Opened 14 years ago

Closed 13 years ago

Last modified 13 years ago

#5969 closed enhancement (fixed)

implement computation of rational cuspidal subgroups of modular abelian varieties

Reported by: was Owned by: was
Priority: major Milestone: sage-4.4.1
Component: number theory Keywords:
Cc: boothby Merged in: sage-4.4.1.alpha2
Authors: William Stein Reviewers: Alex Ghitza
Report Upstream: N/A Work issues:
Branch: Commit:
Dependencies: Stopgaps:

Status badges

Description

This will depend on #5882.

Attachments (5)

trac_5969-part1.patch (17.8 KB) - added by was 14 years ago.
trac_5969-part2.patch (2.9 KB) - added by was 13 years ago.
trac-5969-part3.patch (4.8 KB) - added by was 13 years ago.
trac_5969-part4.patch (9.6 KB) - added by was 13 years ago.
trac_5969-part5.patch (1.3 KB) - added by was 13 years ago.

Download all attachments as: .zip

Change History (13)

Changed 14 years ago by was

Attachment: trac_5969-part1.patch added

Changed 13 years ago by was

Attachment: trac_5969-part2.patch added

Changed 13 years ago by was

Attachment: trac-5969-part3.patch added

comment:1 Changed 13 years ago by was

Cc: boothby added
Report Upstream: N/A

comment:2 Changed 13 years ago by was

Status: newneeds_review

Hi,

Note that trac-5969-part4.patch removes the abvarsub modular symbols functions for torsion, since I found that they are buggy and not finished. The same functionality is already available in the modular abelian varieties code anyways, so this is no real loss.

Changed 13 years ago by was

Attachment: trac_5969-part4.patch added

comment:3 Changed 13 years ago by was

I just checked that all four patches apply fine to sage-4.3.5 still with no rebasing necessary.

comment:4 Changed 13 years ago by AlexGhitza

Authors: William Stein
Reviewers: Alex Ghitza
Status: needs_reviewneeds_work

The "part2" patch changes some things in matrix/matrix_integer_dense.pyx, and that causes two doctest failures:

sage -t -long "devel/sage/sage/modules/fg_pid/fgp_module.py"
**********************************************************************
File "/mnt/usb1/scratch/ghitza/sage-4.3.5-sage.math.washington.edu-x86_64-Linux/devel/sage/sage/modules/fg_pid/fgp_module.py", line 1131:
    sage: phi = Q.hom([0,V.0,V.1]); phi
Expected:
    Morphism from module over Integer Ring with invariants (2, 0, 0) to module with invariants (0, 0, 0) that sends the generators to [(0, 0, 0), (0, 0, 1), (0, 1, 0)]
Got:
    Morphism from module over Integer Ring with invariants (2, 0, 0) to module with invariants (0, 0, 0) that sends the generators to [(0, 0, 0), (1, 0, 0), (0, 1, 0)]
**********************************************************************
File "/mnt/usb1/scratch/ghitza/sage-4.3.5-sage.math.washington.edu-x86_64-Linux/devel/sage/sage/modules/fg_pid/fgp_module.py", line 1139:
    sage: phi(Q.1)
Expected:
    (0, 0, 1)
Got:
    (1, 0, 0)
**********************************************************************

It was not obvious to me whether this was harmless or an actual problem.

The rest looks good, there are a couple of docstring fixes but I have a reviewer patch that can take care of them.

Changed 13 years ago by was

Attachment: trac_5969-part5.patch added

comment:5 Changed 13 years ago by was

Status: needs_workneeds_review

It turns out that part 2 fixes a *MAJOR* bug in SNF for matrices over ZZ in an edge case. The doctest in finitely generated modules was just wrong (ouch). I carefully checked through this with Robert Bradshaw, and posted a patch that updates the doctest.

comment:6 Changed 13 years ago by AlexGhitza

Status: needs_reviewpositive_review

This looks good to me, and passes tests.

Note that the part1 patch applies with some fuzz to sage-4.4.rc0, but it's fine.

comment:7 Changed 13 years ago by was

Merged in: 4.4.1.alpha2
Resolution: fixed
Status: positive_reviewclosed

comment:8 Changed 13 years ago by mvngu

Merged in: 4.4.1.alpha2sage-4.4.1.alpha2
Note: See TracTickets for help on using tickets.