Re: [sage-devel] Re: Sage python3 doctests

2019-01-27 Thread Frédéric Chapoton
random example, containing both #py2 and #py3 tags Le dimanche 27 janvier 2019 19:24:08 UTC+1, John Cremona a écrit : > > > > On Sun, 27 Jan 2019 at 17:10, Frédéric Chapoton > wrote: > >> ok, then --optional=sage,py3 is ok (all doctests pass) whereas >> >> ./sage -t --optional=sage,py2 src/sage/m

Re: [sage-devel] Re: Sage python3 doctests

2019-01-27 Thread Jeroen Demeyer
On 2019-01-27 19:23, John Cremona wrote: On Sun, 27 Jan 2019 at 17:10, Frédéric Chapoton mailto:fchapot...@gmail.com>> wrote: ok, then --optional=sage,py3 is ok (all doctests pass) whereas ./sage -t --optional=sage,py2 src/sage/modular/cusps_nf.py Is there a particular reason for th

Re: [sage-devel] Re: Sage python3 doctests

2019-01-27 Thread John Cremona
On Sun, 27 Jan 2019 at 17:10, Frédéric Chapoton wrote: > ok, then --optional=sage,py3 is ok (all doctests pass) whereas > > ./sage -t --optional=sage,py2 src/sage/modular/cusps_nf.py > Is there a particular reason for this file being used as an example here? I only say this as it was written by

Re: [sage-devel] Re: Sage python3 doctests

2019-01-27 Thread Vincent Delecroix
Le 27/01/2019 à 18:12, Vincent Delecroix a écrit : PS: without a decent "sage -t" I would not advice any developer to switch to Python3 Sage... https://trac.sagemath.org/ticket/27150 -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscrib

Re: [sage-devel] Re: Sage python3 doctests

2019-01-27 Thread Vincent Delecroix
But I don't understand - why the py2 / py3 flag is not set correctly in --optional (it should have been set by sage/doctest/control.py) - what is this python2 appearing in --optional? where it is set? Vincent PS: without a decent "sage -t" I would not advice any developer to switch to Pyt

Re: [sage-devel] Re: Sage python3 doctests

2019-01-27 Thread Frédéric Chapoton
ok, then --optional=sage,py3 is ok (all doctests pass) whereas ./sage -t --optional=sage,py2 src/sage/modular/cusps_nf.py runs the py2 doctest and fail as expected same thing for ./sage -t --optional=sage,py2,py3 src/sage/modular/cusps_nf.py Le dimanche 27 janvier 2019 18:06:19 UTC+1, vdelecro

Re: [sage-devel] Re: Sage python3 doctests

2019-01-27 Thread Vincent Delecroix
This should be $ ./sage -t --optional=sage,py3 src/sage/modular/cusps_nf.py ie don't forget "sage". Otherwise the doctest without # are not run. Le 27/01/2019 à 18:03, Frédéric Chapoton a écrit : This clearly needs a cleanup. For example, ./sage -t --optional=py2 src/sage/modular/cusps_nf.py

Re: [sage-devel] Re: Sage python3 doctests

2019-01-27 Thread Frédéric Chapoton
This clearly needs a cleanup. For example, ./sage -t --optional=py2 src/sage/modular/cusps_nf.py ./sage -t --optional=py3 src/sage/modular/cusps_nf.py have unexpected effects (when tested using python3) Le dimanche 27 janvier 2019 17:54:54 UTC+1, Frédéric Chapoton a écrit : > > Good question.

Re: [sage-devel] Re: Sage python3 doctests

2019-01-27 Thread Frédéric Chapoton
Good question. Needs to be checked Le dimanche 27 janvier 2019 17:47:17 UTC+1, vdelecroix a écrit : > > What about doctests marked # python2 / # python3? > > Le 27/01/2019 à 17:46, Frédéric Chapoton a écrit : > > Strange, but still this runs the correct python > > > > ./sage -t --long src/sage

Re: [sage-devel] Re: Sage python3 doctests

2019-01-27 Thread Vincent Delecroix
What about doctests marked # python2 / # python3? Le 27/01/2019 à 17:46, Frédéric Chapoton a écrit : Strange, but still this runs the correct python ./sage -t --long src/sage/categories/finite_posets.py # 15 doctests failed F Le dimanche 27 janvier 2019 17:35:41 UTC+1, vdelecroix a écrit :