GAP3 does not have IsIdenticalObject

[vbraun]

Looks like a regression introduced in #19585. Reported at ​https://groups.google.com/d/msg/sage-devel/MrXp43v3WP4/ju2LwTuvBAAJ

sage: gap=Gap3('/Applications/gap3r4p4/bin/gap')
sage: gap.RequirePackage('"specht"')
---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
<ipython-input-22-c539883f0214> in <module>()
----> 1 gap.RequirePackage('"specht"')

/Applications/sage/local/lib/python2.7/site-packages/sage/interfaces/interface.pyc in __call__(self, *args, **kwds)
    605 
    606     def __call__(self, *args, **kwds):
--> 607         return self._parent.function_call(self._name, list(args), kwds)
    608 
    609     def _sage_doc_(self):

/Applications/sage/local/lib/python2.7/site-packages/sage/interfaces/gap.pyc in function_call(self, function, args, kwds)
    919             self.eval(marker)
    920             res = self.eval(cmd)
--> 921         if self.eval('IsIdenticalObj(last,__SAGE_LAST__)') != 'true':
    922             return self.new('last2;')
    923         else:

/Applications/sage/local/lib/python2.7/site-packages/sage/interfaces/gap.pyc in eval(self, x, newlines, strip, split_lines, **kwds)
    567             if not input_line.endswith(';'):
    568                 input_line += ';'
--> 569         result = Expect.eval(self, input_line, **kwds)
    570         if not newlines:
    571             result = result.replace("\\\n","")

/Applications/sage/local/lib/python2.7/site-packages/sage/interfaces/expect.pyc in eval(self, code, strip, synchronize, locals, allow_use_file, split_lines, **kwds)
   1237                 elif split_lines:
   1238                     return '\n'.join([self._eval_line(L, allow_use_file=allow_use_file, **kwds)
-> 1239                                         for L in code.split('\n') if L != ''])
   1240                 else:
   1241                     return self._eval_line(code, allow_use_file=allow_use_file, **kwds)

/Applications/sage/local/lib/python2.7/site-packages/sage/interfaces/gap.pyc in _eval_line(self, line, allow_use_file, wait_for_prompt, restart_if_needed)
    765                     return ''
    766             else:
--> 767                 raise RuntimeError(message)
    768 
    769         except KeyboardInterrupt:

RuntimeError: Gap3 produced error output
Error, Variable: 'IsIdenticalObj' must have a value

   executing IsIdenticalObj(last,__SAGE_LAST__);

[tscrim]

So GAP3 has IsIdentical, which was changed to IsIdenticalObj in GAP4b3. I think what we need to do is either:

• a subclass of the GAP interpreter for backwards compatibility with GAP3 which just has the one difference in function_call(),
• check the GAP version in the __init__ and set some private attribute which gets used in function_call(), or
• store the GAP version as an attribute and use an if statement in function_call().

@aschilling For now you should be able to make that one line change mentioned above to sage.interfaces.gap.function_call (but I haven't tested it).

[jaanos]

Hi!

• a subclass of the GAP interpreter for backwards compatibility with GAP3 which just has the one difference in function_call(),

There already is a separate class for GAP3 which inherits from Gap_generic. The cleanest solution IMO would be to have a method _identical_function in Gap_generic (parallel to _equality_symbol etc.), say:

def _identical_function(self):
    return "IsIdenticalObj"

and then have it overridden in Gap3.

[tscrim]

Replying to jaanos:

• a subclass of the GAP interpreter for backwards compatibility with GAP3 which just has the one difference in function_call(),

There already is a separate class for GAP3 which inherits from Gap_generic.

Good; let us use that.

The cleanest solution IMO would be to have a method _identical_function in Gap_generic (parallel to _equality_symbol etc.), say:

def _identical_function(self):
    return "IsIdenticalObj"

and then have it overridden in Gap3.

I think it should be a class-level attribute instead of a method because it speed cost and a method is IMO somewhat overkill (because no work is done).

Anne, can you check to see if this works?

---

New commits:

​8ff1fe8

Adding an identical function class attribute.

[aschilling]

Replying to tscrim:

Replying to jaanos:

• a subclass of the GAP interpreter for backwards compatibility with GAP3 which just has the one difference in function_call(),

There already is a separate class for GAP3 which inherits from Gap_generic.

Good; let us use that.

The cleanest solution IMO would be to have a method _identical_function in Gap_generic (parallel to _equality_symbol etc.), say:

def _identical_function(self):
    return "IsIdenticalObj"

and then have it overridden in Gap3.

I think it should be a class-level attribute instead of a method because it speed cost and a method is IMO somewhat overkill (because no work is done).

Anne, can you check to see if this works?

---

New commits:

​8ff1fe8

Adding an identical function class attribute.

With this branch applied, I still get an error

sage: gap=Gap3('/Applications/gap3r4p4/bin/gap')
sage: gap.RequirePackage('"specht"')
---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
<ipython-input-2-c539883f0214> in <module>()
----> 1 gap.RequirePackage('"specht"')

/Applications/sage/local/lib/python2.7/site-packages/sage/interfaces/interface.pyc in __call__(self, *args, **kwds)
    605 
    606     def __call__(self, *args, **kwds):
--> 607         return self._parent.function_call(self._name, list(args), kwds)
    608 
    609     def _sage_doc_(self):

/Applications/sage/local/lib/python2.7/site-packages/sage/interfaces/gap.pyc in function_call(self, function, args, kwds)
    920             self.eval(marker)
    921             res = self.eval(cmd)
--> 922         if self.eval(self._identical_function + '(last,__SAGE_LAST__)') != 'true':
    923             return self.new('last2;')
    924         else:

/Applications/sage/local/lib/python2.7/site-packages/sage/interfaces/gap.pyc in eval(self, x, newlines, strip, split_lines, **kwds)
    568             if not input_line.endswith(';'):
    569                 input_line += ';'
--> 570         result = Expect.eval(self, input_line, **kwds)
    571         if not newlines:
    572             result = result.replace("\\\n","")

/Applications/sage/local/lib/python2.7/site-packages/sage/interfaces/expect.pyc in eval(self, code, strip, synchronize, locals, allow_use_file, split_lines, **kwds)
   1237                 elif split_lines:
   1238                     return '\n'.join([self._eval_line(L, allow_use_file=allow_use_file, **kwds)
-> 1239                                         for L in code.split('\n') if L != ''])
   1240                 else:
   1241                     return self._eval_line(code, allow_use_file=allow_use_file, **kwds)

/Applications/sage/local/lib/python2.7/site-packages/sage/interfaces/gap.pyc in _eval_line(self, line, allow_use_file, wait_for_prompt, restart_if_needed)
    766                     return ''
    767             else:
--> 768                 raise RuntimeError(message)
    769 
    770         except KeyboardInterrupt:

RuntimeError: Gap3 produced error output
Error, Variable: 'IsIdenticalObj' must have a value

   executing IsIdenticalObj(last,__SAGE_LAST__);

I will be traveling for at least 15 hours now and won't have any internet on the plane ....

[git]

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

​efdffd5

Putting the class attribute in the correct place.

[tscrim]

Whoops. I put it in the wrong place. Try it now.

Have a great flight!

[aschilling]

Replying to tscrim:

Whoops. I put it in the wrong place. Try it now.

Yes, now it works!

sage: gap=Gap3('/Applications/gap3r4p4/bin/gap')
sage: gap.RequirePackage('"specht"')
------------------------------------------------------
SPECHT 3.1 - development edition                      
    A package for calculating decomposition numbers of
    Hecke algebras of the symmetric groups and q-Schur
    algebras.

    This is an unofficial release of SPECHT. There are
    a significant number of  *undocumented*  functions
    in this version. Fewer old bugs, and more new bugs
    are also included.

(C) Andrew Mathas  mathas@maths.usyd.edu.au     Sydney
------------------------------------------------------

[tscrim]

Can we consider that a positive review then?

[aschilling]

Replying to tscrim:

Can we consider that a positive review then?

Thank you for fixing this! Happy New Year!