Re: [sage-devel] snappy in Sage

2018-10-02 Thread Nathan Dunfield
> > > The idea is that cython produces fully compliant C code, so that no > tuning > > of the generated C code is required. Distributing the cythonized c-files > > has the advantage that the installing user does not need cython > installed. > > In sage we have run into trouble with that due t

Re: [sage-devel] snappy in Sage

2018-10-02 Thread Nathan Dunfield
> > Thanks, Thierry. Sounds like we need a ticket for this. In particular, I > wasn't even able to get sage -pip install mercurial to work on OS X. > I was just helping a colleague install SnapPy into SageMath on OS X and hit the above issue. I just downloaded the source tarball from BitBuck

Re: [sage-devel] snappy in Sage

2018-10-02 Thread Jeroen Demeyer
On 2018-10-02 23:01, Dima Pasechnik wrote: So you need to ship generated C files for C99, C11, and C++ files for c++11, c++14, c++17, blah++foo, etc etc etc? No. You *do* need to tell Cython the language (C or C++). But beyond that, it generates code compatible with any version of the C/C++ st

Re: [sage-devel] snappy in Sage

2018-10-02 Thread Dima Pasechnik
On Tue, Oct 2, 2018 at 9:51 PM Jeroen Demeyer wrote: > > On 2018-10-02 17:53, Dima Pasechnik wrote: > > Letting Cython to do necessary adjustments for the target C/C++ > > compiler seems like the most natural thing to do... > > Cython doesn't need to do any such "necessary adjustments". It's > com

Re: [sage-devel] snappy in Sage

2018-10-02 Thread Jeroen Demeyer
On 2018-10-02 17:53, Dima Pasechnik wrote: Letting Cython to do necessary adjustments for the target C/C++ compiler seems like the most natural thing to do... Cython doesn't need to do any such "necessary adjustments". It's completely independent of C/C++ compilers (and it should be). -- You

Re: [sage-devel] snappy in Sage

2018-10-02 Thread Thierry
On Tue, Oct 02, 2020 at 08:39:42AM -0700, kcrisman wrote: > > > On Tuesday, October 2, 2018 at 10:55:14 AM UTC-4, Samuel Lelievre wrote: > > > > One can also get the "Computational topology Sage Docker image": > > > > https://hub.docker.com/r/computop/sage/ > > > > It ships among other things:

Re: [sage-devel] snappy in Sage

2018-10-02 Thread Vincent Delecroix
Le 02/10/2018 à 18:31, Nils Bruin a écrit : On Tuesday, October 2, 2018 at 8:54:12 AM UTC-7, Dima Pasechnik wrote: On Tue, Oct 2, 2018 at 4:44 PM Nathan Dunfield > wrote: It's bad packaging by upstream: it's running Cython but the Cython source files are not in the snappy source tarball.

Re: [sage-devel] snappy in Sage

2018-10-02 Thread Nils Bruin
On Tuesday, October 2, 2018 at 8:54:12 AM UTC-7, Dima Pasechnik wrote: > > On Tue, Oct 2, 2018 at 4:44 PM Nathan Dunfield > wrote: > >> > >> It's bad packaging by upstream: it's running Cython but the Cython > >> source files are not in the snappy source tarball. > > > > > > Yes, we ship t

Re: [sage-devel] snappy in Sage

2018-10-02 Thread Dima Pasechnik
On Tue, Oct 2, 2018 at 4:44 PM Nathan Dunfield wrote: >> >> It's bad packaging by upstream: it's running Cython but the Cython >> source files are not in the snappy source tarball. > > > Yes, we ship the Cython generated C/C++ files rather than Cython code itself. > My understanding from the Cyt

Re: [sage-devel] snappy in Sage

2018-10-02 Thread Nathan Dunfield
> > It's bad packaging by upstream: it's running Cython but the Cython > source files are not in the snappy source tarball. > Yes, we ship the Cython generated C/C++ files rather than Cython code itself. My understanding from the Cython docs is that this is the recommended approach for non-S

[sage-devel] Re: snappy in Sage

2018-10-02 Thread Nathan Dunfield
This is a known bug and we are working on it (there's even a ticket). It should be fixed soon. It's some issue with our configuration script. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving em

Re: [sage-devel] snappy in Sage

2018-10-02 Thread kcrisman
On Tuesday, October 2, 2018 at 10:55:14 AM UTC-4, Samuel Lelievre wrote: > > One can also get the "Computational topology Sage Docker image": > > https://hub.docker.com/r/computop/sage/ > > It ships among other things: > > - SageMath > - SnapPy > - Regina > - PHCpack > Absolutely! I actually

Re: [sage-devel] snappy in Sage

2018-10-02 Thread kcrisman
On Tuesday, October 2, 2018 at 10:52:43 AM UTC-4, Jeroen Demeyer wrote: > > On 2018-10-02 16:50, kcrisman wrote: > > Thanks, Thierry. Sounds like we need a ticket for this. > > I assume you mean a Snappy ticket since the problem with Snappy is > completely unrelated to Sage. > > Oh! I guess

Re: [sage-devel] snappy in Sage

2018-10-02 Thread Samuel Lelievre
One can also get the "Computational topology Sage Docker image": https://hub.docker.com/r/computop/sage/ It ships among other things: - SageMath - SnapPy - Regina - PHCpack There is a link to it from the "Distribution" page on the Sage wiki: https://wiki.sagemath.org/Distribution -- You

Re: [sage-devel] snappy in Sage

2018-10-02 Thread Jeroen Demeyer
On 2018-10-02 16:50, kcrisman wrote: Thanks, Thierry. Sounds like we need a ticket for this. I assume you mean a Snappy ticket since the problem with Snappy is completely unrelated to Sage. Note also (in general) that You are using pip version 9.0.3, however version 18.0 is available. Yo

Re: [sage-devel] snappy in Sage

2018-10-02 Thread kcrisman
> I had a similar issue. The fix was > > - to install mercurial on your distro or with: > > sage -pip install mercurial > > No way! Hg rises from the dead! Love it. > - to install snappy from the repo > > sage -pip install hg+https://bitbucket.org/t3m/snappy > > > Tha

Re: [sage-devel] snappy in Sage

2018-10-02 Thread Thierry
Hi, On Tue, Oct 02, 2018 at 07:59:34AM +0200, Jeroen Demeyer wrote: > On 2018-10-01 20:19, kcrisman wrote: > > new = max(os.path.getmtime(file) for file in > >glob('cython/core/*.pyx')) > > > > ValueError: max() arg is an empty sequence > > It's bad packaging by upstream: it's running