Fix missing synchronisation upon starting gap3

[nthiery]

Bug reported by Jean Michel:

┌────────────────────────────────────────────────────────────────────┐
│ SageMath version 8.0.beta9, Release Date: 2017-05-31               │
│ Type "notebook()" for the browser-based notebook interface.        │
│ Type "help()" for help.                                            │
└────────────────────────────────────────────────────────────────────┘
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃ Warning: this is a prerelease version, and it may be unstable.     ┃
┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛

sage: gap3.eval("1+1")
''
sage: gap3.eval("1+2")
2
sage: gap3.eval("1+3")
3

Equivalently, the above can be reproduced with the %%gap3 magic.

Analysis: unlike the gap interface, no synchronization was done upon starting the gap3 interface. This did not show up when calling gap3(cmd) as the latter triggers an explicit synchronization.

This ticket adds the missing synchronization and, upon suggestion by Jean, adds the option -b in the call to gap3 (no banner).

[tscrim]

Is there suppose to be a branch or marked as invalid?

[tscrim]

Okay, it seems the commit got mixed in with #23132.

[nthiery]

Oops; thanks for spotting this. Fixed!

---

New commits:

​fda6e26

23132: Fix missing synchronisation upon starting gap3 + doc + skip gap3 banner

[tscrim]

LGTM.

[nthiery]

Thanks!

[chapoton]

hey, guys.. doctests do not pass...

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​5079877

23142: trivial doctest update

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​529e4d3

23142: another trivial doctest update

[nthiery]

Arr, sorry! Apparently I am getting lazy due to being used to having a continuous integration tool immediately running the tests for me ... I can't wait until we really have this for Sage (e.g. more patchbots; we really need to set up a couple of them in Orsay). In the mean time, I indeed should have waited for my Sage to compile and run the tests for real before setting this for review.

[nthiery]

The following failures already appear with develop:

mistral-/opt/sage/src/sage/interfaces>sage -t --optional gap3 gap3.py 
too many failed tests, not using stored timings
Running doctests with ID 2017-06-06-16-24-24-c343198f.
Git branch: develop
Using --optional=gap3
Doctesting 1 file.
sage -t gap3.py
**********************************************************************
File "gap3.py", line 441, in sage.interfaces.gap3.Gap3.help
Failed example:
    m                                        #optional - gap3
Expected:
    [ [ 1, 2, 3 ], [ 4, 5, 6 ] ]
Got:
    <BLANKLINE>
**********************************************************************
File "gap3.py", line 443, in sage.interfaces.gap3.Gap3.help
Failed example:
    m.Print()                                #optional - gap3
Expected:
    [ [ 1, 2, 3 ], [ 4, 5, 6 ] ]
Got:
    "__SAGE_LAST__"
**********************************************************************
File "gap3.py", line 447, in sage.interfaces.gap3.Gap3.help
Failed example:
    m                                        #optional - gap3
Expected:
    [ [ 1, 2, 3 ], [ 4, 5, 6 ] ]
Got:
    <BLANKLINE>
**********************************************************************
File "gap3.py", line 449, in sage.interfaces.gap3.Gap3.help
Failed example:
    m.Print()                                #optional - gap3
Expected:
    [ [ 1, 2, 3 ], [ 4, 5, 6 ] ]
Got:
    "__SAGE_LAST__"
**********************************************************************
1 item had failures:
   4 of  10 in sage.interfaces.gap3.Gap3.help
    [95 tests, 4 failures, 11.90 s]
----------------------------------------------------------------------
sage -t gap3.py  # 4 doctests failed
----------------------------------------------------------------------
Total time for all tests: 18.0 seconds
    cpu time: 4.2 seconds
    cumulative wall time: 11.9 seconds

Something should be done about them, but it's an independent problem. Hence back to needs review.

[nthiery]

Out of curiosity, I just tried fixing the above problem by adding a self._expect.expect("@i") at the end of Gap3.help. This in case there would be a trivial fix to include in this ticket. This seemed at first to work, but not quite; it triggers another error

sage -t gap3.py
**********************************************************************
File "gap3.py", line 717, in sage.interfaces.gap3.GAP3Element.__getitem__
Failed example:
    a[1]                                     #optional - gap3
Expected:
    1
Got:
    2

I am therefore dropping the case.

[tscrim]

Ah, right, gap3 is not optional but experimental. I also get the same errors with

./sage -tp --optional=sage,gap3 src/sage/interfaces/gap3.py

on both develop and this branch.