docs(r): name the one other backtick normaliser in the bare-name header

The header said every comparison of a definition spelling with a symbol
goes through rBareName. The NAMESPACE export check in import-resolvers/r.ts
compares through its own stripBackticks, which removes every leading and
trailing backtick, not one pair. Say so. Comment only.
This commit is contained in:
Mick Cooney 2026-10-17 15:03:29 +01:00
parent 963eb359a7
commit f93c6df7d6

View file

@ -2,9 +2,13 @@
* R lets any name be written in backticks (`` `%+%` ``, `` `my fn` ``,
* `` `helper` ``); the symbol is the text between them. The graph and the scope
* model keep a definition's name exactly as written, while NAMESPACE entries,
* roxygen tags and rename requests carry the symbol. Every place that compares
* the two spellings goes through {@link rBareName}, so there is one definition of
* "the same name".
* roxygen tags and rename requests carry the symbol. The scope, qualified-call,
* post-parse, rename and type code compare the two spellings through
* {@link rBareName}, so there is one definition of "the same name" there. The one
* other normaliser is `stripBackticks` in `import-resolvers/r.ts`, which removes
* every leading and trailing backtick from a NAMESPACE entry (a quoted
* `` "`x`" `` entry keeps its backticks after `normaliseArg`) before comparing it
* with a symbol.
*
* One pair of surrounding backticks is removed, in constant time. An escaped
* backtick inside the name stays as written, as `normaliseArg` in