Skip to content

Conversation

topolarity
Copy link
Member

Accidental regression from #57289. Due to the different specialization rules for Function types, the @noinline bodies were inferring very poorly treating once as a Function instead of the much more specific OncePerX{T,F}.

Adds a trimming test case so that we don't regress again.

Accidental regression from JuliaLang#57289, due to the different specialization
rules for `Function` types. The `@noinline` bodies were inferring very
poorly treating `once` as a `Function` instead of the much more
specific `OncePerX{T,F}`.
Just to have a slightly more interesting "Hello, world!" and make sure that
OncePerX trim-compatibility is covered in our tests somewhere.
@topolarity topolarity added trimming Issues with trimming functionality or PR's relevant to its performance/functionality backport 1.12 Change should be backported to release-1.12 labels Mar 6, 2025
@topolarity topolarity requested a review from vtjnash March 6, 2025 18:51
@topolarity topolarity merged commit c93dd2d into JuliaLang:master Mar 7, 2025
9 checks passed
KristofferC pushed a commit that referenced this pull request Mar 20, 2025
Accidental regression from
#57289. Due to the different
specialization rules for `Function` types, the `@noinline` bodies were
inferring very poorly treating `once` as a `Function` instead of the
much more specific `OncePerX{T,F}`.

Adds a trimming test case so that we don't regress again.

(cherry picked from commit c93dd2d)
@KristofferC KristofferC removed the backport 1.12 Change should be backported to release-1.12 label Mar 24, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

trimming Issues with trimming functionality or PR's relevant to its performance/functionality

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants