Make the support matrix Basis column say what the basis is - #11
Merged
Conversation
Two rows of the table were not true. The network row read "the audit detects it" under Basis. That is a mechanism, not a basis, and it read as if nobody had measured the row - while the suite runs four network scenarios and passes all four on both architectures. The sentence itself was worth keeping, so it moved to Notes where it belongs, and Basis now says what it is for. The header claimed the suite last ran on 2026-09-03. The run data says 2026-09-08 on both architectures. Nothing about the promise changes: every status in the table is untouched, and Python and Electron keep saying in the Basis column that a person checked them and the suite did not. That distinction is the one this column exists to make. Consistency with the generated matrix is now checked rather than asserted in a comment. The check lives with the generator, outside this repository, so CI cannot run it - it compares the set of environments, the status of each, and whether the basis is the suite or a person, and deliberately not the prose, the per-architecture counts or the row order, which differ by design because the two tables are written for different readers. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Two rows of the support matrix were not true, and the drift had gone unnoticed because nothing compared the published table with the one the test harness generates.
What was wrong
The network row read
the audit detects itunder Basis. That is a mechanism, not a basis, and it read as if nobody had measured the row - while the suite runs four network scenarios and passes all four on both architectures. The sentence was worth keeping, so it moved to Notes where it belongs.The header claimed the suite last ran on 2026-09-03. The run data says 2026-09-08 on both architectures.
Nothing about the promise changes. Every status in the table is untouched, and Python and Electron keep saying in the Basis column that a person checked them and the suite did not - the distinction that column exists to make.
Why it drifted
The generator carried a comment demanding that its rows "must agree with the table in the README". That was unachievable, so it was ignored: the two tables are written for different readers. The README is prose for users - no counts, no ADR references. The generated table reports a measurement - per-architecture counts, shorthand, ADR references. Keeping them identical serves nobody, so nobody did it, and they drifted until one row contradicted itself.
Three things have to agree, and only three: the set of environments, the status of each, and whether the basis is the suite or a person. Prose, counts and row order are free to differ.
That is now checked rather than asserted in a comment. The check lives with the generator, outside this repository, so CI cannot run it - a limitation stated plainly rather than papered over. Reverting either fixed row makes it fail and name the disagreement, for example:
Gates 12/12, Rust 419, C# 365.
🤖 Generated with Claude Code