On Thu, 2005-02-17 at 18:44 -0500, Mark E. Mallett wrote:
My prediction is that something returning the entire match string will
be much more useful than something returning the count of matches.
I agree, I can't really think of a use case for the count, so I probably
shouldn't have included it in the first place.
I'd rather have both, though, and providing them is really a minor thing,
no matter how late it is.
anyone else have an opinion?
[${0} is the match, ${#} is the count]
Or keep ${0} as the count, and add something else (e.g. ${*}) as the
complete match string.
${*} (shell) or ${&} (sed, Perl), I guess. ${0} as match corresponds to
e.g. awk. ${:} as in Python's [:] could work, too, but is strange
without the range. so many options ...
explicit is better than implicit -- users can just add an extra pair of
parentheses. (this only applies to :regex. for :matches, the match
string is always equal to the source string.)
Indeed it is- it's still useful to be able to access it.
you're right, that's actually awkward to do with :matches without this
feature:
if header :matches ["To", "Cc"] ["kj*", "za*"]
the above needs to be replaced by four individual statements. whether
that is _compelling_, I'll leave for someone else to say :-)
4. Action set
> An illegal name MUST cause a syntax error.
"MUST be detected as a syntax error" ? (at compile time? runtime?)
Just to be clear, my main comment was about the wording "must cause a
syntax error" being wrong. The parenthesised "compile time or runtime"
was a side comment and yeah, it's probably obvious that it's compile
time.
what's wrong? you mean s/cause/handled as/ ?
--
Kjetil T.