Permutations to get alignment of ordering with Magma

[hulpke]

These have been explicitly checked by comparing CodePcGroup.

Program used: In Magma, fetch data through

p:=3;e:=7;
l:=SearchPGroups(p,e);;
SetLogFile("gnum3e7.g");
print "l:=[\n";
for i in [1..#l] do
  a,b:=SmallGroupEncoding(l[i]);
  print a,",";
end for;
print "];";

(Then edit the file to throw out junk) and compare in GAP.

[fingolfin]

Very good, thank you very much. This should ideally be released ASAP, so that it is ready for GAP 4.10.

It would be great if there also was a CHANGES or NEWS file in the release, which points out this change (and any other that may have happened since 1.3).

CC @alex-konovalov

[olexandr-konovalov]

You removed a comment sign, and now tests fail - please restore it or remove an empty line at all, otherwise tests fail.

[hulpke]

Thanks. Will be fixed.

[codecov]

Codecov Report

Merging #39 into master will increase coverage by 11.03%.
The diff coverage is 98.31%.

@@             Coverage Diff             @@
##           master      #39       +/-   ##
===========================================
+ Coverage   88.75%   99.79%   +11.03%     
===========================================
  Files          22      491      +469     
  Lines        4589   325067   +320478     
===========================================
+ Hits         4073   324395   +320322     
- Misses        516      672      +156

Impacted Files	Coverage Δ
gap/small.gd	100% <100%> (ø)	⬆️
gap/small.gi	95.74% <33.33%> (-1.23%)	⬇️
small11/smlgp11.g	13.29% <0%> (-81.73%)	⬇️
small5/smlgp5.g	100% <0%> (ø)	⬆️
gap/idgrp1.g	100% <0%> (ø)	⬆️
gap/smlgp1.g	100% <0%> (ø)	⬆️
id5/idgrp5.g	100% <0%> (ø)	⬆️
id2/id896.zan	100% <0%> (ø)
id2/id256.l	100% <0%> (ø)
... and 483 more

[hulpke]

@fingolfin
Yes, there needs to be more documentation and Bettina was planning to write explicitly to the gap-forum. I just created the branch now there is no duplication of work, as I saw @markuspf had started as well.

changes have been made

[hulpke]

Please do not merge yet -- I'm still in discussion with Bettina and Eamonn on the correct approach

[hulpke]

@alex-konovalov Thanks for creating the do not merge label!

[markuspf]

@hulpke I had started, but then abandoned doing stuff, because I noticed that the fix would be much more complicated than what I first thought (and I saw you doing the more complicated fix already).

[markuspf]

As an additional note:

I would quite like to know how the comparison between MAGMA and GAP happened; Do we have code to re-run this/validate this?

If we want to insist that the numberings between GAP and MAGMA are consistent, it would be good if we had a way to verify this, at least in terms of some random samples.

I tried but couldn't really be bothered finding out how to even compare groups between MAGMA and GAP (that is probably entirely my own laziness/inability), because I couldn't even easily figure out how to get a pcp out of MAGMA.

[olexandr-konovalov]

Would be nice to have some progress and make a release in October to ensure it goes into GAP 4.10.

[hulpke]

@alex-konovalov Do not delay 4.10 for catching this update. The change will be larger as it will also have to involve Magma, there will be an update of the package in due course, but not synchronized with the GAP release.

[hulpke]

@markuspf The comparison was done by printing out the encoding (equivalent to CodePcGroup in GAP), as listed on the very top of this PR.

[olexandr-konovalov]

@hulpke do you have any updates, by any chance? This PR still has "do not merge" label, and the branch has now conflicts with master...

[hulpke]

No update from me. In fact the problem will likely be dealt with by explicitly not promising an order for p^7 (as part of it might be generated on the fly in Magma and thus could change order).

I will thus close this PR.

[olexandr-konovalov]

A new attempt is now suggested in #46 - thanks @hulpke