Do not compare types in coercion model richcmp()

[embray]

Though small, the fix to richcmp is significant and fixes a lot of bugs.

Component: coercion

Keywords: structure coerce

Author: Erik Bray, Jeroen Demeyer

Branch/Commit: de73aad

Reviewer: Jeroen Demeyer, Frédéric Chapoton

Issue created by migration from https://trac.sagemath.org/ticket/24968

[jdemeyer]

comment:2

I don't like this change. It will make it harder to port Sage to Python 3 by hiding errors in an try/except block.

It would be much better to just not compare types in the first place.

[jdemeyer]

Changed branch from u/embray/python3/sage-structure-coerce to u/jdemeyer/python3/sage-structure-coerce

[embray]

comment:4

If you say so--does this actually break anything?

FWIW I have personally written code before that implements comparisons between metaclasses (!). I don't understand the implications when it comes to Sage though.

---

New commits:

7654b19

Don't compare types at all

[embray]

Changed commit from 082ad7b to 7654b19

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:

de73aad

Don't compare types at all

[sagetrac-git]

Changed commit from 7654b19 to de73aad

[jdemeyer]

Changed author from Erik Bray to Erik Bray, Jeroen Demeyer

[jdemeyer]

comment:6

Let's see what the patchbot says.

[embray]

comment:7

Looks like the comparison between two objects' types dates as far back to f112feb0. It's probably not relevant. Theoretically it could be useful, but probably not for any currently existing code.

[jdemeyer]

comment:8

It seems to work on the patchbot. Erik, if you agree, you can set this to positive review.

[jdemeyer]

Reviewer: Jeroen Demeyer

[fchapoton]

Changed reviewer from Jeroen Demeyer to Jeroen Demeyer, Frédéric Chapoton

[fchapoton]

comment:10

Let us move on.

[vbraun]

Changed branch from u/jdemeyer/python3/sage-structure-coerce to de73aad