Faster listing of number field homsets

[sagetrac-fwclarke]

The patch speeds up the calculation of homomorphisms between number fields. This is mainly achieved by avoiding unnecessary checking. Doctests are also updated.

CC: @lftabera @rharron @sagetrac-tfeulner @simon-king-jena

Component: number fields

Author: Francis Clarke

Branch/Commit: e2ee8a3

Reviewer: Jeroen Demeyer

Issue created by migration from https://trac.sagemath.org/ticket/10843

[sagetrac-fwclarke]

Attachment: trac_10843.patch.gz

rebased for 5.9

[sagetrac-fwclarke]

comment:3

The patch has been rebased for Sage-5.9.

[videlec]

comment:8

Hi,

Looks nice.

I uploaded your patch in a git branch and added three commits above:

• the first one to remove the trailing whitespaces
• the second one to fits with Sage standard for the docs and make the code more compatible with Python 3
• the third to have a better caching

Please, tell me what you think.

Vincent

---

New commits:

20f21a6	#10843 faster lising of number field homsets
4040bfd	trac #10843: remove trailing whitespaces
a2abefa	trac #10843: documentation, indentation
51569d9	trac #10843: improved caching

[videlec]

Commit: 51569d9

[videlec]

Branch: public/10843

[sagetrac-fwclarke]

comment:9

Thanks for this.  It all seems to be working perfectly.

[jdemeyer]

comment:11

Why the separate branches for D absolute/relative? The following works in both cases:

roots = D.polynomial().roots(ring=C, multiplicities=False)

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

946525b	trac #10843: remove separate absolute/relative branches
e2ee8a3	trac #10843: add check keywords and an associated doctest

[sagetrac-git]

Changed commit from 51569d9 to e2ee8a3

[sagetrac-fwclarke]

comment:13

Replying to @jdemeyer:

Why the separate branches for D absolute/relative?

Assuming that you meant C rather than D, I've changed the code as suggested. I don't know why this was done separately. Perhaps the simpler approach didn't work when this ticket began four years ago.

I've made one other change which should have been made before. It solves a problem which arose in http://ask.sagemath.org/question/24173/homomorphisms-for-relative-number-fields/.

[jdemeyer]

Reviewer: Jeroen Demeyer

[vbraun]

Changed branch from public/10843 to e2ee8a3