Repository navigation
index refresh and test counts: a refresh beside another process waits while an index the home reads has no packages (mcpp#927); a suite is counted by its cases - #999
Merged
Conversation
… reads has no packages, judged as a reader judges it (mcpp#927)
…line, as the summary counts it
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 free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
mcpp#927 item 3 reported a process that ended with
no local package indexwhile another was refreshing the index. 2026.11.13.1 made such a process wait for the refresh it skipped when an index it declares is absent (design 2026-11-12 §18 P3), but that judgement differed from the reader's:<index>/pkgsis missing (index_status_for); the refresh judged a declared index's directory by its existence, so a directory the other process had made and not yet filled let it skip;xim,data/xim-pkgindex) is read too, and was not judged at all, so a reader beside a first fetch of it skipped.Changes:
index_presentis the one judgement of an index's presence:<index>/pkgs. The index status, the refresh marker, the freshness check and the refresh beside another process use it.every_index_present: the provider's own and each declared); otherwise it waits for the lock, as before.tests/unit/test_xlings.cpp(XlingsIndexLock): the test holds the index lock (a second holder in the same process is refused, so no sync is reached) and asserts that a refresh returns at once when every index is present, and waits until the lock is released when a declared index has no packages or the provider's own index is absent. Against the previous code the two waiting cases fail.docs/05-dependencies.md(both languages) states the behaviour; CHANGELOG[Unreleased]; the design record's §18.3 states the change and the check that holds it.Item 6 of the issue (a hint for
mcpp clean --stale) is in place since 2026.11.13.1 (P6): a build whose configuration opens a new directory warns with the size of the one beside it and the command that reclaims it (tests/e2e/cache/directories.cpp).A suite's count on the program's line
The release's ecosystem check ran mcpp-index's
mtestexample: its program reported8 caseson its line and the summary7 passed. The summary counts a suite run as a case by its cases (a case that declares cases is told by them); the program's line counted every reported case, the suite included.declares_casesis the one judgement; the program's line, the summary and the failures' index use it.e2e/test, protocol): a program of one suite of two cases and one case alone reads3 caseson its line andcases: 3 passedin the summary.