On Wed, 2005-07-20 at 10:51 -0700, Ned Freed wrote:
Basic idea looks good, however, a couple of typos have crept in. First, the
syntax at the beginning of 4.1 fails to list the two new modifiers.
doh! thanks.
Second, the plain text draft now says:
4.1.2.1. Modifier ":quotewildcard"
Every character with special meaning for :matches ("*", "?" and " is
prefixed with "
doh! it was supposed to say:
Every character with special meaning for :matches ("*", "?" and
"\") is prefixed with "\" in the expansion.
4.1.2.2. Modifier ":quoteregex"
Every character with special meaning for :regex (".", "*", "?" etc.)
is prefixed with "
I assume you meant '"\"' and not '"'. Looks like you need to apply some
quoting
to the draft source...
yeah, I always forget that \ must be written \e in roff.
The lack of a full list of regex specials is a bit awkward here, but I
guess I can live with it.
there is no ABNF in [REGEX] to refer to, and a full list will be prone
to inaccuracies. not that I expect any changes to the syntax allowed by
it, but ...
Finally, the section ends with "Using two or more modifiers of the same
precedence is a syntax error.". This isn't quite right: These aren't syntax
errors. How about saying "It is an error to use two or more modifiers of the
same precedence in a single SET operation" or something similar.
thanks. I used "command" rather than "operation" to avoid using new
terminology.
I also downgraded refering to an unknown namespace to just an "error"
rather than a "syntax error".
I did not remove the text about interaction with regex yet. I feel the
text belongs in the variables draft, so if it is procedurally possible,
I prefer it stays there rather than move into the regex draft.
I also would like to keep the text here if possible.
I guess we'll decide this in Paris?
--
Kjetil T.