Skip to content

Forward the new juliati options through the reusable workflow - #7

Draft
davidanthoff wants to merge 1 commit into
mainfrom
forward-inputs
Draft

Forward the new juliati options through the reusable workflow#7
davidanthoff wants to merge 1 commit into
mainfrom
forward-inputs

Conversation

@davidanthoff

Copy link
Copy Markdown
Member

julia-actions/julia-run-testitems gained seven inputs (julia-actions/julia-run-testitems#2), but this workflow forwarded only testitem-timeout, filter, env and github_job_prep_script — so anyone using the shared workflow could not reach any of them.

Found while documenting the new options: the docs agent correctly declined to write them up for the shared workflow, because doing so would have documented inputs that silently did nothing.

Forwarded

junit-path, coverage-lcov-path, output-mode, threads, gc-between-testitems, memory-threshold, schedule.

Added as plain inputs, not per-trigger

The draft-pr- / pr- / main- / manual-trigger- override pattern exists for varying how much gets tested depending on the trigger — matrix breadth, the test filter, the timeout. These seven describe how the test processes behave, and do not sensibly differ between a draft PR and main. Adding 35 more input declarations for a need that has not appeared would be noise; the pattern is there to extend if it ever does.

Verification

The workflow parses, all seven inputs are declared, and all seven reach the julia-run-testitems step — 13 inputs forwarded in total, checked programmatically rather than by eye. README documents each one and notes why they have no per-trigger variants.

🤖 Generated with Claude Code

julia-run-testitems gained junit-path, coverage-lcov-path, output-mode,
threads, gc-between-testitems, memory-threshold and schedule, but this workflow
forwarded only testitem-timeout, filter, env and github_job_prep_script — so
anyone using the shared workflow could not reach any of them.

Added as plain inputs rather than getting the per-trigger override treatment.
Those overrides exist for varying how much gets tested on a draft PR versus
main; these describe how the test processes behave and do not sensibly differ
by trigger.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant