from GitHub.
configure: error: in `/build/ncbi-igblast-1.19.0/c++':
configure: error: SRA/VDB build failed with status 127
See `config.log' for more details
make[1]: *** [debian/rules:7: override_dh_auto_configure] Error 1
make[1]: Leaving directory '/build/ncbi-igblast-1.19.0'
make: *** [debian/rules:4: binary] Error 2
dpkg-buildpackage: error: debian/rules binary subprocess returned exit status 2
I: copying local configuration
It would be great if someone could have a deeper look.
> > I wanted to see it compile and thus withouted as much as possible in a
> > first run to reduce the complexity.
>
> That's fair. FWIW, most of these options have no bearing on the package
> anyway and are just artifacts of IgBLAST's origin in a larger tree.
> Also, there's one option I'd suggest actively enabling: --with-openmp.
> If you specify neither --with-openmp nor --without-openmp, the build
> system uses OpenMP on a limited basis, but explicit --with-openmp
> enables it across the board, from which (Ig)BLAST generally benefits.
I've cleared most of the --without options and added --with-openmp with
the according Build-Depends.
Unfortunately I can't test this since the configure step fails.
Kind regards
Andreas.
--
http://fam-tille.de
Reply to: