Release 0.37.1: language owner switch and rollback correctness, bare package lang filters #365

Merged
buildagent merged 17 commits from train-0371 into master 2026-10-13 22:00:07 +02:00
Member

Fixes what dogfooding v0.37.0 on this repository found (1,355 files, 879 .rs at v0.37.0). Every defect here predates 0.37.0; 0.36.1 reproduces them. The version is bumped in this PR with version_bump.sh. No schema change, and package bytes are unchanged.

Switching a language to its package, and back

1. The rollback buffer kept the wrong files (c2d5d477)

  • After languages owner rust package, the superseded generation held the 476 files that did not change hands, and none of the builtin .rs rows.
  • Cause: activation::write_generation created a new package key with host_epoch NULL. packages::reconcile_extraction_epoch then marked every file of the package changed. The re-extract ran writer::write_upsert, which deletes and re-inserts the files row and so cascaded away the rows the switch had just withdrawn.
  • Fix: contributions::mint_package_for_host creates the key at this host's epoch.

2. An owner flip re-extracted the whole tree (6cfd094d, b86b7e35)

  • Cause: the [languages] route, and the daemon startup, called build_full_observed with old == new. That runs startup_rebuild.
  • Fix:
    • Recover the old activation from the recorded digest: hash each subset of the consented [[displaces]]; a match is exact.
    • Reuse plugin enable's bounded dirty domain, with the new rule Cause::Displacement.
    • The routing pair no longer counts as a total project-config change.
    • When the record cannot be recovered, it still does the full rebuild, and logs why (NotRecovered::{NoPackageSet,NoRecord,CapExceeded,NoMatch}).

3. A file edited after a switch kept the package's rows after a rollback (6cfd094d)

  • Cause: write_upsert gives an edited file a new files.id, and the switch joined the build queue's stale file_id.
  • Fix: the switch resolves queued paths to today's id first. A re-extracted file the incoming generation holds nothing for is marked owed (Switch::owed).
  • Both statements read only the build queue, so the promotion lock stays linear in the dirty set (#307).

4. plugin enable of a package that owns nothing re-extracted every file it claims (bdf652c9, d55456dc)

  • Fix: in dirty_domain, the claim and container rules now count a file only where the package is its owner on that side. They compare the component_of triple, not the key string.
  • A file's rows come only from its routed owner, and embedded regions only from files the owner has.
  • Over an empty index the identity is not forced, so enabling before the first index stays Unchanged and does not fail at the gate.

Language filters (ca3451d7, 73455afe)

  • Bare ids match package languages: lang: "vue" / "svelte" / any package-shaped id now matches <pkg>/<id> in search_text, search_symbols and resolution_gaps, exactly as the full spelling does. The shared rule is lang_profile::filter_glob, with its Rust twin filter_admits, graded against SQLite.
  • Ambiguity: two packages declaring one bare id is refused as ambiguous_lang, with did_you_mean listing both. The spellings come from the new IndexAccess::lang_spellings.
  • Empty results: a lang filter that empties a result names the spellings present (spellings_present).
  • Payload cost: the tool descriptions are unchanged, and the startup payload is back above its reserve.

Measured on this repository (release binaries, isolated CODE_INDEX_HOME)

v0.37.0 this PR
plugin enable rust (rust stays builtin) 37.9 s, 879 reparsed 12.7 s, 0 parsed
owner flip rust → package 163 s, 2,234 extractions 45–70 s, 881
pass after the flip parse 0 (after the double extraction) parse 0
rollback buffer after the flip 476 non-Rust files, 0 .rs 881 builtin .rs, nothing else
package live binds vs builtin — 412,022 identical
edit then plugin rollback package rows stay live under the builtin 880 restored + 1 owed

Tests (each mutation was run, RED, then restored GREEN; I re-ran one mutation per lane myself)

  • crates/indexer/tests/owner_switch_rollback.rs (13 tests):
    • enable, owner flip and reverse, each bounded and rolling back byte-identically;
    • edit and create after a switch;
    • fallback to the total rebuild, and the BuiltinsOnly first activation;
    • enable/disable of a package that owns nothing, update of an owning package, enable over an empty index.
  • crates/daemon/tests/language_owner_e2e.rs::a_restart_with_a_moved_owner_re_extracts_only_the_files_that_change_hands.
  • crates/mcp-server/tests/lang_alias_e2e.rs: real vue and svelte packages, both legs; a bare id answers like the full one; an ambiguous id is refused.
  • Core and parity: lang_profile units, lang_profile_parity, rpc_coverage_e2e leg.
  • Two independent reviews: no blockers. Their should-fixes are all in, including the empty-index regression found in review.

Records that moved, attributed

  • release_gate_e2e step 8: now counts the superseded generation, (2, 3).
    • Master's total of (2, 3) came from the epoch defect: it wiped the rows of a generation already past retention. Reverting only c2d5d477 reproduces the old total.
    • Step 10's rollback is unchanged.
  • tests/corpus/host-package-lifecycle.json: ruby/inert and ruby/removed were re-measured with a reason.
    • Exactly 14 whole-table keys moved (count.*, influence.builtin_only, rule.tier1a_unique_own_file): the redundant retained copy of the one builtin-owned file (test/yajl_test.rb), which change 4 no longer re-extracts.
    • Every active.* and gen.* value is unchanged. Bisected: reverting only the narrowing reproduces the old record.
  • tests/corpus/baseline.json and cost-baseline.json are unchanged (md5 checked).
  • Registered: DISPLACEMENT_HYPOTHESES_MAX_CONSENTS in bounding_site_registry (capacity: it changes no answer). The leg-routed suite population is 58.

Verification

  • One full local run on the train, then a re-run of every step that was red, on the final tree 15da7dbd:
    • fmt; clippy (workspace, first-party, windows-gnu, guest); rustdoc --document-private-items -D warnings;
    • workspace nextest 5119/5119, daemon leg 932/932, ci-timing 34/34; guest gates;
    • the CI corpus job (11 suites), psr4_vs_builtin, tier 3;
    • precision_gate 7/7, phantom 0;
    • all 11 package_leg.sh <pkg> full and all 7 overhead legs.
  • Dogfood on this repo with the 15da7dbd release binaries: the measurement table above.

Not changed, by design: plugin rollback puts the index back but leaves the approval and [languages] as they were, so the next pass re-applies them (its help says so). Making a rollback stick takes disable or languages owner <lang> builtin.

🤖 Generated with Claude Code

https://claude.ai/code/session_01YDDBdWU4yThi8Zi51E6FCj

Fixes what dogfooding v0.37.0 on this repository found (1,355 files, 879 `.rs` at v0.37.0). Every defect here predates 0.37.0; 0.36.1 reproduces them. The version is bumped in this PR with `version_bump.sh`. No schema change, and package bytes are unchanged. ## Switching a language to its package, and back **1. The rollback buffer kept the wrong files** (c2d5d477) - After `languages owner rust package`, the superseded generation held the 476 files that did *not* change hands, and none of the builtin `.rs` rows. - **Cause:** `activation::write_generation` created a new package key with `host_epoch` NULL. `packages::reconcile_extraction_epoch` then marked every file of the package changed. The re-extract ran `writer::write_upsert`, which deletes and re-inserts the `files` row and so cascaded away the rows the switch had just withdrawn. - **Fix:** `contributions::mint_package_for_host` creates the key at this host's epoch. **2. An owner flip re-extracted the whole tree** (6cfd094d, b86b7e35) - **Cause:** the `[languages]` route, and the daemon startup, called `build_full_observed` with old == new. That runs `startup_rebuild`. - **Fix:** - Recover the old activation from the recorded digest: hash each subset of the consented `[[displaces]]`; a match is exact. - Reuse `plugin enable`'s bounded dirty domain, with the new rule `Cause::Displacement`. - The routing pair no longer counts as a total project-config change. - When the record cannot be recovered, it still does the full rebuild, and logs why (`NotRecovered::{NoPackageSet,NoRecord,CapExceeded,NoMatch}`). **3. A file edited after a switch kept the package's rows after a rollback** (6cfd094d) - **Cause:** `write_upsert` gives an edited file a new `files.id`, and the switch joined the build queue's stale `file_id`. - **Fix:** the switch resolves queued paths to today's id first. A re-extracted file the incoming generation holds nothing for is marked owed (`Switch::owed`). - Both statements read only the build queue, so the promotion lock stays linear in the dirty set (#307). **4. `plugin enable` of a package that owns nothing re-extracted every file it claims** (bdf652c9, d55456dc) - **Fix:** in `dirty_domain`, the claim and container rules now count a file only where the package is its owner on that side. They compare the `component_of` triple, not the key string. - A file's rows come only from its routed owner, and embedded regions only from files the owner has. - Over an empty index the identity is not forced, so enabling before the first index stays `Unchanged` and does not fail at the gate. ## Language filters (ca3451d7, 73455afe) - **Bare ids match package languages:** `lang: "vue"` / `"svelte"` / any package-shaped id now matches `<pkg>/<id>` in `search_text`, `search_symbols` and `resolution_gaps`, exactly as the full spelling does. The shared rule is `lang_profile::filter_glob`, with its Rust twin `filter_admits`, graded against SQLite. - **Ambiguity:** two packages declaring one bare id is refused as `ambiguous_lang`, with `did_you_mean` listing both. The spellings come from the new `IndexAccess::lang_spellings`. - **Empty results:** a `lang` filter that empties a result names the spellings present (`spellings_present`). - **Payload cost:** the tool descriptions are unchanged, and the startup payload is back above its reserve. ## Measured on this repository (release binaries, isolated `CODE_INDEX_HOME`) | | v0.37.0 | this PR | |---|---|---| | `plugin enable` rust (rust stays builtin) | 37.9 s, 879 reparsed | **12.7 s, 0 parsed** | | owner flip rust → package | 163 s, 2,234 extractions | **45–70 s, 881** | | pass after the flip | parse 0 (after the double extraction) | parse 0 | | rollback buffer after the flip | 476 non-Rust files, 0 `.rs` | **881 builtin `.rs`, nothing else** | | package live binds vs builtin | — | **412,022 identical** | | edit then `plugin rollback` | package rows stay live under the builtin | 880 restored + 1 owed | ## Tests (each mutation was run, RED, then restored GREEN; I re-ran one mutation per lane myself) - **`crates/indexer/tests/owner_switch_rollback.rs`** (13 tests): - enable, owner flip and reverse, each bounded and rolling back byte-identically; - edit and create after a switch; - fallback to the total rebuild, and the BuiltinsOnly first activation; - enable/disable of a package that owns nothing, update of an owning package, enable over an empty index. - **`crates/daemon/tests/language_owner_e2e.rs::a_restart_with_a_moved_owner_re_extracts_only_the_files_that_change_hands`.** - **`crates/mcp-server/tests/lang_alias_e2e.rs`:** real vue and svelte packages, both legs; a bare id answers like the full one; an ambiguous id is refused. - **Core and parity:** `lang_profile` units, `lang_profile_parity`, `rpc_coverage_e2e` leg. - **Two independent reviews:** no blockers. Their should-fixes are all in, including the empty-index regression found in review. ## Records that moved, attributed - **`release_gate_e2e` step 8:** now counts the *superseded* generation, (2, 3). - Master's total of (2, 3) came from the epoch defect: it wiped the rows of a generation already past retention. Reverting only c2d5d477 reproduces the old total. - Step 10's rollback is unchanged. - **`tests/corpus/host-package-lifecycle.json`:** ruby/inert and ruby/removed were re-measured with a reason. - Exactly 14 whole-table keys moved (`count.*`, `influence.builtin_only`, `rule.tier1a_unique_own_file`): the redundant retained copy of the one builtin-owned file (`test/yajl_test.rb`), which change 4 no longer re-extracts. - Every `active.*` and `gen.*` value is unchanged. Bisected: reverting only the narrowing reproduces the old record. - `tests/corpus/baseline.json` and `cost-baseline.json` are unchanged (md5 checked). - **Registered:** `DISPLACEMENT_HYPOTHESES_MAX_CONSENTS` in `bounding_site_registry` (capacity: it changes no answer). The leg-routed suite population is 58. ## Verification - **One full local run on the train**, then a re-run of every step that was red, on the final tree 15da7dbd: - fmt; clippy (workspace, first-party, windows-gnu, guest); rustdoc `--document-private-items -D warnings`; - workspace nextest 5119/5119, daemon leg 932/932, ci-timing 34/34; guest gates; - the CI corpus job (11 suites), `psr4_vs_builtin`, tier 3; - `precision_gate` 7/7, phantom 0; - all 11 `package_leg.sh <pkg> full` and all 7 `overhead` legs. - **Dogfood on this repo with the 15da7dbd release binaries:** the measurement table above. **Not changed, by design:** `plugin rollback` puts the index back but leaves the approval and `[languages]` as they were, so the next pass re-applies them (its help says so). Making a rollback stick takes `disable` or `languages owner <lang> builtin`. 🤖 Generated with [Claude Code](https://claude.com/claude-code) https://claude.ai/code/session_01YDDBdWU4yThi8Zi51E6FCj
An owner switch withdrew the replaced rows into the superseded
generation correctly, and the very next index pass destroyed them.
`activation::write_generation` minted a new package's key through
`contributions::ensure_package`, leaving `plugin_packages.host_epoch`
NULL ("derived under an unknown epoch") although every row under it was
about to be written by this host. `packages::reconcile_extraction_epoch`
then marked every file the package had just produced as changed, the
pass extracted them a second time, and `writer::write_upsert`'s
delete-and-reinsert of each `files` row cascaded the withdrawn rows of
exactly the files that changed hands out of the rollback buffer. A
rollback restored nothing for the switched language and left the
package's rows live under a builtin activation.

`contributions::mint_package_for_host` inserts a NEW key at
`HOST_EXTRACTION_EPOCH` and leaves an existing key's epoch alone; NULL
stays reserved for keys minted over rows an older binary wrote.

`tests/owner_switch_rollback.rs` drives both operator routes (`plugin
enable`, and the index pass that finds a moved `[languages]` owner) plus
the reverse switch, and grades the buffer and the rollback
byte-identically.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01YDDBdWU4yThi8Zi51E6FCj
`files.lang` for a .vue/.svelte file is `de.h-dv.vue/vue` /
`de.h-dv.svelte/svelte`, so `search_text {lang: "vue"}` answered 0 while
the full spelling answered the file. `lang_profile::filter_glob` already
widened a bare PROFILE name (`ruby` -> `*/ruby`, #112); it now widens
any local-shaped bare id, so every tool built on `sql_lang_filter`
(search_text incl. the #355 region arm, search_symbols symbols+imports,
the token tier, resolution_gaps) answers `vue` exactly as
`de.h-dv.vue/vue`. Full spellings and builtins are unchanged.

- ambiguity: a bare non-profile id two packages declare is refused
  (`ambiguous_lang`, candidates in did_you_mean) by the MCP server for
  search_text / search_symbols / resolution_gaps, over the indexes the
  call searches. Spellings come from a new `IndexAccess::lang_spellings`
  RPC (files.lang + live extraction_components.lang); an older daemon
  answers "unknown method" -> not checked, never refused.
- resolution_gaps' unknown_lang check uses the same rule (it refused
  `vue`, and `typescript` held only in a region).
- empty_population: when `lang` alone emptied a result, its probe
  carries `spellings_present` (nearest first) and the note names them.
- Rust twin `filter_admits` graded against SQLite in
  lang_profile_parity; `is_local_shaped` graded against LocalLangId.

Tests: lang_alias_e2e (real frozen vue+svelte packages; two fixture
packages sharing `tmpl`), both legs. Mutations run: restoring the
profile-only gate -> RED `lang "vue" must answer exactly what
"de.h-dv.vue/vue" answers` left ([], 0) right (["src/Panel.vue"], 1);
ambiguous_lang -> None -> RED left None right Some("ambiguous_lang").

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01YDDBdWU4yThi8Zi51E6FCj
THE OWNER FLIP. An index pass that finds the recorded activation digest
moved rebuilt every walked file (`DirtySet::startup_rebuild`): it had
only the digest, not the old claim tables. `languages owner rust
package` on this repository re-extracted 1,355 files to move 879 and
copied the 476 carried ones into the rollback buffer.
`activation::recorded_displacement` recovers the old activation when
`[languages]` is all that moved (one digest per subset of the packages'
consented displacements) or when the package set arrived whole
(the builtins-only digest), and the pass then builds
`dirty::dirty_domain`'s bounded set through `build::build_bounded_observed`.
`dirty_domain` gains `Cause::Displacement`: a container on both sides
whose displacement moved dirties exactly the walked files whose owner
differs. Anything else (host, worker, package bytes or grants, project
config) still rebuilds everything, because for those the record really
cannot name the old claim tables.

AN EDIT BETWEEN A SWITCH AND ITS ROLLBACK. `writer::write_upsert`
re-creates the `files` row, so the build queue's recorded `file_id` went
stale and the rollback's withdraw missed the file: the package's rows
stayed live under the builtin activation and the next pass stat-skipped
them. The switch now resolves each queued PATH to today's `files` row,
and marks a re-extracted file the incoming generation holds nothing for
as changed (`Switch::owed`), so the next pass extracts the current bytes
under the restored activation. Both statements are over the queue only.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01YDDBdWU4yThi8Zi51E6FCj
The daemon's startup activation had its own copy of the index pass's
reconcile and still rebuilt every walked file on an owner flip;
`activation::recorded_activation` is now the one derivation both routes
call. And `dirty::dirty_domain` no longer treats the `[languages]`
routing pair as a project-config move: it is a function of the packages'
displacements, which the `Cause::Displacement` rule already bounds. The
running daemon's re-arm hands `dirty_domain` the old host's config and
the new one's, so a live `languages owner` flip was a whole-tree rebuild
there too.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01YDDBdWU4yThi8Zi51E6FCj
recorded_displacement / recorded_activation return NotRecovered
(NoPackageSet, NoRecord, CapExceeded{consents}, NoMatch{candidates})
instead of None, and both callers (index::reconcile_activation_before_index,
the daemon's startup activation) log it at info, so a full rebuild is
never unexplained. The consent cap is per (package, builtin) pair and the
doc now says so. The project config is read once per recovery; only the
routing pair varies between candidates.

A file created or renamed into the switched domain after a switch has no
build-queue row, but needs no new rule: the next pass's route gate
(reconcile::classify_stat, producer half for first-party ids) reclassifies
it. owner_switch_rollback now grades that on a first-party shape in both
directions, plus the BuiltinsOnly recovery and the NoMatch fallback.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01YDDBdWU4yThi8Zi51E6FCj
The startup activation's build queue (generation_build_files of the new
active generation) is the durable record of what it re-extracted; it must
hold exactly the two .rb files. Mutation (run): always build_full_observed
in startup_activation -> RED, the queue also holds src/main.rs.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01YDDBdWU4yThi8Zi51E6FCj
plugin enable of a package that claims files it owns none of (the
first-party de.h-dv.rust under auto) re-extracted every claimed file with
the builtin: 879 .rs files on this repository, most of a ~36 s enable.
The claim and container/grant rules in dirty::dirty_domain now count a
file only where the package's claim routes it AND the package is the
file's owner on that side (Activation::owner, compared as the
component_of triple). Rows a package writes come only from files the
router hands it, embedded regions included, so a file it owns on
neither side has the same producer before and after. prior_producers is
not narrowed. owner answers are memoised per (side, path), so the
precedence pass's call count is unchanged.

An empty domain is not an unchanged activation: build_with_window forces
the identity when the two sides' package activations differ, so the
generation still records the package. And an EMPTY index is never
recovered: it carries nothing, a bounded set over it can be empty, and
the gate refuses a generation that produced and carries nothing — the
daemon's first start retried that forever (caught by three daemon e2es).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01YDDBdWU4yThi8Zi51E6FCj
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01YDDBdWU4yThi8Zi51E6FCj
An empty dirty set with moved package activations forced a generation;
over an EMPTY files table (before a first index) that generation
produced and carried nothing and the gate recorded gate_failed, so
plugin enable reported Failed where it used to be Unchanged.
build_with_window now forces only when files has rows; the first index
pass activates the package as part of the first extraction. The index
route's empty-index branch says so at info, and its comment matches the
Result it now handles.

owner_switch_rollback adds: enable over an empty index (Unchanged, the
first pass activates), an update of an owning package (new digest, same
key and claims: exactly its files, as ExtractionIdentity), and a disable
of a package that owns nothing (re-extracts nothing).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01YDDBdWU4yThi8Zi51E6FCj
The bare-id rule needs no schema prose: `lang: "vue"` simply works and an
empty result names the spellings held. Dropping the three added `lang`
doc lines returns the startup payload to master's (15878 of 16515
platform-invariant tokens, 637 left, reserve 622 — was 15945/570, into
the reserve). `harness_settle_barrier`: 57 -> 58 leg-routed suites
using the shared `initialize`, the arrival named.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01YDDBdWU4yThi8Zi51E6FCj
release_gate_e2e step 8 counted the claimed file's rows across EVERY
generation and expected (2, 3). Steps 4, 5 and the disable each
re-extract the markup, so two withdrawn generations hold a copy: the
superseded one (step 5's, which step 10 restores) and step 4's, already
deleting past SUPERSEDED_RETENTION_LIMIT and held only until collection.
The old (2, 3) existed because step 4's copy had been destroyed by the
epoch defect fixed in c2d5d477 (measured: reverting mint_package_for_host
to ensure_package alone brings the total back to (2, 3); master shows
generation 2 holding no row for the file). The assertion now asks the
superseded generation, (2, 3); step 10's rollback still restores the
package rows. Mutation (run): SUPERSEDED_RETENTION_LIMIT = 0 -> RED (0, 0).

bounding_site_registry: DISPLACEMENT_HYPOTHESES_MAX_CONSENTS, capacity,
NotAgentVisible - exceeding it changes no answer, only costs a full
rebuild, and is logged as NotRecovered::CapExceeded.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01YDDBdWU4yThi8Zi51E6FCj
A claimed file the package owns on neither side is no longer re-extracted
(dirty_domain), so the retained generation no longer holds a byte-identical
second copy of the ladder's one builtin-owned file (test/yajl_test.rb:
10 symbols, 62 refs, 2 imports, 7 resolved). Exactly 14 whole-table keys
moved (count.*, influence.builtin_only, rule.tier1a_unique_own_file); every
active.* and gen.* value is unchanged. Bisected: reverting only the
narrowing reproduces the old record. baseline.json / cost-baseline.json
untouched.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01YDDBdWU4yThi8Zi51E6FCj
Merge lane R gate fixes (release_gate retention count, cap registered) into train-0371
All checks were successful
Package CI / package (typescript) (pull_request) Successful in 27s
Package CI / package (php) (pull_request) Successful in 36s
Package CI / package (rust) (pull_request) Successful in 29s
Package CI / package-corpus (python) (pull_request) Successful in 20s
Package CI / package-corpus (javascript) (pull_request) Successful in 29s
Package CI / package-corpus (rust) (pull_request) Successful in 30s
Package CI / package (vue) (pull_request) Successful in 45s
Package CI / package-corpus (csharp) (pull_request) Successful in 45s
Package CI / package (ruby) (pull_request) Successful in 45s
Package CI / package-corpus (php) (pull_request) Successful in 38s
Package CI / package-corpus (vue) (pull_request) Successful in 37s
Package CI / package-corpus (xaml) (pull_request) Successful in 41s
Package CI / package-corpus (svelte) (pull_request) Successful in 47s
Package CI / package-corpus (timeline) (pull_request) Successful in 54s
Package CI / package-corpus (typescript) (pull_request) Successful in 53s
Package CI / package-corpus (ruby) (pull_request) Successful in 56s
Package CI / package overhead budgets (nightly, one package at a time) (pull_request) Has been skipped
CI / guest crates (fmt, clippy, doc) (pull_request) Successful in 1m38s
CI / CI lane wall-clock headroom (pull_request) Successful in 1m48s
CI / cargo doc (intra-doc links) (pull_request) Successful in 5m26s
CI / cargo deny (pull_request) Successful in 7m14s
CI / cargo test (abi, 32-bit + wasm32) (pull_request) Successful in 7m17s
CI / cargo check (MSRV 1.98) (pull_request) Successful in 7m34s
CI / cargo clippy (pull_request) Successful in 7m40s
CI / cargo check (windows-gnu) (pull_request) Successful in 7m48s
CI / cargo test (pull_request) Successful in 21m51s
CI / OSS corpus (tier 1) (pull_request) Successful in 32m19s
CI (Windows) / fmt + clippy + build + test (windows) (pull_request) Successful in 40m8s
CI / cargo test (daemon transport) (pull_request) Successful in 13m0s
CI / Plugin path cost + pool throughput (nightly) (pull_request) Has been skipped
15da7dbd62
Sign in to join this conversation.
No reviewers
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
h-dv/code-index!365
No description provided.