On 10/13/2021 03:38 PM, Tracey Emery wrote:
On Wed, Oct 13, 2021 at 07:27:56PM +0000, Brian Callahan wrote:
On 10/13/2021 02:57 PM, Tracey Emery wrote:
> On Wed, Oct 13, 2021 at 03:42:53PM +0000, Brian Callahan wrote:
> >
> >
> > On 10/13/2021 11:24 AM, Tracey Emery wrote:
> > > On Wed, Oct 13, 2021 at 09:06:22AM -0600, Tracey Emery wrote:
> > > > Hello,
> > > >
> > > > Attached is a new port, math/isl. This is a new library to go with the
> > > > new xtensa-lx106-elf port build. The port has been tested successfully
> > > > on both amd64 and sparc64.
> > > >
> > > > pkg/DESCR:
> > > >
> > > > isl is a library for manipulating sets and relations of integer points
> > > > bounded by linear constraints. Supported operations on sets include
> > > > intersection, union, set difference, emptiness check, convex hull,
> > > > (integer) affine hull, integer projection, computing the lexicographic
> > > > minimum using parametric integer programming, coalescing and
> > > > parametric
> > > > vertex enumeration. It also includes an ILP solver based on
> > > > generalized
> > > > basis reduction, transitive closures on maps (which may encode
> > > > infinite
> > > > graphs), dependence analysis and bounds on piecewise step-polynomials.
> > > >
> > > > Ok?
> > > >
> > > > --
> > > >
> > > > Tracey Emery
> > >
> > > Now with 100% more attachment. Thanks tb@ for the eyes.
> >
> > Will this get picked up by egcc if it is installed?
> >
> > ~Brian
>
> Well, after building 8.4 and 11.2 successfully, I looked at the ports
> and both are configured for --without-isl. Guess I should have looked
> there first! hahahahaha!
>
> So, no, isl doesn't get picked up at this point.
Great. A couple of small things I noticed:
* Wikipedia claims there is a homepage at http://isl.gforge.inria.fr/
but it
appears to be down at the moment.
That's why I didn't add it. What do you think about using
https://guix.gnu.org/packages/isl-0.18/ for HOMEPAGE instead? It has
info and at least references the above URL.
No argument here.
* You only use the VERSION variable once, in DISTNAME, so it can
likely be
consolidated.
Yes, duh. Easy change. It was a quickly-thrown-together port. :D
* In the PLIST there is a lib/libisl.so.0.0-gdb.py entry. Should that
be
lib/libisl.so.${LIBisl_VERSION}-gdb.py?
This was autogenerated, so I'm not sure. Changing it makes sense, but
update-plist just changes it back, so future updates will have to be
manually maintained. That seems like a big hassle. Thoughts?
The gcc package manually maintains a similar thing in its PLIST:
PLIST-libs:13:lib/libestdc++.so.${LIBestdc++_VERSION}-gdb.py
~Brian
