[Debian-science-sagemath] Overview of doctest failures

[François Bissey]

> On 20/10/2017, at 10:36, Ximin Luo <infinity0 at debian.org> wrote:
> 
> Ximin Luo:
>> [..]
>> 
>>> I didn't use the new flint-arb because it FTBFS for me so I got 67 failing doctests in sage/rings/complex_arb.pyx.
>>> 
> 
> Me too, I filed it here: https://github.com/fredrik-johansson/arb/issues/194
> 
> BTW I just noticed that ./configure has some options we're not using. Do we want to use them?
> 
>     --with-mpir=<path>   Specify location of MPIR (default: /usr/local)
>     --with-gmp=<path>    Specify location of GMP (default: /usr/local)
>     --with-mpfr=<path>   Specify location of MPFR (default: /usr/local)
>     --with-flint=<path>  Specify location of FLINT (default: /usr/local)
>     --with-blas[=<path>] Use BLAS and specify its location (default: /usr/local)
>     --with-ntl[=<path>]  Build NTL interface and specify its location (default: /usr/local)
> 
> Sage are passing --with-gmp and --with-mpfr:
> 
> https://git.sagemath.org/sage.git/tree/build/pkgs/arb/spkg-install
> 
> In Debian it's "traditional" to pass everything, and I wonder why Sage aren't passing the other ones. Does any of the Sage devs/packagers here know?
> 
> 
> Also: Tobias could you upload cypari2 1.1.3 to deb-sci-sage or experimental? Thanks in advance.
> 

I am not sure about every single one. sage probably should also pass —with-flint.
The thing is arb’s build system is copied from flint and has useless leftovers like
—with-blas and —with-ntl that are not relevant to arb. It is quite likely that spkg-install
from sage is also a partial copy of the flint one. And blas is not used in flint, I am not sure
why it is there.

François