remove 0.6 deprecations #25312
remove 0.6 deprecations #25312
Conversation
base/client.jl
Outdated
| if ex isa Symbol && all(equalto('_'), string(ex)) | ||
| # remove with 0.7 deprecation | ||
| Meta.lower(Main, ex) # to get possible warning about using _ as an rvalue | ||
| end |
There was a problem hiding this comment.
I think this one needs to stay.
There was a problem hiding this comment.
Good catch! Fixed. Thanks! :)
|
Dates stdlib has a deprecation file. |
Good catch! The IterativeEigensolvers and SuiteSparse deprecations are for 0.7 (I should add a note), and going by the note at the top of the file I think the Dates deprecations are also for 0.7. Please correct me if I am incorrect! |
|
Right, I forgot I removed the 0.6 deprecations for Dates. |
|
There are also a few in the front end; I'll try to get to those. |
|
Ah yes, of course --- this will require fixing Documenter. |
|
Indeed so :). I plan to restore the problematic deprecations here and address them separately. (Apologies for the absence of motion on this front! Focusing on #25364.) |
|
Rebased and fixed the doc build failure. What say you now, CI? |
|
The AV failures appear unrelated (one timeout after successfully completing the main test suite, one remote exception deserialization failure in the Perhaps someone could give this pull request a final sweep and merge? Absent that and objections, I plan to merge this pull request Sunday evening PT at the earliest and pick off the tricker remaining removals in downstream work (unless someone happily beats me to it! :) ). Best! |
|
Rebased. Absent objections or requests for time, I plan to merge these changes once CI clears again. Best! |
|
The Travis i686 failure appears unrelated (one libgit2 remote exception deserialization failure) as do the AV failures (one timeout without apparent errors prior, one immediate failure due to a download/configuration issue). Thanks all! :) |
| return Matrix(uninitialized, 0, 0) | ||
| end | ||
|
|
||
| # TODO: remove warning for using `_` in parse_input_line in base/client.jl |
There was a problem hiding this comment.
this todo wasn't done, should be put back until it is
There was a problem hiding this comment.
one underscore was deprecated in 0.6. all-underscores in 0.7.
There was a problem hiding this comment.
You are suggesting to modify the linked check such that only multiple-underscore instances trigger a warning?
There was a problem hiding this comment.
Not urgent though, since we haven't yet reclaimed _ for anything else.
There was a problem hiding this comment.
Delaying the warning-into-error removal of the single underscore case from 0.7 to 1.0 is probably fine. But the TODO about client.jl should at least be put back in the 0.7 section so it doesn't get forgotten and left as only a warning in 1.0.
There was a problem hiding this comment.
I assumed existence of a relevant TODO in the 0.7 section. If no such TODO exists, then indeed, moving this TODO to the 0.7 section sounds wise. I'll check and whip together a pull request if necessary. Thanks! :)
This pull request removes (almost all) 0.6 deprecations. (Three or so deprecations remain; I hope to remove those tomorrow but might need a pointer or two.) Let's see what CI says. Best!