search

edgarcosta / endomorphisms

Rigorous computation of the endomorphism ring of a Jacobian
GNU General Public License v2.0
10 stars 8 forks source link

Sage doesn't load .magmarc at startup #19

Closed edgarcosta closed 5 years ago

edgarcosta commented 5 years ago

This week @kedlaya was trying to use our package through sage, and the magma was behaving as if Molin--Neurohr's code was not installed. After a while, we realized that magma hadn't loaded .magmarc file, and thus the mismatch in expectations.

@JRSijsling, have you ever noticed this before? Should we add a note?

JRSijsling commented 5 years ago

A first, kind of unhelpful, remark is that by now you do not have to use it through Sage any longer. See examples/Buttons.m.

I though that this was taken care of in __init__.py. So this is strange. You can add a line magma.load("~/.magmarc") to ~/.sage/init.sage. But after that I still get errors on toby. This sucks.

On the other hand, if your own error is about the computation of period matrices plane quartics, that is as it should be, because that functionality is not in Molin--Neurohr. I will write Kiran on how to include this.

JRSijsling commented 5 years ago

I kicked out all my config files and started from scratch, reinstalling the package and adding that line to ~/.sage/init.sage. After that, for some reason everything worked again.

The magmarc file should contain references to the spec of hcperiods as well as the spec of endomorphisms itself. That triggers everything. Shall we add a line to this effect to the README?

(Why what I did made any difference to SageMath is another question. It should not have, to be honest.)

JRSijsling commented 5 years ago

Just shamelessly pushed an improved README.md. If you like it, then this can be closed, if not, then we talk further.

edgarcosta commented 5 years ago

Thanks!

I was going to ask what you thought was the best solution, but I agree with that one :)