py3: phase out lexico cmp in real_mpfi

[fchapoton]

Currently cmp(a,b) for two real-interval field elements performs a lexicographic comparison. And rich comparison has a different semantic.

We rename _cmp_ to lexico_cmp to put it outside the comparison framework. This means that cmp will not work anymore. The documentation is modified accordingly, to warn users not to use cmp at all on these objects.

Helpful for the major ticket #22297

Depends on #18303

CC: @tscrim @jdemeyer @a-andre @dkrenn @cheuberg @behackl

Component: python3

Author: Frédéric Chapoton

Branch/Commit: ea97bfc

Reviewer: Daniel Krenn

Issue created by migration from https://trac.sagemath.org/ticket/22907

[fchapoton]

Branch: u/chapoton/22907

[fchapoton]

Commit: 20cd81e

[dkrenn]

Replying to @fchapoton:

Currently cmp(a,b) for two real-interval field elements performs a lexicographic comparison. And rich comparison has a different semantic.

We rename _cmp_ to lexico_cmp to put it outside the comparison framework. This means that cmp will not work anymore. The documentation is modified accordingly, to warn users not to use cmp at all on these objects.

Do we need lexico_cmp at all or could we simply delete it?

[fchapoton]

comment:3

well, lexico_cmp is useful to check that pickling works. And if somebody did use cmp, it could serve as a replacement

[dkrenn]

comment:4

Replying to @fchapoton:

well, lexico_cmp is useful to check that pickling works. And if somebody did use cmp, it could serve as a replacement

Ok, fine for me.

[dkrenn]

comment:5

Replying to @dkrenn:

Replying to @fchapoton:

well, lexico_cmp is useful to check that pickling works. And if somebody did use cmp, it could serve as a replacement

As cmp is removed, do we need a deprecation warning for this?

[dkrenn]

Reviewer: Daniel Krenn

[dkrenn]

comment:7

Apart from the question above and modulo a successful run of a patchbot, this patch looks good.

[fchapoton]

comment:8

Hmm, trying to introduce a deprecation seems to uncover some problems with QQbar. Investigating, maybe in relation with #18303

[fchapoton]

comment:9

en experimental branch with deprecation is available as "u/chapoton/experiment-22907"

[dkrenn]

comment:11

Replying to @fchapoton:

en experimental branch with deprecation is available as "u/chapoton/experiment-22907"

This looks fine for me.

[fchapoton]

Changed branch from u/chapoton/22907 to u/chapoton/experiment-22907

[fchapoton]

Dependencies: #18303

[fchapoton]

Changed commit from 20cd81e to ea97bfc

[fchapoton]

comment:12

let us wait for #18303 and then check than nothing else is broken by deprecation

---

New commits:

ea97bfc

py3 deprecation of call to cmp on RIF elements

[fchapoton]

comment:13

looks good, bot is morally green

[fchapoton]

comment:14

back to needs review, please double check

[vbraun]

Changed branch from u/chapoton/experiment-22907 to ea97bfc