Hi Dima,
On 2018-12-26, Dima Pasechnik <dimp...@gmail.com> wrote:
> E.g. do not do
>
> G = libgap('DihedralGroup(8)')
>
> instead do
>
> G = libgap.DihedralGroup(8)
Sure, that's what I do. I also used to do gap.DihedralGroup(8), which works
just as well. I just chose an example to point out that the two call methods
have a different semantics.
>> sage: L = gap('[1,2,3]')
>> sage: L[1]
>> 1
>
> this is guaranteed constant headache, leading to madness or more
> likely abandonment and bitrot, having in Python list-like 1-based
> objects,
> next to normal 0-based objects... Pythonisti fed devs to hungry serpents for
> lesser degrees of non-Pythonness...
Sure. I am saying that GAP did a wrong choice, and arguably it was a bad choice
in gap to follow GAP's convention. libgap being 0-based is good!
However, it means libgap cannot be a drop-in replacement for gap, and
the change from gap to libgap will, for some time, create trouble for
users. So, one cannot just remove gap all of a sudden, in particular not
without a deprecation. And the same should hold for r.
>> - libgap.__call__ on a string does not interprete that string but
>> returns a GAP string.
>
> I don't understand - you get a Python wrapper to a GAP object if you
> call libgap appropriately,
> not a string.
No. See the examples I gave. libgap('bla') gives GAP string (which of
course is my abbreviation for "Python wrapper of a GAP string").
gap('bla') tries to evaluate 'bla' as a command.
Similarly to the above 1- versus 0-based, I am saying that libgap
arguably works better than gap (so, please don't change libgap now!).
But again, it means a headache for users and thus removing gap and
replacing it with libgap must not happen without a deprecation.
> I'd like to have a look at this code: perhaps half, if not all, of
> libgap.eval()
> calls can be replaced by something more Pythonic.
Currently I use it for the following purposes:
1. Creating examples in the doc strings. Reason: I simply don't know
how to express the GAP command
Group( [ (1,2)(3,8)(4,6)(5,7), (1,3)(2,5)(4,7)(6,8) ] )
in libgap. I tried
libgap.Group( [ ((1,2),(3,8),(4,6),(5,7)), ((1,3),(2,5),(4,7),(6,8)) ] )
but it fails.
2. Testing if a construction fails, by a comparison such as
if phi == libgap.eval('fail')...
I would prefer "if phi == libgap.fail...", but it fails.
3. Calling GAP's `InstallValue`. I tried to do it via
libgap.function_factory, but without success.
4. Pickling. GAP has, as far as I can see, very poor support for
serialisation. In my application, I am dealing with permutation
groups, that can be reconstructed from their string representation.
So, I store this and do unpickling by libgap.eval(...).
I would be happy to learn of more efficient ways of pickling a group
in GAP, but for backwards compatibility (opening old pickles) I need to keep
evaluating strings.
Best regards,
Simon
--
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+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.
