chore: promote develop to main (0.12.0-dev alignment) - #204
Merged
Conversation
…out (mcb-v5an.11)
- Box CallToolResult in handler Result types to reduce Err variant size - Box loco_rs::Error in config and MCP server initializers - Unbox errors at call sites where protocol expects concrete type - Add kubernetes-validate dependency for Python test suite
…Result - pyproject: flext-core/flext-cli now come from the GitHub 0.12.0-dev branch (operator directive: the venv always carries the pushed branch, never stale local snapshots); uv.lock regenerated. - annotate analyze_issues/analyze/summarize/generate and the register_result_command handler contract as p.Result[T]: the canonical flext factory protocol now that FlextResultConstruction.ok/fail declare p.Result[V]. Concrete FlextResult return annotations rejected the protocol values (mypy). - remove the lib.agent_pointers/lib.gitops mypy ignore_errors override: the root cause is fixed, suppression no longer needed. Validated: make check WHAT=python (ruff + mypy + 64 pytest + guard) green with flext-core git@acae76c72.
Same serde field-ordering refresh as the earlier snapshot updates, now covering every remaining domain (index, memory, project, search, session, validate, vcs, entity invalid_args). Verified: full mcb-server contract suite 28/28 green.
Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
Replace proprietary boot/ship dual-surface Makefile with flext-infra codegen conform output and domain hooks in custom.mk, preserving MCB dispatch via makefiles/domain.mk mapped to flext public verbs. Update surface checks, hook entrypoints, and pre-commit unit scope for the flext verb matrix; exclude slow surface-matrix tests from test-staged. Co-authored-by: Cursor <cursoragent@cursor.com>
Unblocks flext-managed pre-push without mass-rewriting scripts/ typing. Remaining docstring/typing debt stays tracked as ignores. Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
Pre-push was blocking on the 15-minute surface matrix (work WHAT=tags timeout). Keep full surface via make check WHAT=surface; default python gate matches pre-commit with pytest -m not slow. Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
Parent boot WHAT=hook leaked into recursive make test and flext rejected unsupported test WHAT=hook. Co-authored-by: Cursor <cursoragent@cursor.com>
Recoverable checkpoint of in-flight work. Not a delivery; gates deferred. Excludes .arbor/cache (exceeds the remote file size limit). Origin branch: feature/v0-4-0-multitenant-weaviate
Recoverable checkpoint of in-flight work. Not a delivery; gates deferred. Origin branch: feature/v0-4-0-multitenant-weaviate
Recoverable checkpoint of in-flight work. Not a delivery; gates deferred. Origin branch: feature/v0-3-2-ci-gates
…-2-ci-gates-20260809T174917Z Recovery/mcb local feature v0 3 2 ci gates 20260809 t174917 z
…ns/cache-6.1.0 build(deps): bump actions/cache from 4.3.0 to 6.1.0
- Register /alive (public liveness) and /ready (dependency-aware readiness) - Register /metrics (conditional on MetricsConfig.endpoint_enabled) - Fix E0382: compute metrics_enabled before moving settings - Bump crossbeam-epoch 0.9.18 -> 0.9.20 (RUSTSEC-2026-0204) - Remaining quick-xml 0.38.4 blocked by loco-rs/opendal pin
merge: sync develop with FileHashService dyn-compat fix
PR #178 ("Recovery/mcb sanitized wip") merged 619 files / +135,696 lines into develop without review, versioning one workstation's configuration and session residue as if it were source. The live consequence was a red CI: .cargo/config.toml pins linker = "clang" with -fuse-ld=mold, and mold does not exist on GitHub runners, so run 31430407703 failed at exit code 2 with even libc, serde, quote and proc-macro2 build scripts unable to link. The mold error was the symptom; the disease is that CI ran under a build contract that only exists on the operator machine. Untrack, keeping files on disk: - machine-local config: .cargo/, .vscode/, .waza.yaml, opencode.jsonc - tool output: *violations*.txt (6), qlty.*.sarif, build_err.txt, clippy_output.txt, validate_output.txt, test_003_warnings.txt, pr116_comments.txt, staged_files.txt, coverage/lcov.info - artifacts: libdefine_validator_test.rlib, data/vectors/, tests/mcb.2026-02-10 (server stdout), test_file.txt - scratch: tmp_sisyphus_test.txt, .beads.gate.lock - context/: parallel docs tree outside the docs/ SSOT that documented dill and figment, both removed by ADR-050 and ADR-051 Harden .gitignore, the barrier that let the dump in. It carried an explicit !.vscode/settings.json exception forcing personal editor state to be tracked, and a dead !.github/workflows/release.yml rule for a file since renamed to release-binaries.yml. Both corrected, and each purged class now has a blocking rule, verified by an empty git status --untracked-files=all. Net: -82,181 lines. Refs: mcb-okll
run_git set current_dir but left the inherited git environment intact. git resolves GIT_DIR, GIT_WORK_TREE, GIT_INDEX_FILE and GIT_CONFIG ahead of the working directory, so whenever the suite ran inside a git operation — a pre-commit or pre-push hook, a rebase, a merge — every helper call was redirected from its TempDir onto the surrounding real repository. Reproduced by exporting only GIT_DIR before the suite: 10 of 17 vcs tests fail with "could not lock config file /home/marlonsc/mcb/.git/config" and "Unable to create .git/worktrees/purge-pr178-wip-dump/index.lock". The decisive symptom is a TempDir `git commit` executing the real repository's pre-commit pipeline, running the project's own Python suite and reformatting a source file from inside a unit test. run_git now clears the six inherited path variables, disables system and global config, and points core.hooksPath at /dev/null, so a test repository cannot reach outside itself or execute project tooling. Same command and same exported GIT_DIR after the fix: 17 passed, and the run drops from 22.68s to 0.07s now that tests no longer trigger the host hook pipeline. Covered by tests asserting the helper repository is self-contained, carries its initial commit through git's own surface, and is not captured by an ambient work tree — exercised through the public helper API, with no process-wide environment mutation so the suite stays parallel-safe. Refs: mcb-k5yl
test_gap2_vcs_list_repositories_discovers_repos drove git through its own
raw Command::new("git"), duplicating logic the domain test helper already
owns. That copy inherits GIT_DIR, GIT_WORK_TREE and GIT_INDEX_FILE, so
under a pre-push hook the fixture repository was created against the
surrounding repository instead of the temp dir, and the assertion found no
discovered repo.
Route the fixture through run_git, the single owner, which clears the
inherited variables and disables hooks and system/global config.
Reproduced with the hook environment exported: the test failed before and
now passes; gap_fixes_e2e is 6 passed under the same conditions.
Refs: mcb-k5yl
… helper
create_git_repo_fixture drove git through its own raw Command::new("git"),
the third copy of logic the domain test helper owns. That copy inherits
GIT_DIR, GIT_WORK_TREE and GIT_INDEX_FILE, so under a pre-push hook the
fixture was created against the surrounding repository and
test_vcs_index_repository_success plus the three
test_vcs_list_repositories_cases found no repository.
Route the fixture through run_git, the single owner, and set the commit
identity on the repository itself now that the helper no longer reads
global config.
Reproduced with the hook environment exported: the four tests failed
before and handlers::vcs_tests is now 7 passed under the same conditions.
Refs: mcb-k5yl
enumerate_with_git passed `-C workspace_root`, but git resolves GIT_DIR, GIT_WORK_TREE, GIT_INDEX_FILE and GIT_COMMON_DIR ahead of `-C`. Whenever validation runs from inside a git operation — a pre-commit hook, a rebase, a merge — the inherited environment redirected `git ls-files` to the surrounding repository, so the inventory described a different tree than the one the caller asked about. This is a production defect, not a test-only concern: validation is invoked from the pre-commit path, exactly where the environment is set. Clear the inherited variables before invoking git, so the inventory always describes workspace_root. The fixture in run_context_tests drove git through its own raw Command as well; it now routes through the shared run_git helper. Reproduced with the hook environment exported: the test failed before and run_context_tests is now 2 passed under the same conditions. Refs: mcb-k5yl
fix(repo): purge PR #178 WIP dump and stop git escaping into the real repository
The installed hooks (.git/hooks/pre-commit, .git/hooks/pre-push) dispatch to the private handlers in custom.mk, and neither handler stated a CI token. Both tiers therefore ran under "CI absent", so the gate set each tier executed was whatever the caller environment happened to carry rather than the tier's own contract. CI is ternary (flext-infra config/codegen.yaml, RULING 1): CI=Y omits the gates the CI workflows own (lint, format, pyrefly, markdown) and revokes pytest, CI=N runs the full suite with coverage and keeps every blocking gate, and absent is testmon incremental. The two hook tiers want the two explicit states: pre-commit is the fast tier and declares CI=Y; pre-push is the last complete gate before work leaves the machine and declares CI=N. The token is stated per command because each recipe line is its own process. Covered by two tests that expand the handlers through make's own dry-run, so they assert what the hook really executes instead of copying the Makefile text: test_pre_commit_hook_runs_under_ci_yes test_pre_push_hook_runs_under_ci_no RED before the change listed every command as an offender (5 for pre-commit, 9 for pre-push); GREEN after. `make run WHAT=mcb-hook-pre-commit` exits 0 with clippy clean across all 7 workspace crates under -D warnings.
`make gen WHAT=check` reported drift on freshly-absorbed develop and wanted to rewrite files that were already correct. The projections were right; the generator was stale. mcb resolves FLEXT_INFRA_RUNTIME_ROOT to itself, so codegen runs the flext_infra installed in .venv, which carries its own config/codegen.yaml. uv.lock pinned that package to b76af238, 139 commits behind 0.12.0-dev, and its config still held the pre-bump action pins. Regenerating from it would have downgraded the GitHub Action pins that dependabot already landed here through PRs #180, #182 and #183: jdx/mise-action v4.2.4 -> v4.2.3 actions/cache v6.1.0 -> v4.3.0 actions/upload-artifact v7.0.1 -> v4.6.2 Upstream had already fixed this in 98642b2d3 ("fix(codegen): preserve current action pins"). Moving the pin to 0.12.0-dev HEAD (c5a5f2ec6) adopts it, and .github/workflows/ci.yml consequently drops off the drift list. The managed dependency bumps in pyproject.toml (pyrefly, pytest, ruff, rumdl) come from the same refresh and are emitted by the generator, not hand-written.
`.gitignore` is a generated projection, but the rules PR #201 added to stop the PR #178 dump from recurring were only ever present in the projection. Any `make gen` therefore deleted them: the machine-local config block, the tool output block, the compiled/captured artifact block, the runtime log, scratch and parallel-docs blocks all disappeared, and `!.vscode/settings.json` came back, so `.vscode/settings.json` was recreated as a tracked candidate. The barrier against re-committing 82k lines of workstation state lived in the one file the generator is free to overwrite. flext-infra already provides the seam for this (mro-jnm1.3): a repository policy overlay contributes `extra_ignored_patterns`, appended after the fleet-wide scaffold sections, precisely so a project never hand-edits the generated file. mcb declared an overlay but no patterns, so the seam was unused. Declaring them in config/workspace.yaml makes regeneration reproduce the barrier instead of stripping it, and because the overlay section is appended last, `/.vscode/` now correctly overrides the scaffold's `!.vscode/settings.json`. Covered by test_generated_gitignore_keeps_declared_project_exceptions, which reads both sides from their real files, so the test fails if the overlay ever stops reaching the rendered artifact. RED before the change (no patterns declared); GREEN after, with all 9 spot-checked patterns present and `git check-ignore -v .vscode/settings.json` resolving to `.gitignore:318:/.vscode/`. Regenerating also adopts the current upstream projections: per-verb pre-commit and pre-push hooks (the runner now reports each make verb separately), the `docs` verb removal, and the pytest case timeout default.
The file held only comments after the third-party forks became pinned Cargo
dependencies: no [submodule] section, and no gitlink in the index either.
flext-infra reads it before any lifecycle step. Its own contract already
treats an ABSENT .gitmodules as no submodules and returns ok({}), but a
PRESENT file makes it shell out to git config -f .gitmodules --get-regexp,
which exits 1 when nothing matches. That non-zero exit surfaced as a hard
failure and stopped make work WHAT=start from provisioning any lane.
Removing the file restores the state the tooling already handles, and it is
what the generator renders for a project with zero gitlinks: base/gitmodules.j2
emits nothing when workspace_gitlinks is empty, so the comment block was a
hand-edit of a generated projection.
…utover ERRADICATED: .claude/settings.json marketplace, mcb_scripts.egg-info, .agents/provider.toml + flext-law, orphan submodules, .gitmodules. CANONICAL LAYOUT: scripts/lib -> src/mcb_scripts (14 modules + qlty + docs), 57 imports rewritten, sys.path bypass removed, tests -> tests/python. OWNER-FIRST: cli.py DELETED (reimplemented flext_cli.cli), 8 consumers rewired, app()->cli.execute_app, model_rebuild on 6 models, toml->tomllib. GATES RESTORED: _custom_* overrides removed, builtin wins, surface.py removed. DOC REPAIR: deleted docs restored from git, broken links repointed. setup=0 gen=0 fmt=0 fix=0 check=0 pytest=67p/0f
fix: eradicate local flext, adopt canonical src layout, owner-first cutover
The PR #202 src-layout cutover moved scripts/lib into src/mcb_scripts, activating coverage source=["src"] over modules that have no live mcb consumer: - cosmos_command.py: only imported by scripts/dispatch.py, which the generated Makefile never invokes (SCRIPT_VERBS is empty). - workspace_command.py + workspace.py: ai-hub workspace template machinery with zero mcb consumers. - scripts/analyze_qlty.py + qlty/main|parser|runner: standalone qlty CLI wrapper not wired to any make verb or hook. qlty/model|report|strategies stay: gitops.py consumes them for SARIF policy issues (test_policy_issues_reuse_qlty_report_model proves it). make test WHAT=full: 71 passed, coverage 23.79% -> 49.95% (fail_under=45). custom.mk also gains _custom_work_{start,land,finish} that dispatch the builtins with env -u MAKEFLAGS: the work saga pushes via GitPython, which inherited WHAT=land/APPLY=Y into the pre-push hook's make invocations and aborted every land with 'unsupported gen WHAT=land'.
The env -u carrier clear was insufficient: make exports command-line variables (WHAT=land APPLY=Y) into every recipe's environment directly, so GitPython's git push still delivered them into the pre-push hooks and every land aborted with 'unsupported gen WHAT=land'. Unset both the dispatch tokens and the make flag carriers in the same shell that launches the builtin saga.
The pinned flext-infra rev (26c848a8) had lost the 'unset WHAT MAKEFLAGS APPLY' hook scrub, so every 'make work WHAT=land APPLY=Y' leaked its selectors into the pre-push hooks and the land aborted with 'unsupported gen WHAT=land'. Upstream flext-infra tips also carried committed conflict markers (cli/core/tests pyprojects, infra templates), which made 'make deps lock' fail at TOML parse. Upstream repair (separate repos, already pushed): - flext-cli f56a29dd, flext-core 8748e2df, flext-tests 29f4c493: resolve the committed pytest-timeout conflict markers. - flext-infra 532eab82: resolve markers across templates/tests, keeping the hook scrub and R12 verb ownership. Here: re-lock onto those tips, regenerate, and let the fresh runtime project the scrub into .pre-commit-config.yaml (12 entries). custom.mk drops its now-reserved _custom_work_* and _custom_build_artifacts handlers (the new parse-time monopoly guard rejects them); the Rust build moves to a pre-build hook. make test WHAT=full: 71 passed, coverage 49.95% (fail_under=45) including the two env-leak tests.
…ct-markers hotfix/resolve-committed-conflict-markers: lane land
Contributor
|
ⓘ Qodo reviews are paused because the subscription is no longer active. Ask your workspace admin to reactivate the subscription to resume reviews. Manage billing |
|
Important Review skippedAuto reviews are disabled on base/target branches other than the default branch. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: Organization UI Review profile: ASSERTIVE Plan: Pro Plus Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
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
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.
Promote develop to main after full gate alignment, conflict marker resolution, and green CI.
Summary by cubic
Promotes develop to main and aligns the repo with flext‑infra 0.12.0‑dev. Removes the accidental WIP dump, restores clean CI, hardens git/test isolation, and upgrades
git2with a small domain API change.develop; stop persisting checkout tokens; docs job uses theflext_infraCLI; update action pins; rename release workflow file..envrcenables Python bytecode caching; adopt managed ruff per-file ignores;config/workspace.yamldeclares package + script dispatch; newconfig/managed-artifacts.yaml.GIT_*and disables hooks/global config; validator clears inherited git env so inventories target the requested workspace; add unit tests for isolation.FileHashService::compute_hashnow takes&self; service modules renamed to*_serviceand error paths adjusted; bumpgit2to0.21andcrossbeam-epochto0.9.20..gitignoreand add.gitops-cache/.gitignore; drop empty.gitmodules.Actions
pre-commit installanduv syncafter pulling to pick up new hooks and Python deps.FileHashService, update the signature tofn compute_hash(&self, path: &Path) -> Result<String>.Written for commit ca95dbd. Summary will update on new commits.