-
Notifications
You must be signed in to change notification settings - Fork 165
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
Two fixes for fp group homomorphisms and Improvement for Fitting free matrix group setup #3103
Conversation
You write 2x
"
gi:=MappingGeneratorsImages(hom)[2]{ListPerm(PermList(hom!.genpositions)^-1,
Length(hom!.genpositions))};". .
I think it would be much more efficient as
gi{hom!.genpositions} := MappingGeneratorsImages (hom)[2];
I.e. without creating extra permutations and lists.
…On Wed, Dec 12, 2018, 09:33 Alexander Hulpke ***@***.*** wrote:
@hulpke <https://github.com/hulpke> requested your review on: #3103
<#3103> Two fixes for fp group
homomorphisms and Improvement for Fitting free matrix group setup.
—
You are receiving this because your review was requested.
Reply to this email directly, view it on GitHub
<#3103 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/ACIIUMzp8rEm8GS46PNH3tgCtlawYUrvks5u4L9JgaJpZM4ZN0xg>
.
|
@@ -28,7 +28,17 @@ local cache,ffs,pcisom,rest,it,kpc,k,x,ker,r; | |||
|
|||
pcisom:=ffs.pcisom; | |||
|
|||
rest:=RestrictedMapping(ffs.factorhom,U); | |||
#rest:=RestrictedMapping(ffs.factorhom,U); | |||
if IsPermGroup(U) and AssertionLevel()>1 then |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The commit message says:
using
RestrictedMapping
as it will not work for certain matrix group homomorphisms
So, does that mean that there is a bug in RestrictedMapping
? Or is this more a general problem where e.g. RestrictedMapping
doesn't have enough information to do something sensible in all cases? Or perhaps a third option?
Also, what exactly is a "matrix group homomorphisms" in this context? A hom with a matrix group as source? Range? Both? Either? And, based on that, I'd have expected that only these kinds of homs would be treated separately here, and the rest handled by RestrictedMapping
anyway, but that's not the case. I am sure there is a good reason, but what is it?
In any case, IMHO there should be a comment here that briefly explains these things.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Changed text.
lib/fitfree.gi
Outdated
List(GeneratorsOfGroup(U),x->ImagesRepresentative(ffs.factorhom,x))); | ||
RUN_IN_GGMBI:=false; | ||
fi; | ||
if rest=fail then Error("can't build homomorphism"); fi; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
So when would this ever fail? Can we perhaps give a slightly more helpful error message?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I've replaced the error with an assertion.
Codecov Report
@@ Coverage Diff @@
## master #3103 +/- ##
==========================================
+ Coverage 83.56% 83.57% +0.01%
==========================================
Files 686 686
Lines 336759 336772 +13
==========================================
+ Hits 281396 281466 +70
+ Misses 55363 55306 -57
|
Codecov Report
@@ Coverage Diff @@
## master #3103 +/- ##
==========================================
+ Coverage 83.55% 83.57% +0.01%
==========================================
Files 686 686
Lines 336665 336777 +112
==========================================
+ Hits 281317 281460 +143
+ Misses 55348 55317 -31
|
@laurentbartholdi |
1ae7729
to
1c23d1c
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I didn't pull the changes, but reviewing the code it seems all good.
for the epimorphism from a group to its fitting free factor, as no method for constructing this restriction might exist. (E.g. in the case of matrix groups.)
This fixes #3097, #3100 and a nuisance error when working with matrix groups under the fitting free setup (avoid
RestrictedMapping
which runs into trouble).