-
Notifications
You must be signed in to change notification settings - Fork 161
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
StructureDescription(G:nice) breaks if group has different presentations because cannot compute centralizer in a group given by a matrix #973
Comments
Ok, I figured out the main problem, but I do not know how to resolve it. GAP computes the normal subgroups, then for each normal subgroup
In particular, after going out from the break loop, this is what I find:
The strange thing is that the other way around GAP has no problem determining that the two are different:
I have no idea how to resolve this. |
What is happening is that the nice monomorphism for a subgroup (defined over a smaller field) -- the `nice' in the example -- does not test fully whether an element of the whole group is in its domain. The reason behind this problem is the use of a set of subgroups in the code -- unless groups are permutation group (and thus naturally subsets of one larger group) the implementation of the As an aside: Does this output of structure description really help a user? I think it would be far more helpful (if it really is intended to call
This of course will not work for all groups either and not scale well. |
In case it may be useful to know where this test comes from: this is the test for the bug reported by Nilo de Roock (http://mail.gap-system.org/pipermail/forum/2006/001338.html ) and fixed by @Stefan-Kohl on 14 Feb 2006. Description of the fix: "For certain matrix groups, the function So the test was not checking for a specific structure description for the group, but was in fact checking that it works without an error (unrelated to |
@alex-konovalov Thank you. Well, it helps knowing how these groups arise, but this is still an error. For such groups @hulpke Thank you, what you say about sets is indeed problematic. For now I have replaced all of my
However, there still seems to be some other glitch, because the error still occurs. (In fact, it even occurs if I only use About good descriptions: it has been already challenged (e.g. in #160) that for p-groups this may not be a usable description. I do not mind opening an issue about discussing how |
…ot avoid method selection. Also nice method for Struct.Desc. This fixes gap-system#973
…ot avoid method selection. Also nice method for Struct.Desc. This fixes gap-system#973 Impies subsequent fix in test file
Ok, I think I have found the culprit. GAP cannot compute the centralizer of a subgroup in another one for this example:
Is this intended behaviour or a bug? (Note that I updated the title once again to reflect this.) |
…ot avoid method selection. Also nice method for Struct.Desc. This fixes gap-system#973 Impies subsequent fix in test file Instead of testing for the explict structure (which might be x or :), test whether the Struct.Desc. command runs thorugh.
…ot avoid method selection. Also nice method for Struct.Desc. This fixes gap-system#973 Implies subsequent fix in test file -- do not test for explicit string value of `StructureDescription` (which can change) but only whether the function succeeds (which is after all what this test was supposed to do).
…ethod selection. Also nice method for Struct.Desc. This fixes gap-system#973 Implies subsequent fix in test file -- do not test for explicit string value of `StructureDescription` (which can change) but only whether the function succeeds (which is after all what this test was supposed to do).
This splits the current general StructureDescription code for an abelian for a finite and for a finite simple method, with redispatches. Further, it adds nice monomorphism in the proper way (into grpnice.gi), while removes it from grpnames.gi. Corrects problematic tests, and fixes the StructureDescription part of #973. Added some more tests to increase the code coverage of grpnames.gi. Now most lines are covered for StructureDescription, DirectFactors and SemidirectDecompositions. Manual entries are cleaned up in these three functions and in some of the functions they use in grpnames.gd (for easier future documentation).
Fixed by #985 |
This affects current master.
Update2: The problems is really about GAP not being able to calculate a particular centralizer in a matrix group. See my third post below.
Update: The problem is really about adding list of groups to a set, and determining if a new list of groups is equal to any one of them in the set. See the comment below.
This was the original bug:
Somehow
StructureDescription(G:nice)
seems to break depending on the presentation of the group. Here are two particular groups with this problem:I am trying to debug this. It looks to me that nice monomorphisms are involved here, and I have no idea about them (browsing the manual just now), so any help would be appreciated. Thanks.
The text was updated successfully, but these errors were encountered: