You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit 78f118d
Browse filesBrowse the repository at this point in the historyBrowse files
crates/socket-patch-core/src/utils/purl.rs 1428 audit-core c07af61 2026-10-29 PR #1920 make parse_name_version private; inline parse_jsr_purl/build_jsr_purl into deno_crawler; drop JsrPurlParts alias and is_purl; 7 build_*_purl are bare format! wrappers; ~25 comment lines (strip_purl_qualifiers essay, REGRESSION/B73/#1024 tags, banner); delete 7 test_build_*_purl + 7 *_round_trip + test_is_purl tautologies; move PurlKey asserts to purl_key.rs; fold 3 test mods; three splitters: A14
527
527
crates/socket-patch-core/src/utils/purl_key.rs 633 audit-core c07af61 2026-10-29 - inline release_identity/nuget_identity/composer_identity into PurlKey::new and is_canonical_base into canonical_base_purl; hand purl splits -> A14; CLI scan/rollout.rs L18-141 and policy.rs L451 key maps by into_string(): key by PurlKey; test nuget_key_agrees_with_the_vendored_version_match recomputes expected via normalize_nuget_version; ~8 comment lines; imports crawlers+vendor: A08
528
528
crates/socket-patch-core/src/utils/purl_key_nuget_vendor_tests.rs 151 audit-core c07af61 2026-10-29 - misplaced: end-to-end vendor_nuget + vex test with nothing purl_key-specific; move beside vendor nuget tests (test_support::vendor_nuget); drop history module doc (#1202 story)
529
-
crates/socket-patch-core/src/utils/python_lock.rs 2283 audit-core c07af61 2026-10-29 - delete dead pub complete_python_lock_metadata(+_in); fold PythonLockProbe into PythonLockSession::plan, drop probe_equivalence_tests; inline check_python_lock_source_scope; one repoint(table, artifact) helper for 4 copies; source_identity is a 2nd uv-source reader beside UvSource; ~15 comment lines; 'ours' test asserts is_ok() (passes on Ok(None)); fold 5 test mods; see A24, A25
530
-
crates/socket-patch-core/src/utils/python_script.rs 1768 audit-core c07af61 2026-10-29 - drop always-None script param of refuse_uv_no_sources/no_binary and fold the 4 refuse_uv_* wrappers into refuse_package_setting; make is_hosted_override private; one tool_uv(doc) for 4 chains; one direct-deps fn for 3 copies (L319, L383, L891); inline push_hosted_override; ancestor_uv_config_files: use relpath (A19); trim SourcesLayout doc; ~12 bare #refs; untested no_binary; see A15, A22, A24
529
+
crates/socket-patch-core/src/utils/python_lock.rs 2283 audit-core c07af61 2026-10-29 PR #1990 delete dead pub complete_python_lock_metadata(+_in); fold PythonLockProbe into PythonLockSession::plan, drop probe_equivalence_tests; inline check_python_lock_source_scope; one repoint(table, artifact) helper for 4 copies; source_identity is a 2nd uv-source reader beside UvSource; ~15 comment lines; 'ours' test asserts is_ok() (passes on Ok(None)); fold 5 test mods; see A24, A25
530
+
crates/socket-patch-core/src/utils/python_script.rs 1768 audit-core c07af61 2026-10-29 PR #1990 drop always-None script param of refuse_uv_no_sources/no_binary and fold the 4 refuse_uv_* wrappers into refuse_package_setting; make is_hosted_override private; one tool_uv(doc) for 4 chains; one direct-deps fn for 3 copies (L319, L383, L891); inline push_hosted_override; ancestor_uv_config_files: use relpath (A19); trim SourcesLayout doc; ~12 bare #refs; untested no_binary; see A15, A22, A24
531
531
crates/socket-patch-core/src/utils/redact.rs 552 audit-core c07af61 2026-10-29 PR #1923 dead pub url_userinfo/strip_url_credentials + test: A20; fix module doc L22-24; private is_uuid_shaped twin: A15; move the_service_download_advisory_has_one_builder to vendor/service_fetch.rs (it lints that file)
532
532
crates/socket-patch-core/src/utils/relpath.rs 232 audit-core c07af61 2026-10-29 - code clean; module doc 'every normalizer in the crate' is false (4 copies, A19); trim history in regression-test doc L217-220
533
533
crates/socket-patch-core/src/utils/repo_root.rs 363 audit-core c07af61 2026-10-29 PR #1923 inline GitRepo::dot_git (1 caller, policy/mod.rs); home(): A10; module doc 'the one walk' is false (A21); drop bare 'B22' tag L240
`main` @ `cbbbdb9`, unchanged since the last run. The previous entry was structural (#1988), so this run is a sweep. None of my 28 open PRs has merged or closed, and the discussion has no maintainer steering.
4
+
5
+
I took core `utils/{python_lock,python_script}.rs`, the next free audited rows on the queue. No open PR touches them; I checked by diffing all 73 open heads locally. The callers I changed in `hosted/engine.rs` and `patch/redirect/mod.rs` are held by 10 PRs, and `git merge-tree` shows the branch merges cleanly with each. No open PR adds a caller of a removed or changed item.
6
+
7
+
**PR:**[#1990](https://github.com/SocketDev/socket-patch/pull/1990), ready for review, with Bugbot triggered. CI is green on all 82 checks, and Bugbot found no issues.
8
+
9
+
### Per file
10
+
-**`python_lock.rs`:**
11
+
- 2 dead fns deleted (`complete_python_lock_metadata`, `complete_python_lock_metadata_in`).
12
+
-`PythonLockProbe` deleted. The wheel gate now calls `PythonLockSession::plan`, and the probe-equivalence test went with it.
13
+
- New `names` and `repoint` helpers replace 5 name checks and 4 repoint copies.
14
+
- The planner reuses `lock_package_collection`, and new `artifact_filename` replaces 2 copies.
15
+
-`python_lock_paths` is one iterator chain.
16
+
- 4 test modules are folded into one.
17
+
- About 20 comment lines removed (a banner, 6 bare issue numbers, docs).
18
+
- 1 fake test fixed: the "ours" case asserted only `is_ok()`, which passes on `Ok(None)`.
19
+
-**`python_script.rs`:**
20
+
- New `tool_uv` replaces 3 lookup chains, and new `declares_directly` replaces 3 copies.
21
+
- 1 fn inlined (`push_hosted_override`), and `is_hosted_override` is now private.
22
+
-`refuse_uv_no_sources`/`no_binary` lose their always-`None``script` parameter.
23
+
- About 30 comment lines removed (15 bare issue numbers, the `SourcesLayout` doc).
24
+
- 1 test added for no-binary's combine rules.
25
+
-**Total:** +226 / −531. Production is about −195 and tests about −110.
26
+
27
+
### Evidence
28
+
- Clippy (workspace, all features) is clean.
29
+
- Core lib passes 6,995. The 4 failures are the known root-sandbox ones.
30
+
- Red→green:
31
+
- "ours" with `plan` returning `Ok(None)`: red.
32
+
- The PEP 751 hash kept as the old digest: red.
33
+
- The no-binary script bool not winning: red, and redirect's own test is red too.
34
+
-`/code-review high` found no correctness bugs and raised 9 findings. 3 are fixed: the "ours" assert still passed because the input already held the URL, a local shadowed `tool_uv`, and the transitive-revert caveat had been dropped. The PR gives a reason for each of the other 6.
35
+
36
+
### Merges recorded
37
+
None.
38
+
39
+
### Lessons
40
+
- When you tighten a fake test, check that the new assert can't pass on the input itself. Here the input already contained the URL the assert looked for.
41
+
- A `/// #NNN: text` doc rewritten with a regex that capitalizes the first letter turns "uv" into "Uv". Lowercase tool names need a hand check.
42
+
- The audit noted `python_script`'s project no-binary refusal as untested. Redirect's `no_binary_refuses_a_wheel_uv_would_not_install` does cover it end to end, so grep the other modules' tests before you call something untested.
|#1939| sweep |`commands/vendored_backend/{mod,repair}.rs`| ready; green |
24
24
|#1944| structural | T51 #1941: 11 `tests/apply` suites + `common/envelope.rs` onto `apply_fixture` (test-only, −635 net) | ready; `hosted-e2e` red on a live 403, not this diff |
0 commit comments