coderaiser / putout-editor / 36734206768

100%

push

github

docs: memory: correct the parseFilesystem note, and the two index summaries

The note I wrote overstated its own finding, and the index summaries did not
mention the new content at all.

Overstated: I claimed the fixture shape made a rule match nothing in two ways,
one of which was my own probe being wrong. A crawlFile mask is not a path - it
tests value === name or the regexp against the basename - so tokens.css still
finds an absolute tokens.css, and remove-undefined-token-file is fine. I checked
rather than argued: a planted var(--color-planted-not-a-token) in codemirror.css
comes back out of redlint scan with the rule reporting it.

So the lesson is the one difference that is real, getFilename being absolute, plus
the part that actually generalises: what the rule does with the returned path is
the thing to check, not whether the mask matched. Recorded as the negative result
so it is not chased again.

The sibling part of the finding stands, and is now concrete rather than
structural: start:dev, report and gen:diagrams exist in no package but the root's,
so a document naming one would have been a false positive against
packages/client/package.json. Verified gen:diagrams is reported by neither the old
nor the new code, because the old code never got as far as comparing - which is
why the unit test, not the real run, is what pins that half.

Both index rows gained the finding they now have to describe, and the memory row
names the trap rather than the file.

1081 of 1081 branches covered (100.0%)

Branch coverage included in aggregate %.

6670 of 6670 relevant lines covered (100.0%)

61.48 hits per line