L2/03-304 Date/Time: Mon Aug 25 17:56:15 EDT 2003 Contact: henry@spsystems.net Report Type: Technical Report or Tech Note issues Regarding TR18-8 (last-call draft of Regular Expressions), section 0.1: Several later examples use \s in its Perl meaning, but there is no mention of it here, and no explanation when it occurs later. Suggest defining it here. -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- Date/Time: Mon Aug 25 18:05:54 EDT 2003 Contact: henry@spsystems.net Report Type: Technical Report or Tech Note issues Regarding TR18-8 (last call for Regular Expressions update), section 1.2, subsection "Other Properties": The discussion of string properties is missing the reference to the place where the name property is discussed. -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- Date/Time: Mon Aug 25 18:16:24 EDT 2003 Contact: henry@spsystems.net Report Type: Technical Report or Tech Note issues Regarding TR18-8 (last call for Regular Expressions update), section 2.2: The discussion at the end, of the alternative expression of literal clusters, is not really quite right. In traditional Spanish, [a-z \p{ch}] should not match the initial c in "char", even if later parts of the same regular expression would otherwise force it to, and thus it is not truly equivalent to ([a-z] | ch). The POSIX spec calls for doing this right: the real equivalence is to ([a-bd-z] | ch | c(?!h)), using negative lookahead to force "c" followed by "h" to match only "ch". -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- Date/Time: Mon Aug 25 18:18:19 EDT 2003 Contact: henry@spsystems.net Report Type: Technical Report or Tech Note issues Regarding TR18-8 (last call version of Regular Expressions update), section 2.5: The wording of the "The ISO names..." sentence has gotten garbled somewhere along the way. Possibly "the names at some point" should read "the names have changed at some point". -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- Date/Time: Mon Aug 25 18:26:41 EDT 2003 Contact: henry@spsystems.net Report Type: Technical Report or Tech Note issues Regarding TR18-8 (last call version of Regular Expressions update), section 3.9: (Repeating a comment I made on 18-7...) The wording here is confusing and ambiguous. Should getFirstMatchSet yield the code points that could appear anywhere in a match, or only those which could appear at its BEGINNING? The wording, taken literally, says the former, but there are implications of the latter, and the latter is far more useful in practice. The example does not resolve this.