The TRE documentation defines a range as
Two characters separated by -. This is shorthand for the full range of characters between those two (inclusive) in the collating sequence.
(here in the repository)
However, testing with the Estonian locale (in R's imported version of TRE) shows that T is incorrectly matched by [A-Z] ... this comment says
/* XXX - Should use collation order instead of encoding values in character ranges. */
Would it be correct to change the documentation to say
The characters to include are determined by Unicode code point ordering.
as in the ICU documentation ... ?
The TRE documentation defines a range as
(here in the repository)
However, testing with the Estonian locale (in R's imported version of TRE) shows that T is incorrectly matched by [A-Z] ... this comment says
Would it be correct to change the documentation to say
as in the ICU documentation ... ?