Uh oh!
There was an error while loading. Please reload this page.
Test debug-inline capture of a local function from a nested closure - #20299
Open
T-Gro wants to merge 3 commits into
Open
Test debug-inline capture of a local function from a nested closure#20299T-Gro wants to merge 3 commits into
T-Gro wants to merge 3 commits into
Conversation
…callsite Regression coverage for #20297. The reduced repro keeps the reported shape - a captured local function, a callsite inside a nested closure, and '>>' composition - and crashes with AccessViolationException on release/10.0.4xx, where <addEnum>__debug@N is emitted as a two-argument static method that does 'ldarg.0; ldfld tee' against the string parameter. It passes once #20089 is present. SRTP 30-38 all call the inline function directly from the enclosing method, so none of them cover a callsite inside a closure. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
Contributor
|
This was referenced Aug 19, 2026
auduchinok
approved these changes
Aug 19, 2026
Comment on lines
+1
to
+6
| let f () = | ||
| let tee g (x: int) = g x; x | ||
| let addName (key: string) (v: string) = tee (fun x -> printfn "%s=%s" key v) | ||
| let inline addEnum (key: string) value = tee (fun x -> printfn "%s=%d" key (x + int value)) | ||
| let pipeline v = addName "a" "n" >> addEnum "b" v | ||
| pipeline 1uy 41 |
Member
There was a problem hiding this comment.
I'd personally try to minimize the test, so the dumps are easier to follow if needed, but otherwise it's fine.
MemberAuthor
There was a problem hiding this comment.
Minimized the repro in ac6ae82: dropped the printfn formatting and the extra addName closure, keeping only the essential ingredients (inline function whose body is a lambda capturing the outer local tee, used first-class inside the >> nested closure) plus the int value/ExplicitDynamic SRTP aspect. The baseline is now 108 lines (was 166) and still shows <addEnum>__debug@5 receiving the captured tee through a closure field.
T-Gro
enabled auto-merge (squash)
August 20, 2026 11:04
Address review feedback: strip printfn and addName noise from the repro so the emitted-IL baseline is easier to follow. The minimized source keeps all essential ingredients (inline function whose body is a lambda capturing an outer local 'tee', used first-class inside a nested closure via >>) and the SRTP 'int value'/ExplicitDynamic aspect, so <addEnum>__debug@N still reaches the captured local through a closure field. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…-inline-captured-locals-t
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for freeto join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Regression test only — #20297 is already fixed on
mainby #20089.Existing coverage applies the
inlinefunction directly from the enclosing method. This test applies it inside a nested closure, so the captured local reaches the synthesized<name>__debug@Nmethod through a closure field. Before #20089 that emittedldarg.0; ldfldagainst astringparameter and crashed withAccessViolationException.release/10.0.4xx(SDK 10.0.400) has #19548 but not #20089, and is still affected.