Fix global import of sage.repl.rich_output.display_manager.get_display_manager

[mkoeppe]

This function leaks in via sage.interfaces.all.

Since it is being used in doctest, we import it properly through sage.repl.all.

CC: @egourgoulhon

Component: user interface

Author: Matthias Koeppe

Branch/Commit: d6cf9f6

Reviewer: Kwankyu Lee

Issue created by migration from https://trac.sagemath.org/ticket/33016

[mkoeppe]

Branch: u/mkoeppe/deprecate_global_import_of_sage_repl_rich_output_display_manager_get_display_manager

[mkoeppe]

Commit: 5c06046

[mkoeppe]

New commits:

5c06046

src/sage/interfaces/all.py: Deprecate re-export of get_display_manager

[mkoeppe]

Author: Matthias Koeppe

[kwankyu]

comment:3

I have used this function only for testing. But it seems people use get_display_manager to set display preferences in jupyter. It would be inconvenient if it should be imported first from an obscure place.

[sagetrac-git]

Changed commit from 5c06046 to d6cf9f6

[sagetrac-git]

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

738b703	src/sage/repl/all.py: Also import get_display_manager
d6cf9f6	src/sage/interfaces/all.py: Do not re-export get_display_manager; do not fail if get_display_manager cannot be imported

[mkoeppe]

Description changed:

--- 
+++ 
@@ -1,2 +1,4 @@
 This function leaks in via `sage.interfaces.all`.
 
+Since it is being used in doctest, we import it properly through `sage.repl.all`.
+

[mkoeppe]

comment:6

Here's a new version

[kwankyu]

comment:7

Perhaps we still need a deprecation warning about the change of import location?

[kwankyu]

Reviewer: Kwankyu Lee

[mkoeppe]

comment:9

Replying to @kwankyu:

Perhaps we still need a deprecation warning about the change of import location?

I'd say we don't, according to the leak theory

[kwankyu]

comment:10

Replying to @mkoeppe:

Replying to @kwankyu:

Perhaps we still need a deprecation warning about the change of import location?

I'd say we don't, according to the leak theory

Sorry, but what is the leak theory?

One argument for no deprecation warning might be that we already made an announcement that no library code should import from all modules. Is this what you meant?

[kwankyu]

comment:12

Replying to @kwankyu:

Perhaps we still need a deprecation warning about the change of import location?

I still don't know the leak theory. But I now agree that we don't need a deprecation warning.