Re: Bug#692000: ITP: liblbfgs -- L-BFGS solver for dense nonlinear optimization problems
> On Sun, 4 Nov 2012 23:03:29 +0100
> Andreas Tille <andreas@fam-tille.de> wrote:
>
> Hi,
>
> Debian Med was a simple typo - I intended to write Debian Science.
>
> Kind regards
>
> Andreas.
>
>
> On Sat, Nov 03, 2012 at 04:23:49PM -0700, Dima Kogan wrote:
> > > On Sat, 3 Nov 2012 18:38:05 +0100
> > > Andreas Tille <andreas@an3as.eu> wrote:
> > >
> > > Hi Dima,
> > >
> > > you have submitted a series of ITPs which are all quite interesting for
> > > Debian Science. Are you considering to maintain these package in the
> > > Debian Med team (and its repositories on Alioth)?
> > >
> > > Kind regards
> > >
> > > Andreas.
> >
> > Hi Andreas.
> >
> > The packages in question are general-purpose solvers; i.e. there's nothing
> > inherently medical about them. For the purposes of maintainership, it doesn't
> > really matter, however. If you're offering to sponsor those packages (I think
> > we're talking about liblbfgs and libdogleg), debian-med is certainly fine with
> > me.
> >
> > Thanks
> >
> > dima
OK. Thanks, Andreas.
I just finished the packaging of those two libraries. Their gitwebs are
http://anonscm.debian.org/gitweb/?p=debian-science/packages/libdogleg.git
http://anonscm.debian.org/gitweb/?p=debian-science/packages/liblbfgs.git
libdogleg is more or less ready to go (i.e. I don't anticipate you'll find any
major issues).
liblbfgs has something broken with its upstream build system which causes it to
ignore external CFLAGS and LDFLAGS. This makes the hardening options be ignored,
and things like DEB_BUILD_OPTIONS=noopt don't work. I looked into fixing it, but
my automake knowledge isn't sufficient to do this yet. I don't know if this
issue precludes the package from being worthy of being uploaded. In either case,
help with this would be appreciated.
Thanks again.
dima
Reply to: