Temporarily pin aesara to 2.1.0 to fix CI

[michaelosthege]

The latest Aesara release caused a test to break.

The latest CI run on main was with Aesara 2.1.0, but I ran into this problem while working on #4867.

Since I don't have the leisure to diagnose/fix, this PR pins the Aesara version.
If someone wants to fix it, feel free to push this branch!
Otherwise I think we should merge ASAP to prevent other PRs from stalling on unrelated CI failures.

[codecov]

Codecov Report

Merging #4868 (32dd28f) into main (25da9e2) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##             main    #4868   +/-   ##
=======================================
  Coverage   72.82%   72.82%           
=======================================
  Files          86       86           
  Lines       13795    13795           
=======================================
  Hits        10046    10046           
  Misses       3749     3749           

[brandonwillard]

The latest CI run on main was with Aesara 2.1.0, but I ran into this problem while working on #4867.

Since I don't have the leisure to diagnose/fix, this PR pins the Aesara version.
If someone wants to fix it, feel free to push this branch!
Otherwise I think we should merge ASAP to prevent other PRs from stalling on unrelated CI failures.

Please clearly identify the problem before asking others to merge something ASAP.

First, is there a specific test that's failing in a way that's identifiably related to changes in Aesara? Can you at least link to one of the failures in #4867 that you think is relevant?

Right now you're asking others to sift through your PR's CI output, yet you've likely done this yourself in order to confirm that your changes aren't the cause of the/some failures, no?

[brandonwillard]

Superseded by #4869.