DEP: improve deprecation of remaining private modules named without underscore prefix #19585
Conversation
90cb371 to
7ed4e61
Compare
7ed4e61 to
0ac76b6
Compare
|
Down to 2 failures on one of the CI runs, with |
| ('scipy.linalg.basic', None), | ||
| ('scipy.linalg.decomp', None), | ||
| ('scipy.linalg.decomp_cholesky', None), | ||
| ('scipy.linalg.decomp_lu', None), | ||
| ('scipy.linalg.decomp_qr', None), | ||
| ('scipy.linalg.decomp_schur', None), | ||
| ('scipy.linalg.decomp_svd', None), | ||
| ('scipy.linalg.flinalg', None), |
There was a problem hiding this comment.
originally I didn't touch flinalg because, after the next release, there would be nothing left in it. In which case it seems logical at that point just to remove it however officially it is slated for 2.0.0 so there is an awkward contradiction. Any thoughts?
There was a problem hiding this comment.
@h-vetinari do you have an opinion here?
There was a problem hiding this comment.
@h-vetinari friendly ping here as we are a few days away from branching
@j-bowhay is there any harm in making the change here even if we do end up removing flinalg?
There was a problem hiding this comment.
Plan on merging tomorrow if there are no further comments
I think I've fixed them (see if CI agrees) - |
[docs only]
|
CI is all green, your call on |
[skip ci]
Co-authored-by: Jake Bowhay <60778417+j-bowhay@users.noreply.github.com>
[skip ci]
Reference issue
Closes gh-18279
What does this implement/fix?
Clears up all of the remaining instances for this issue, namely in
io,misc,signalandfftpack. Also, one inlinalg(which I assume just slipped through?)Additional information
The only tricky module was
scipy.misc.doccerwhich has been moved toscipy._lib.doccer. IIUC we can't use the helper for that, so I copy-pasted and modified appropriately.I might have missed something, let's see what CI thinks.