Make math-mode=fast no op, it's too dangerous #41638
Conversation
Fixes: JuliaLang#25028 Competing PR to my other JuliaLang#41607
|
I don't think it is worth opening a bunch of different variation PRs about this. The difficulty here isn't really to just remove the relevant lines, it is to come to a conclusion. When that has been done, the PR can be updated to whatever that conclusion is. Having multiple versions of a PR just risks splitting the discussion making it harder to follow the conversation. |
|
Ok, let's keep the discussion at the other PR. I was trying to address your concern that that other PR was a breaking change (while only for the command-line option, which was the point of that PR). |
|
@JeffBezanson, if I understood you correctly at the other PR, then according to triage, this one should be merged (or possibly without the "added "broken" option"). |
|
Bump. It seemed like Jeff approved the way this PR works in #41607 (comment) |
|
As I read it, the PR currently deletes the option rather than make it a no-op. |
|
No, it keeps the --math-mode=fast, just disables it, while, yes, dropping (only for now, not other docs) from --help. It also adds undocumented "broken" option. |
|
Since this was reopened and is being (re)considered, I'm not sure what the "14 failing" are (false alarms?): My code in the PR seems trivial enough that I think it's correct, so maybe it can just be merged for 1.8? If people regret it could be undone in e.g. 1.9 or even in 1.8.x. |
|
we're past the point of merging this into 1.8. I think the main thing this needs is a rebase on the latest master. |
|
From triage: OK. See comments. Also need to update the manual. |
|
(and a rebase) |
|
a) "(and a rebase)", I sort of know it means outside of GitHub, and even though GitHub also has a page on it, it seems impossible here in the web-only (where I made the edit/PR...): https://docs.github.com/en/get-started/using-git/about-git-rebase When asked to do a rebase previously, I want down a rabbit-hole, made some mistakes, and had to start over. Can anyone else do it for me on this PR number? Outdated (not sure why it went away...): b) The code seems fine and the timeout(s) seem like false alarms: |
The behavior was changed and the CLI doc was removed in JuliaLang#41638, though we current still allow users to selectively use the `@fastmath` macro. This adds back the CLI docs and makes a minor clarification about behavior.
The behavior was changed and the CLI doc was removed in JuliaLang#41638, though we current still allow users to selectively use the `@fastmath` macro. This adds back the CLI docs and makes a minor clarification about behavior.
The behavior was changed and the CLI doc was removed in JuliaLang#41638, though we current still allow users to selectively use the `@fastmath` macro. This adds back the CLI docs and makes a minor clarification about behavior. Co-authored-by: Matt Bauman <[email protected]>
The behavior was changed and the CLI doc was removed in JuliaLang#41638, though we current still allow users to selectively use the `@fastmath` macro. This adds back the CLI docs and makes a minor clarification about behavior. Co-authored-by: Matt Bauman <[email protected]>
The behavior was changed and the CLI doc was removed in #41638, though we current still allow users to selectively use the `@fastmath` macro. This adds back the CLI docs and makes a minor clarification about behavior. Co-authored-by: Matt Bauman <[email protected]>
Fixes: #25028
Competing PR to my other #41607