Deprecate sage.misc.misc.uniq

[jdemeyer]

uniq(X) is just a shorthand for sorted(set(X)): I don't think that we need a separate function for that. More seriously, the sorting is problematic since arbitrary sets may not be sortable in general, especially in Python 3 or after applying #22029.

Depends on #26933

CC: @tscrim

Component: misc

Author: Jeroen Demeyer

Branch/Commit: b4f67ca

Reviewer: Martin Rubey, Travis Scrimshaw

Issue created by migration from https://trac.sagemath.org/ticket/27014

[jdemeyer]

Dependencies: #26933

[jdemeyer]

Branch: u/jdemeyer/deprecate_sage_misc_misc_uniq

[jdemeyer]

Commit: 17ee0d0

[jdemeyer]

New commits:

0d45f46	Do not sort in Subsets_s
dba4c44	Make _stable_uniq private
17ee0d0	Deprecate uniq()

[mantepse]

Reviewer: Martin Rubey

[mantepse]

comment:4

I am all for it! (I ran the patchbot and looked at the code)

[tscrim]

comment:5

I am not so sure about actually deprecating it. The resulting behavior after _stable_uniq -> uniq is done will be the same is what it currently is. Plus the message seems like it will not be there after the deprecation, and I believe this is a useful function from the (non-expert) user perspective. So I am somewhat inclined to also just do the _stable_uniq -> uniq now. Thoughts?

[jdemeyer]

comment:6

Replying to @tscrim:

The resulting behavior after _stable_uniq -> uniq is done will be the same is what it currently is.

No. The existing uniq() sorts the list of objects, while _stable_uniq() keeps the existing order.

[tscrim]

comment:7

Replying to @jdemeyer:

Replying to @tscrim:

The resulting behavior after _stable_uniq -> uniq is done will be the same is what it currently is.

No. The existing uniq() sorts the list of objects, while _stable_uniq() keeps the existing order.

Ah, right, and the sorting of the output is documented. However, given the (eventual) goal, I think we should replace uniq with _stable_uniq and issue a deprecation warning about the sorting behavior change. Or at least in the current deprecation message say something like

the output of uniq(X) being sorted is deprecated; use sorted(set(X)) instead
if you want the output sorted

The current deprecation message seems to indicate that the function will be removed altogether.

[mantepse]

Changed reviewer from Martin Rubey to none

[mantepse]

comment:8

Why not simply introduce the function stable_uniq, and remove uniq after deprecation?

Is there a good reason to call what's now _stable_uniq in a year's time uniq?

[tscrim]

comment:9

Replying to @mantepse:

Why not simply introduce the function stable_uniq, and remove uniq after deprecation?

stable_uniq does not sound discoverable or natural IMO. Also for those people who only want the uniq behavior, it means they would not have the change their code.

Is there a good reason to call what's now _stable_uniq in a year's time uniq?

Someone might be relying on this sorting behavior in the wild, so we need to deprecate it.

[mantepse]

comment:10

Replying to @tscrim:

Replying to @mantepse:

Why not simply introduce the function stable_uniq, and remove uniq after deprecation?

stable_uniq does not sound discoverable or natural IMO.

OK, although I wouldn't be looking for uniq either. In fact, I was surprised by it's documentation, because the well known unix utility does something else.

I was looking for remove_duplicates once, before it occurred to me that I want set :-)

Someone might be relying on this sorting behavior in the wild, so we need to deprecate it.

I think that we all agree on that.

[sagetrac-git]

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

b4f67ca

Deprecate uniq()

[sagetrac-git]

Changed commit from 17ee0d0 to b4f67ca

[jdemeyer]

comment:12

Reworded deprecation message as Travis suggested.

[tscrim]

comment:13

Thank you.

[tscrim]

Reviewer: Martin Rubey, Travis Scrimshaw

[embray]

comment:14

Retarging tickets optimistically to the next milestone. If you are responsible for this ticket (either its reporter or owner) and don't believe you are likely to complete this ticket before the next release (8.7) please retarget this ticket's milestone to sage-pending or sage-wishlist.

[vbraun]

Changed branch from u/jdemeyer/deprecate_sage_misc_misc_uniq to b4f67ca