OK.  But clearly it is a bug that the CoxeterGroup package raises:
  RuntimeError: coxeter3 must be installed
when coxeter3 *is* installed but sagemath_coxeter3 is not installed.

- Marc

On Sun, Jan 26, 2025 at 1:28 PM <dimp...@gmail.com> wrote:

> On Sun, Jan 26, 2025 at 11:03:42AM -0800, Marc Culler wrote:
> > Yes, I have a log file for coxeter3 which built fine.
> >
> > The problem turned out to be that I had to use BOTH ---enable-coxeter3
> AND
> > -enable-sagemath_coxeter3.  With the latter option, the .so file gets
> build
> > and installed as expected.
> >
> > Why?  What is this about?  And why does the traceback tell our user to
> > install the coxeter3 package when the coxeter3 package is installed
> already
> > and what is really needed is the sagemath_coxeter3 package?
>
> coxeter3 is just a C++ library.
> ---- (cf. its SPKG.rst)
> This package wraps Fokko Ducloux's Coxeter 3 C++ library
> ----
>
>
> sagemath_coxeter3 is a Python inteface to it,
> ---- (cf. its SPKG.rst)
> It provides a Cython interface to the ``coxeter3`` library.
> ----
> so that's akin to fplll vs fpylll, only optional, so they aren't
> installed by default. And yes, it's not an error to only install
> coxeter3.
>
> HTH
> Dima
>
> >
> > - Marc
> > On Sunday, January 26, 2025 at 12:49:02 PM UTC-6 dim...@gmail.com wrote:
> >
> > > On Sun, Jan 26, 2025 at 11:51 AM Marc Culler <marc....@gmail.com>
> wrote:
> > > >
> > > > In the venv of my recently built SageMath-10.6beta4 I find:
> > > > % ls lib/python3.12/site-packages/sage/libs/coxeter3
> > > > __init__.py coxeter.pxd decl.pxd __pycache__ coxeter.pyx
> > > > all__sagemath_coxeter3.py coxeter_group.py
> > > >
> > > > In other subdirectories of sage/libs, such as flint, one finds that
> > > every .pyx file has an associated .so file in the same subdirectory.
> So I
> > > would expect to find
> > >
> lib/python3.12/site-packages/sage/libs/coxeter3/coxeter-cpython-312-darwin.so,
>
> > > but it is not there.
> > > > That seems to suggest that coxeter.pyx is not being compiled by
> cython,
> > > and therefore that the extension module sage.libs.coxeter3.coxeter is
> not
> > > being built and hence is not installed.
> > > >
> > > > So it looks to me like the sagelib build process is skipping the
> step of
> > > cythonizing and compiling cython extension modules that are part of an
> > > optional package.
> > > >
> > > > Am I missing some step which is needed for this to happen? I have
> used
> > > the --enable-coxeter3 flag in the configure command. Do I need to do
> > > something else?
> > >
> > > do you have anything in logs/pkgs/coxeter* ?
> > > If not, and if you use the standard make build (not meson-based one)
> you
> > > may try
> > >
> > > make coxeter3
> > >
> > > and see how it goes.
> > >
> > > Dima
> > >
> > > >
> > > > - Marc
> > > >
> > > >
> > > > --
> > > > You received this message because you are subscribed to the Google
> > > Groups "sage-devel" group.
> > > > To unsubscribe from this group and stop receiving emails from it,
> send
> > > an email to sage-devel+...@googlegroups.com.
> > > > To view this discussion visit
> > >
> https://groups.google.com/d/msgid/sage-devel/0762e04b-33ba-4bfa-8ee4-e145163dd7f7n%40googlegroups.com
> > > .
> > >
> >
> > --
> > You received this message because you are subscribed to the Google
> Groups "sage-devel" group.
> > To unsubscribe from this group and stop receiving emails from it, send
> an email to sage-devel+unsubscr...@googlegroups.com.
> > To view this discussion visit
> https://groups.google.com/d/msgid/sage-devel/097be2ad-5697-476b-b7cf-303a91359644n%40googlegroups.com
> .
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To view this discussion visit 
https://groups.google.com/d/msgid/sage-devel/CALcZXRF69QERhMU97Nww-__F20OyuUH_A0udiZFR0euQz80vrw%40mail.gmail.com.

Reply via email to