test failures with buckygen (optional package)

[Vincent Delecroix]

Dear all,

On sage 9.3 and buckygen installed I got a lot of test failures
in graphs/graph_generators.py of the type

sage -t --long --random-seed=0 src/sage/graphs/graph_generators.py
**********************************************************************
File "src/sage/graphs/graph_generators.py", line 1233, in
sage.graphs.graph_generators.GraphGenerators.fullerenes
Failed example:
len(list(gen)) # optional buckygen
Exception raised:
Traceback (most recent call last):
File
"/home/vdelecro/sage_patchbot/local/lib/python3.6/site-packages/sage/doctest/forker.py",
line 714, in _run
self.compile_and_execute(example, compiler, test.globs)
File
"/home/vdelecro/sage_patchbot/local/lib/python3.6/site-packages/sage/doctest/forker.py",
line 1133, in compile_and_execute
exec(compiled, globs)
File "<doctest
sage.graphs.graph_generators.GraphGenerators.fullerenes[1]>", line 1, in
<module>
len(list(gen)) # optional buckygen
File
"/home/vdelecro/sage_patchbot/local/lib/python3.6/site-packages/sage/graphs/graph_generators.py",
line 1301, in fullerenes
sp.stdout.reconfigure(newline='')
AttributeError: '_io.TextIOWrapper' object has no attribute
'reconfigure'
**********************************************************************

Best
Vincent

[Vincent Delecroix]

Indeed, the method reconfigure was introduced in Python 3.7 (and I do
have Python 3.6 on my system). The proper way to go is to use the
codecs library.

Vincent

[Dima Pasechnik]

we are dropping Python 3.6,  it will reach eol in Dec.

--
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 on the web visit https://groups.google.com/d/msgid/sage-devel/b7508205-ed86-8444-6e72-d4ae429d58e7%40gmail.com.

[Vincent Delecroix]

1. sage-9.3 claims that it supports Python 3.6. My system Python
was by the way accepted by sage configure script.

2. the part of sage source code that I mentioned in my initial e-mail
does not support Python 3.6

Hence this is a bug. Did I miss something?

[Dima Pasechnik]

On Fri, May 14, 2021 at 8:23 PM Vincent Delecroix

<20100.d...@gmail.com> wrote:
>
> 1. sage-9.3 claims that it supports Python 3.6. My system Python
> was by the way accepted by sage configure script.
>
> 2. the part of sage source code that I mentioned in my initial e-mail
> does not support Python 3.6
>
> Hence this is a bug. Did I miss something?

no, you're right, it's a bug in 9.3.
I've argued for dropping Python 3.6 for last 6 months, but
it only made it up to here https://trac.sagemath.org/ticket/30551
- it will be dropped in 9.4.

We've been going back and forth on the support of Python 3.6, which is a pain.

> To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/886a0c10-b126-ddf9-04d4-ed03905113e2%40gmail.com.

[Matthias Koeppe]

On Friday, May 14, 2021 at 12:23:49 PM UTC-7 vdelecroix wrote:

1. sage-9.3 claims that it supports Python 3.6. My system Python
was by the way accepted by sage configure script.

It did also warn you that using system python 3.6 is deprecated.

 

2. the part of sage source code that I mentioned in my initial e-mail
does not support Python 3.6

Yes, this is a bug in Sage 9.3. We have no real infrastructure for systematically testing doctests depending on optional packages. There are tests on GH Actions that test that all  optional packages can be built; but it does not run the Sage testsuite with optional packages installed.