Repository navigation
Conversation
d10c
changed the title
d10c
changed the title
d10c
force-pushed
the
d10c/multi-query-ql-reconciliation
branch
from
November 4, 2026 14:00
5b1df91 to
3baabf9
Compare
Read cohort membership through query metadata and derive producer sets from comments re-extracted between serialized query passes. Render full-cohort atoms bare, qualify partial atoms, and normalize existing qualified sets while preserving foreign and named-column expectations. Add a focused Go-pack fixture for kind-less qualified matching and preservation behavior. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: d6e9cf9f-9b2e-487b-b28d-a79029d09ea8
When a bare atom becomes partial, render the surviving sibling producers with query-ID qualifiers instead of deleting their expectation. Carry forward only tags the current query understands, accept only non-empty single-ID qualifiers as cohort syntax, and leave default-column comma groups or whitespace-bearing qualifiers untouched because they cannot express one producer unambiguously. Treat a complete qualified producer set as equivalent to its collapsed bare atom when deciding whether a comment needs an edit. This preserves useful query provenance on otherwise-correct comments; comments requiring another repair still use canonical bare rendering. Extend the Go fixtures with bare-to-partial, ignored-tag, complete-qualified-set, and unsupported multi-ID mutation cases, then regenerate their expected output. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: d6e9cf9f-9b2e-487b-b28d-a79029d09ea8
Leave unsupported qualified syntax untouched, collapse sibling qualifications only when the rewritten comment emits the covering bare atom, and carry sibling ownership in the original expectation column. Cover invalid qualifier forms and default, MISSING, and SPURIOUS reconciliation through the Go inline-expectation fixtures. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 1adfa307-c48e-43f5-94e3-8286e4a8c083
d10c
force-pushed
the
d10c/multi-query-ql-reconciliation
branch
from
November 12, 2026 16:03
3baabf9 to
0a03bef
Compare
d10c
marked this pull request as ready for review
November 12, 2026 16:08
This branch has not been deployed
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.
What problem does it solve?
Inline-expectation learning needs to reconcile annotations owned by several
.qlreftests that share source files.Why is the problem a problem?
A selected query sees only its own results. Without complete cohort ownership, learning can remove another query's qualification, collapse a shared annotation too early, or rewrite an annotation that belongs to an unselected sibling.
How does the solution work?
The postprocessor accepts repeated
cohortMembermetadata, models each qualified atom as one producer, and renders the annotation from the complete surviving producer set. Bare atoms represent full-cohort ownership. Invalid producer qualifiers, including whitespace-bearing multi-ID forms, are rejected rather than guessed. Ignored annotations and unrelated complete qualified sets are preserved.Tagged
.qlrefmembership is used because it exposes complete directory ownership without compiling every unselected sibling.Why does the solution solve the problem?
Each query updates only the producer facts it owns while the renderer retains facts owned by the rest of the cohort. The complete membership list provides the information needed to decide when qualification can be removed safely.
Why is this a good solution?
The change extends the existing expectation grammar and edit relation with a narrow metadata input. It preserves current single-query behavior, fails safely for ambiguous syntax, and leaves more general producer-list syntax as a separately designed feature.
Testing
Go library fixtures cover shared ownership, selected-query updates, producer qualification and collapse, ignored annotations, problem queries, and invalid qualifier rejection.
This PR is based on an upstream mirror of the exact head of #22228. GitHub does not allow the fork-origin parent PR itself in a Stack object; the follow-up PRs are linked in Stack #23015. Followed by #23013.