Skip to content

Retain bounded alternatives for ambiguous overload calls - #11780

Merged
Rich Chiodo (rchiodo) merged 3 commits into
mainfrom
rchiodo-overload-pr-publication
Sep 21, 2026
Merged

Rich Chiodo (rchiodo) merged 3 commits into
mainfrom
rchiodo-overload-pr-publication

Conversation

@rchiodo

Copy link
Copy Markdown
Collaborator

Why this change

Restore the bounded local overload-conformance case motivating #11601 without repeating the loss of type information and error checking reported in #11731. This is a new implementation following the #11739 rollback, not a reapplication of the original widening behavior.

The overload specification requires a permissive result when gradual arguments leave different return types ambiguous. For example, a list[Any] argument can match overloads returning list[int] and list[str]. Choosing only the first return incorrectly rejects the valid list[str] assignment and string append.

Making that result Any or Unknown is permitted by the specification, but it discards information we still know: both alternatives are lists, their element types are known, and neither has arbitrary members. Earlier attempts exposed the danger of gaining permissiveness by losing this information: real errors could disappear along with the unwanted conformance errors. An ordinary union is not a substitute either; its operations must be valid across the alternatives, which does not provide the required permissiveness for invariant containers.

The goal is therefore to preserve as much information as possible while allowing the operations justified by the surviving overloads, rather than treating fewer diagnostics as success by itself.

How information and error checking are preserved

The internal OverloadResult retains complete ordinary return alternatives, their generic arguments, the ordinary baseline type and the uncertainty kind. It lives in TypeResult.type, so extracting a Type does not silently discard the alternatives. Supported aliases and generic projections carry the alternatives forward; public queries display OverloadResult[list[int], list[str]] honestly rather than presenting a misleading first match or erased container. This is an internal representation, not Python annotation syntax.

Each supported operation requires one complete successful alternative for the entire operation. Partial successes cannot be combined across alternatives. With the example above, both invariant assignments and either int/string append are accepted, but append(3.14), assignment to list[bytes], a nonexistent member, or a mixed-element extend accepted by neither complete alternative remain errors. Different uses may select different witnesses; this is not a claim about runtime mutation correlation.

Preserving a candidate's Type alone is insufficient. Evaluation also carries incomplete/error/contextual/binding and overload metadata, cached results and diagnostics. Candidate checking uses ordinary evaluator validation with isolated cache and diagnostic ownership. A discarded trial must not leak false errors, and diagnostics for ordinary work whose caches survive must not disappear with that trial. The cold-query, query-order, speculative evaluation, cancellation and edit/restore coverage addresses these failure modes. A plausible return type by itself is not proof that an operation succeeded.

Concrete calls retain ordinary precision. Unsupported producers, constructors/bound-method producers, escaping uses and unsupported gradual shapes retain ordinary behavior. Admission/proof/work limits remain bounded; exhausting them is not permission to fall back to a more permissive dynamic type. This does not claim full general overload conformance.

Automatic activation is the normal ConfigOptions/Program default, including CLI, language-server and background analysis. There is no new public setting; experimentalOverloadResults = false remains an internal baseline control, with explicit internal controller precedence. The separate checker live-cache handoff experiment is not adopted: production uses checkerHandoff: false, and handoff requires explicit internal opt-in.

The cumulative change is substantial: 37 files, +6,459/-72 lines, including 20 production files and 15 test/sample files. It includes the original checkpoint, subsequent reviewed activation/cache/diagnostic correctness work, public type printing/transport, default wiring and documentation, not merely the latest default switch.

Validation

  • Fresh integrated validation: core, CLI and LSP test-server builds pass. Targeted: 10 suites / 856 tests. Full: 84 suites / 3,122 tests pass, with 12 existing opt-in tests skipped. Repository-supported transpile-only Jest follows a separate full TypeScript build, including tests. Scoped ESLint/Prettier pass; existing listener/deprecation/force-exit warnings remain. No integration-specific test expectations changed.
  • Fresh ordinary and --threads 2 CLI checks show the exact carrier and successful conformance assignments/appends. Four invalid-operation controls remain errors, including a mixed-element operation with no complete witness; the concrete control remains list[int] with its intentional argument error. Module/function complexity behavior matches main at limits 768/4096.
  • Fresh three-way comparisons use exact a07bbdea rollback, pinned current main, and integrated feature, with identical frozen project files, dependencies, interpreters, options and current-main stubs. Frozen pandas source/wheel retain zero errors and complete matching 935/931 native diagnostics across all arms and historical controls. These are frozen inputs, not latest pandas catalog checkouts.
  • Broader pinned mypy_primer coverage remains partial: 12/153 catalog projects ran, 15 prior dependency setups were TLS-blocked and were not retried, and 126 projects were not attempted. Across 981 tracked files, all 1,954 native diagnostics match across the three arms, with zero CLI/native additions or removals; 659 baseline errors remain. attrs covers its actual one-file catalog command.
  • Compared with historical runs, latest-main stubs add two message_factory errors in packaging and pip's vendored packaging. Both also reproduce using rollback code with current stubs; neither is introduced by this feature. Thirty native replacements change only related stub-location metadata. No historical primary error disappears; complete before/after records are retained.
  • All external projects and pandas controls have zero carrier activations: this establishes ordinary-fallback compatibility only, not active external or all-corpus-clean coverage. A separate default-on local conformance control activates one producer/eight carrier outcomes with no handoffs.

Performance and limits

This deliberately trades additional checking work for retained information. Collapsing an ambiguous result to a dynamic type is cheap, but prevents subsequent checks from using the original alternatives. Here, candidate operations can repeat argument matching, member lookup and code-flow evaluation, with cache isolation and diagnostic handling. The observed active-case cost is not merely the allocation cost of storing several types; that allocation cost was not independently quantified.

To limit ordinary-path cost, activation starts from a native overload match and performs bounded eligibility/proof work before building the local admission graph. Measured natural zero-use controls avoid module scans, owned candidate trials and private restoration. This is not a zero-cost claim: controller setup and request/retirement boundaries remain. Active work is bounded, and unsupported or budget-limited cases retain ordinary behavior rather than skipping required checks.

Historical pre-integration paired measurements against the same implementation with the feature explicitly off:

Workload Initial analysis Complete fixed public-query batch
Ordinary controls Sensitivity-limited; no equivalence claim +0.01175 to +0.17415 ms (+3.89% to +11.63%)
Active-alternatives fixture About +37.06 ms (+23.34%) About +17.00 ms (+742%)
Mixed active/ordinary fixture About +36.10 ms (+19.46%) About +25.63 ms (+727%)

The large active-query percentages have small ordinary baselines, but the absolute added batch costs are real. These are whole matched workload batches, not individual queries, hover latency or IDE-wide estimates. Ordinary initial-analysis estimates were noisy, and an additional dataclass query-batch cost versus the preceding V3 implementation was inconclusive on confirmation. No claim of exact-zero overhead or an ecosystem-wide average is supported.

These costs were accepted provisionally for continuing default-on development because retaining useful error checking is the purpose of the change. Observation counters and timing runs were separate. No fresh integrated-tree timing was performed: these numbers are historical measurements, not measurements or performance certification of the updated-main implementation. External projects in the compatibility run did not activate alternatives and therefore do not establish real-world active-case performance.

Integration status

The complete feature is integrated onto pinned canonical main 916d7dc6 in a fresh worktree. The original reviewed worktree and historical artifacts are preserved.

Three additive conflict resolutions retain both upstream and feature additions in Program options, evaluator options/closure and cache-test imports. Upstream complexity configuration/API, contextual TypeForm lookup/tests, TypedDict union handling and updated typeshed are retained. The cumulative current-main-relative feature footprint remains 37 files, +6,459/-72. No domain, limit, default or handoff-policy change was made during integration. Publication targets only microsoft/pyright:main.

Preserve complete overload return alternatives and whole-operation witnesses without erasing container precision or losing ordinary diagnostics. Enable bounded lazy activation by default while retaining the internal baseline control and leaving checker live-cache handoff disabled.

Carry reviewed regression coverage and documentation onto pinned main, preserving upstream complexity, TypeForm cache, and TypedDict behavior.

Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
@bschnurr

Bill Schnurr (bschnurr) commented Sep 18, 2026 •

Copy link
Copy Markdown
Member

🔒 Automated review in progress — Bill Schnurr (@bschnurr) is auto-reviewing this PR.

@github-actions

This comment has been minimized.

@bschnurr

Copy link
Copy Markdown
Member

Please add runtime coverage for the new type-server protocol conversion path; the current targeted tests do not exercise it.

@bschnurr Bill Schnurr (bschnurr) left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Approved via Review Center.

@bschnurr Bill Schnurr (bschnurr) added the review-auto:approved Automated review: no blocking findings (approval posted). label Sep 18, 2026
@github-actions

This comment has been minimized.

Defer cache eviction until the owning file check or outer request unwinds, bind controllers to their own evaluator, and prevent late asynchronous cleanup from recreating a disposed Program. Add regressions for diagnostic completeness, precise overload alternatives, cancellation, and shutdown.

Classification: B (Pyright evaluator lifecycle limitation). No typeshed changes or type-precision reductions.

Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
@github-actions

This comment has been minimized.

@github-actions

This comment has been minimized.

@bschnurr

Copy link
Copy Markdown
Member

Result: ✅ verified

Verification details

Verification: Isolated verification observed failures that were not classified as caused by this PR: Repository discovery preflight.

Summary: Targeted verification passed: 164 tests across overload-result units, controller behavior, configuration, CLI analysis, and language-server paths. The LSP test-server build also succeeded, including foreground/background and push/pull diagnostic cases. Dependencies were restored successfully with the offline frozen lockfile. An initial repository-discovery probe used an invalid empty working directory, but this was unrelated to the PR and did not prevent test execution.

Test runs: 4 passed, 1 failed

  • ❌ Failed | unrelated to this PR | Repository discovery preflight | printf '%s\n' '== sandbox ==' "$AUTOMATION_SANDBOX_PROFILE" '== head ==' && git log -1 --oneline && printf '%s\n' '== changed files ==' && git diff --name-status HEAD^ && printf '%s\n' '== test scripts ==' && node -e "const p=require('./packages/pyright-internal/package.json'); console.log(JSON.stringify(p.scripts,null,2))" && printf '%s\n' '== dependencies ==' && if [ -x packages/pyright-internal/node_modules/.bin/jest ]; then echo package-jest-present; else echo package-jest-missing; fi && if [ -d node_modules ]; then echo root-node_modules-present; else echo root-node_modules-missing; fi
  • ✅ Passed | Overload-result type, cache isolation, and utility tests | node ./node_modules/jest/bin/jest.js src/tests/overloadResult.test.ts src/tests/testOnlyCacheTransaction.test.ts src/tests/typeCacheUtils.test.ts src/tests/typeUtils.test.ts --runInBand
  • ✅ Passed | OverloadResultController targeted suite | node ./node_modules/jest/bin/jest.js src/tests/overloadResultController.test.ts --runInBand
  • ✅ Passed | OverloadCall11 and automatic-overload configuration tests | node ./node_modules/jest/bin/jest.js src/tests/typeEvaluator6.test.ts src/tests/config.test.ts --runInBand --testNamePattern='OverloadCall11|automatic overload results'
  • ✅ Passed | Automatic overload-result language-server integration | pnpm run webpack:testserver && node ./node_modules/jest/bin/jest.js src/tests/languageServer.test.ts --runInBand --testNamePattern='automatic overload results default'
❌ Repository discovery preflight diagnostic output
fatal: not a git repository (or any parent up to mount point /)
Stopping at filesystem boundary

Define the evaluator-owned isolation contract and rename production cache APIs without changing transaction, speculation, or diagnostic semantics. Preserve all existing type expectations and the evaluator lifetime fix.

Classification: D (runtime coverage gap). Add real JSON-RPC carrier conversion and nested factory/reference coverage, plus exact-node restoration coverage. No typeshed changes or type-precision regressions.

Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
@rchiodo

Copy link
Copy Markdown
Collaborator Author

Fixed the runtime protocol-coverage request in ea2db303.

Added a real in-process JSON-RPC GetComputedType test that normally activates the overload-result carrier and checks both nested generic alternatives, instance/generic flags, any uncertainty, the baseline reference ID, and unchanged concrete-call output. A supplemental real ProtocolTypeFactory test covers a nested unknown-origin carrier and repeated reference identity without broadening automatic admission.

Both new tests failed when the OverloadResult conversion branch was temporarily disabled, and passed after restoring it. There is no persistent conversion implementation change. The final local full suite passed with 3,135 tests; CI for this pushed commit is still pending.

@github-actions

Copy link
Copy Markdown
Contributor

Diff from mypy_primer, showing the effect of this PR on open source code:

sympy (https://github.com/sympy/sympy)
+   .../projects/sympy/sympy/solvers/bivariate.py:135:15 - error: Operator "-" not supported for "None" (reportOptionalOperand)
+   .../projects/sympy/sympy/solvers/bivariate.py:139:17 - error: Operator "-" not supported for type "Basic | Unknown" (reportOperatorIssue)
+   .../projects/sympy/sympy/solvers/bivariate.py:144:23 - error: Operator "-" not supported for "None" (reportOptionalOperand)
+   .../projects/sympy/sympy/solvers/deutils.py:234:14 - error: Operator "not in" not supported for types "str" and "Unknown | int"
+     Operator "not in" not supported for types "str" and "int" (reportOperatorIssue)
+   .../projects/sympy/sympy/solvers/ode/lie_group.py:619:61 - error: Operator "-" not supported for type "Unknown | Basic" (reportOperatorIssue)
-   .../projects/sympy/sympy/solvers/ode/ode.py:1433:17 - error: Argument of type "One | NegativeOne | Zero | Integer | NaN | ComplexInfinity | Rational | Unknown | Expr" cannot be assigned to parameter "value" of type "Zero" in function "__setitem__"
+   .../projects/sympy/sympy/solvers/ode/ode.py:1433:17 - error: Argument of type "One | NegativeOne | Zero | Integer | NaN | ComplexInfinity | Rational | Infinity | NegativeInfinity | Float | Number | Expr" cannot be assigned to parameter "value" of type "Zero" in function "__setitem__"
-     Type "One | NegativeOne | Zero | Integer | NaN | ComplexInfinity | Rational | Unknown | Expr" is not assignable to type "Zero"
+     Type "One | NegativeOne | Zero | Integer | NaN | ComplexInfinity | Rational | Infinity | NegativeInfinity | Float | Number | Expr" is not assignable to type "Zero"
-   .../projects/sympy/sympy/solvers/ode/ode.py:1579:38 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:1580:38 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:1590:9 - error: No overloads for "update" match the provided arguments (reportCallIssue)
-   .../projects/sympy/sympy/solvers/ode/ode.py:1590:12 - error: "update" is not a known attribute of "None" (reportOptionalMemberAccess)
-   .../projects/sympy/sympy/solvers/ode/ode.py:1590:19 - error: Argument of type "Unknown | dict[Unknown, Unknown] | None" cannot be assigned to parameter "m" of type "Iterable[tuple[str, Unknown]]" in function "update"
-     Type "Unknown | dict[Unknown, Unknown] | None" is not assignable to type "Iterable[tuple[str, Unknown]]"
-       "None" is incompatible with protocol "Iterable[tuple[str, Unknown]]"
-         "__iter__" is not present (reportArgumentType)
-   .../projects/sympy/sympy/solvers/ode/ode.py:1597:43 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:1597:64 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:1603:9 - error: No overloads for "update" match the provided arguments (reportCallIssue)
-   .../projects/sympy/sympy/solvers/ode/ode.py:1603:12 - error: "update" is not a known attribute of "None" (reportOptionalMemberAccess)
-   .../projects/sympy/sympy/solvers/ode/ode.py:1603:19 - error: Argument of type "Unknown | dict[Unknown, Unknown] | None" cannot be assigned to parameter "m" of type "Iterable[tuple[str, Unknown]]" in function "update"
-     Type "Unknown | dict[Unknown, Unknown] | None" is not assignable to type "Iterable[tuple[str, Unknown]]"
-       "None" is incompatible with protocol "Iterable[tuple[str, Unknown]]"
-         "__iter__" is not present (reportArgumentType)
-   .../projects/sympy/sympy/solvers/ode/ode.py:1610:43 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:1610:74 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:1616:9 - error: No overloads for "update" match the provided arguments (reportCallIssue)
-   .../projects/sympy/sympy/solvers/ode/ode.py:1616:12 - error: "update" is not a known attribute of "None" (reportOptionalMemberAccess)
-   .../projects/sympy/sympy/solvers/ode/ode.py:1616:19 - error: Argument of type "Unknown | dict[Unknown, Unknown] | None" cannot be assigned to parameter "m" of type "Iterable[tuple[str, Unknown]]" in function "update"
-     Type "Unknown | dict[Unknown, Unknown] | None" is not assignable to type "Iterable[tuple[str, Unknown]]"
-       "None" is incompatible with protocol "Iterable[tuple[str, Unknown]]"
-         "__iter__" is not present (reportArgumentType)
-   .../projects/sympy/sympy/solvers/ode/ode.py:1623:49 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:1623:70 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
+   .../projects/sympy/sympy/solvers/ode/ode.py:1753:36 - error: Cannot access attribute "lhs" for class "Expr"
+     Attribute "lhs" is unknown (reportAttributeAccessIssue)
+   .../projects/sympy/sympy/solvers/ode/ode.py:1753:55 - error: Cannot access attribute "rhs" for class "Expr"
+     Attribute "rhs" is unknown (reportAttributeAccessIssue)
+   .../projects/sympy/sympy/solvers/ode/ode.py:1754:36 - error: Cannot access attribute "lhs" for class "Expr"
+     Attribute "lhs" is unknown (reportAttributeAccessIssue)
+   .../projects/sympy/sympy/solvers/ode/ode.py:1755:17 - error: No overloads for "__setitem__" match the provided arguments (reportCallIssue)
+   .../projects/sympy/sympy/solvers/ode/ode.py:1755:17 - error: Argument of type "Equality | BooleanFalse | BooleanTrue | Unknown | Expr" cannot be assigned to parameter "value" of type "Equality | BooleanFalse | BooleanTrue" in function "__setitem__"
+     Type "Equality | BooleanFalse | BooleanTrue | Unknown | Expr" is not assignable to type "Equality | BooleanFalse | BooleanTrue"
+       Type "Expr" is not assignable to type "Equality | BooleanFalse | BooleanTrue"
+         "Expr" is not assignable to "Equality"
+         "Expr" is not assignable to "BooleanFalse"
+         "Expr" is not assignable to "BooleanTrue" (reportArgumentType)
-   .../projects/sympy/sympy/solvers/ode/ode.py:2961:17 - error: Argument of type "One | NegativeOne | Zero | Integer | NaN | ComplexInfinity | Rational | Unknown | Expr" cannot be assigned to parameter "value" of type "Zero" in function "__setitem__"
+   .../projects/sympy/sympy/solvers/ode/ode.py:2961:17 - error: Argument of type "One | NegativeOne | Zero | Integer | NaN | ComplexInfinity | Rational | Infinity | NegativeInfinity | Float | Number | Expr" cannot be assigned to parameter "value" of type "Zero" in function "__setitem__"
-     Type "One | NegativeOne | Zero | Integer | NaN | ComplexInfinity | Rational | Unknown | Expr" is not assignable to type "Zero"
+     Type "One | NegativeOne | Zero | Integer | NaN | ComplexInfinity | Rational | Infinity | NegativeInfinity | Float | Number | Expr" is not assignable to type "Zero"
-   .../projects/sympy/sympy/solvers/ode/ode.py:3442:7 - error: "update" is not a known attribute of "None" (reportOptionalMemberAccess)
-   .../projects/sympy/sympy/solvers/ode/ode.py:3442:38 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:3443:7 - error: "update" is not a known attribute of "None" (reportOptionalMemberAccess)
-   .../projects/sympy/sympy/solvers/ode/ode.py:3443:38 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:3444:14 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:3445:14 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:3446:14 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:3458:50 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:3459:50 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:3460:50 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:3500:5 - error: No overloads for "update" match the provided arguments (reportCallIssue)
-   .../projects/sympy/sympy/solvers/ode/ode.py:3500:7 - error: "update" is not a known attribute of "None" (reportOptionalMemberAccess)
-   .../projects/sympy/sympy/solvers/ode/ode.py:3500:14 - error: Argument of type "Unknown | dict[Unknown, Unknown] | None" cannot be assigned to parameter "m" of type "Iterable[tuple[str, Unknown]]" in function "update"
-     Type "Unknown | dict[Unknown, Unknown] | None" is not assignable to type "Iterable[tuple[str, Unknown]]"
-       "None" is incompatible with protocol "Iterable[tuple[str, Unknown]]"
-         "__iter__" is not present (reportArgumentType)
-   .../projects/sympy/sympy/solvers/ode/ode.py:3501:18 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:3501:43 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:3502:9 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:3502:16 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:3502:24 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)
-   .../projects/sympy/sympy/solvers/ode/ode.py:3502:31 - error: Object of type "None" is not subscriptable (reportOptionalSubscript)

... (truncated 502 lines) ...

@github-actions

Copy link
Copy Markdown
Contributor

mypy_primer analysis

Advisory AI analysis of primer run 35406802579, attempt 1
for commit ea2db303b8ec8d502c0d9d54f60b81c7cde34180. This is not an approval or proof of correctness.

Only SymPy changed. These differences are treated as non-blocking primer noise; no other project changed.

Recorded changes: 181 added / 152 removed diagnostic headers; 181 added / 61 removed detail lines.

Full evidence and limitations: download mypy-primer-analysis-report from this analysis run.
The original raw primer comment is unchanged. Treat uncertainty and possible regressions as requests for human review.

@github-actions

Copy link
Copy Markdown
Contributor

Compared candidate c2b2151f26ac07728470ce3517ded8d62bc182d9 against its first parent 916d7dc62e063ba6342493e44fc55389d42d5120.

Type checker benchmark

🟢 No performance regressions detected.

Regression threshold: 20.0%
Variance guard: >1.0s time and >100.0 MB memory

Package Checker Files checked Time Time delta Peak memory Memory delta Status
ansible pyright 583 17.675s -18.4% 1119.7 MB +0.8% 🟢 Pass
click pyright 17 2.714s -6.7% 352.6 MB -2.3% 🟢 Pass
homeassistant pyright 9850 127.048s -19.8% 6094.8 MB -1.6% 🟢 Pass
numpy pyright 356 32.296s -25.6% 1350.9 MB -27.0% 🟢 Pass
pandas pyright 1459 886.008s -10.4% 4409.9 MB -0.5% 🟢 Pass
pytest pyright 243 11.275s -10.3% 847.3 MB -9.0% 🟢 Pass
requests pyright 19 2.219s -3.5% 340.0 MB -2.1% 🟢 Pass
torch pyright 2400 125.379s -18.5% 4932.3 MB +0.8% 🟢 Pass
transformers pyright 2901 116.572s -13.4% 5029.3 MB -2.0% 🟢 Pass

Pyright stats

Package Parsed/bound Checked Find Read Tokenize Parse Imports Bind Check Cycles
ansible 989 583 0.040s 0.110s 0.370s 0.720s 0.200s 0.980s 14.290s 0.000s
click 111 17 0.000s 0.010s 0.100s 0.190s 0.030s 0.250s 1.810s 0.000s
homeassistant 11147 9850 0.410s 0.780s 2.720s 4.870s 2.090s 6.390s 106.700s 0.000s
numpy 603 356 0.010s 0.080s 0.360s 0.710s 0.230s 0.830s 29.510s 0.000s
pandas 1895 1459 0.060s 0.250s 1.070s 2.220s 0.290s 2.300s 878.540s 0.000s
pytest 542 243 0.020s 0.050s 0.330s 0.530s 0.110s 0.670s 9.120s 0.000s
requests 168 19 0.000s 0.020s 0.110s 0.200s 0.050s 0.250s 1.280s 0.000s
torch 3079 2400 0.140s 0.380s 2.160s 3.620s 0.610s 4.640s 112.380s 0.000s
transformers 3537 2901 0.120s 0.420s 2.010s 3.780s 0.900s 8.880s 100.210s 0.000s

@rchiodo
Rich Chiodo (rchiodo) merged commit bf07bcb into main Sep 21, 2026
21 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

review-auto:approved Automated review: no blocking findings (approval posted).

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants