Conversation
prefetch:msvc / prefetch:mingw invoked bare `python`, which resolves to the ssl-less 32-bit Pyodide shim when the python guest env is loaded (CI runs the full MISE_ENV set) -- portable-msvc.py died with `ModuleNotFoundError: No module named 'ssl'`. Both tasks now use the pinned CPython via the new py3_win_fwd var. The mingw shim's `__type_info_root_node` was 8-mod-16 aligned, but it anchors the SLIST_HEADER that vcruntime140's __std_type_info_name chains name allocations onto, and x64 SLIST operations require 16-byte alignment (cmpxchg16b) -- the first type_info::name() call access-violated inside ntdll!ExpInterlockedPopEntrySListEnd. Only dotnet-data1 hit it (its dotnet.js is the one module JS that touches Intl, reaching the path via Intl.DateTimeFormat -> v8/ICU LocaleCacheKey::hashCode). .balign 16 fixes it; the dotnet-data1 web-runner test now passes locally under mingw. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The step was carried over from test-msvc.yaml on a stale premise: since win-dotnet2 (#70) merged, build-ws-dotnet-data1-module builds the module under Windows mise (including the cmd 8191-char PATH cap handled as an MSBuild property override), so build-modules already covers it via the build-ws-* glob. Also drop the two README lines claiming the .NET WASM build does not work under Windows mise. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
New semgrep rule comment-summary-line-c applies the summary-first convention to /* */ blocks and // runs (no C linter has this check -- clang-tidy stops at namespace-closing/task-tag comments, cpplint at whitespace), and the existing # rule now also covers *.s GAS comments. The mingw-shim sources are restyled to comply. Also prune comment fragments that restated adjacent code or duplicated rationale documented elsewhere. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
gen:python-ws and gen:python-rest skipped on Windows on the stale premise that ruff/coreutils were off the busybox-ash PATH; they resolve fine now. The real (undocumented) blocker was line endings -- ruff's default `line-ending = auto` wrote CRLF on Windows, so the committed LF trees drifted. Pin `[format] line-ending = lf` in ruff.toml and the tasks regenerate byte-identical output on every OS, so the skips are removed. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
pierre-tenedero
approved these changes
Jul 4, 2026
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.
No description provided.