Immutable entries of mutable attributes of char. tables

[ThomasBreuer]

Character tables use several mutable attributes whose values are
extendible lists:
ComputedClassFusions, ComputedIndicators, ComputedPowerMaps,
ComputedPrimeBlockss.

Up to now, the entries of these lists were mutable.
@frankluebeck suggested to make these entries immutable.
This change makes sense.

Note:
The change may have side-effects w.r.t. users' GAP code.
For example, the object returned by

0 * ComputedPowerMaps( CharacterTable( "A5" ) )[2]`

had been a mutable list before the changes,
and will be an immutable list after the changes.

In order to achieve the immutability,
the following changes were necessary.

• Extended SupportedCharacterTableInfo and DeclareAttributeSuppCT
  such that the mutability of the attribute in question is recorded;
  there seems to be no way (and no need) to access this information
  from the attribute itself,

• changed DeclarePropertySuppCT, here mutability makes no sense,

• let ConvertToCharacterTableNC make the entries of mutable attributes
  immutable.

Other changes in the context of mutability of attributes for character tables:

• Removed the DeclareAttributeSuppCT call for InfoText;
  meanwhile this attribute is declared for arbitrary GAP objects,
  redeclaring it for special objects is logically wrong,

• call MakeImmutable for (entries of) attribute values before the
  conversion of a record to a character table object,
  in order to avoid the creation of a copy in the standard process of
  setting attribute values,

• documented some optional components of an argument for
  ContainedPossibleCharacters.

[frankluebeck]

All these changes look sensible and useful. As mentioned by the author of this pull request, there is a small possibility that code which worked before is broken. But as far as I can see such code is probably buggy anyway, and it will always be easy to fix/correct it.

I have no idea why one of the tests fails, this seems to have to do with the Float package and I cannot see any relation with the changes in this pull request.

The only further change I would recommend is to add a test in some 'tst/testinstall/ctbl*' file, checking immutability of stored powermaps in some table.

[markuspf]

@frankluebeck @ThomasBreuer the test failures were due to the package float which we have disabled right now, as debugging this issue is not on our list of priorities.

[codecov]

Codecov Report

Merging #2657 into master will increase coverage by <.01%.
The diff coverage is 71.42%.

@@            Coverage Diff             @@
##           master    #2657      +/-   ##
==========================================
+ Coverage   75.41%   75.41%   +<.01%     
==========================================
  Files         478      478              
  Lines      241578   241606      +28     
==========================================
+ Hits       182192   182218      +26     
- Misses      59386    59388       +2

Impacted Files	Coverage Δ
lib/ctblgrp.gi	84.62% <100%> (ø)	⬆️
lib/ctblmaps.gi	70.54% <17.64%> (+0.05%)	⬆️
lib/ctblsymm.gi	29.77% <28.57%> (-0.02%)	⬇️
lib/ctblfuns.gi	60.52% <66.66%> (-0.02%)	⬇️
lib/ctbl.gi	71.91% <86.3%> (+0.09%)	⬆️
lib/ctbl.gd	75.86% <90.47%> (+5.02%)	⬆️
hpcgap/lib/hpc/stdtasks.g	64.96% <0%> (+0.51%)	⬆️

[fingolfin]

@frankluebeck you "requested changes", which blocks merging of this PR. But in what you write, I don't see any changes you "request" -- only a suggestion for some changes.

Could you either drop the "request for changes" (using the "Dismiss review" link on the PR page), or else clarify what the requested changes are?

[fingolfin]

The explanation and the changes seem plausibel and sensible to me. Caveat: I am not familiar with the actual, though. But I think Thomas is the expert on it, so I am happy to trust his judgement here.

[ThomasBreuer]

I have now added some tests of (im)mutability, as suggested by @frankluebeck.
This was a useful hint.

[fingolfin]

We need to describe this change for the release notes (see https://github.com/gap-system/gap/wiki/GAP-4.10-release-notes for the current draft), but I am not sure what to write, as this seems to mix various changes. @ThomasBreuer if you could provide a brief draft, that'd be great.

[ThomasBreuer]

@fingolfin I have added an item to the release notes.

[fingolfin]

Thank you!