Problem with old submodule

[davidloeffler]

Weird things happen if you try to compute the old submodule of a Gamma1 modular forms space. E.g.

sage: CuspForms(Gamma1(11), 2).old_submodule() # crashes

This turns out to be because:

sage: ModularForms(Gamma1(11), 2).hecke_module_of_level(1).hecke_module_of_level(11)
Modular Forms space of dimension 2 for Congruence Subgroup Gamma0(11) of weight 2 over Rational Field

so the degeneracy level-raising map is getting computed for the wrong space!

Prerequisite for #11601. (Note that #11601 depends on a bunch of other stuff as well, but this ticket is independent of any of the other patches in the series.)

---

Apply trac_10453_gamma1_old_submodule.patch​ to the Sage library.

[davidloeffler]

patch against 4.6.1.alpha3 + #8716

[davidloeffler]

Here's a patch. It requires #8716 (but is independent of either #10450 or #10452). The fix was to make the various degeneracy matrix commands pass around an explicit target space, rather than just a target level, as the argument.

[davidloeffler]

Depends on #8716

[johanbosman]

The code looks sound and is well documented. All long doctests pass. I played around with it and couldn't produce any false results. I did get the following while applying the patch:

applying trac_10453_gamma1_old_submodule.patch
patching file sage/modular/hecke/ambient_module.py
Hunk #10 succeeded at 883 with fuzz 1 (offset 6 lines).
now at: trac_10453_gamma1_old_submodule.patch

but according to ​this discussion, fuzz 1 is okay. Hence: positive review. ;).

[leif]

Replying to johanbosman:

I did get the following while applying the patch:

applying trac_10453_gamma1_old_submodule.patch
patching file sage/modular/hecke/ambient_module.py
Hunk #10 succeeded at 883 with fuzz 1 (offset 6 lines).
now at: trac_10453_gamma1_old_submodule.patch

but according to ​this discussion, fuzz 1 is okay. Hence: positive review. ;).

In this case, you could re-export the patch and upload the rebased version.

IMHO one should inspect any fuzz anyway.

[davidloeffler]

Replying to leif:

IMHO one should inspect any fuzz anyway.

I have inspected the fuzz, which is caused by patch #10664 having gone in since this was written. It is not a problem; the issue fixed by #10664 is completely orthogonal to this one.

[leif]

Replying to davidloeffler:

I have inspected the fuzz, which is caused by patch #10664 having gone in since this was written. It is not a problem; the issue fixed by #10664 is completely orthogonal to this one.

Thanks, I hadn't [yet].