py3: change doc for comparison in element classes

[fchapoton]

with suggestion to use _richcmp_

may help to fix one of the doctest failures in #22297

CC: @tscrim @jdemeyer @a-andre

Component: python3

Author: Frédéric Chapoton

Branch/Commit: df4c150

Reviewer: Jeroen Demeyer

Issue created by migration from https://trac.sagemath.org/ticket/22890

[fchapoton]

Commit: 89d245d

[fchapoton]

Branch: u/chapoton/22890

[fchapoton]

New commits:

89d245d

py3: change doc for comparison in element classes

[jdemeyer]

comment:2

Why is _richcmp_ recommended?

[fchapoton]

comment:3

Because most of the time, writing either a _cmp_ or a __cmp__ implies a call to cmp..

By the way, Jeroen, could you have a look at #22881, please ?

and maybe also to #22875, if you feel like..

[fchapoton]

comment:5

Jeroen, could you please say what you want here more precisely ?

[fchapoton]

comment:6

ping ?

[jdemeyer]

comment:7

Replying to @fchapoton:

Because most of the time, writing either a _cmp_ or a __cmp__ implies a call to cmp..

I don't really follow this reasoning. It's obviously true that cmp() should be avoided because that doesn't work in Python 3. But just because some (or even most) implementations of _cmp_() call cmp(), that doesn't mean that _cmp_() should be avoided in general.

If you really want to say that cmp() should be avoided, then just say that instead of something else.

[sagetrac-git]

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

da00b47	Merge branch 'u/chapoton/22890' in 8.0.b6
df4c150	trac 22890 better doc

[sagetrac-git]

Changed commit from 89d245d to df4c150

[fchapoton]

comment:9

Better like that ?

[jdemeyer]

Reviewer: Jeroen Demeyer

[vbraun]

Changed branch from u/chapoton/22890 to df4c150