Case-sensitive support with regexp specials ?

Case-sensitive support with regexp specials ?

[Sun Yijiang]

It seems that Emacs only provides a method to do case insensitive
search with "regexp search functions", not with "regexp specials",
like vim does.  This is not convenient when defining mixed
case-sensitive and case-insensitive keywords in some modes.  I wonder
if this is a TODO in GNU Emacs.

Thanks.

Re: Case-sensitive support with regexp specials ?

[Sun Yijiang]

In vim regexp, `\c' means "ignore case diffs from here" and `\C' means
"consider case diffs from here".  Thus, `\chello\CWorld' matches
"HeLloWorld" but does not match "helloWoRLd".  In GNU Emacs, this
regexp should be `[hH][eE][lL][lL][oO]World', which is a little bit
ugly and not very convenient in some cases.


On Wed, 15 Dec 2004 09:57:49 -0500, Richard Stallman <rms@gnu.org> wrote:
>    It seems that Emacs only provides a method to do case insensitive
>    search with "regexp search functions", not with "regexp specials",
>    like vim does.
> 
> Sorry, I don't know what that means.  I do not use vim.
> Could you explain?
>

Re: Case-sensitive support with regexp specials ?

[Kevin Rodgers]

Sun Yijiang wrote:
 > On Wed, 15 Dec 2004 09:57:49 -0500, Richard Stallman <rms@gnu.org> wrote:
 > > > It seems that Emacs only provides a method to do case insensitive
 > > > search with "regexp search functions", not with "regexp specials",
 > > > like vim does.
 > >
 > > Sorry, I don't know what that means.  I do not use vim.
 > > Could you explain?
 >
 > In vim regexp, `\c' means "ignore case diffs from here" and `\C' means
 > "consider case diffs from here".  Thus, `\chello\CWorld' matches
 > "HeLloWorld" but does not match "helloWoRLd".  In GNU Emacs, this
 > regexp should be `[hH][eE][lL][lL][oO]World', which is a little bit
 > ugly and not very convenient in some cases.

Yuck: IMHO a balanced syntax like \(?:...\) would be better than a
       state-switching construction.

Cool: You could combine that with \sw to match upper- and lower-case
words in any character set, without having to enumerate the character
ranges.

-- 
Kevin Rodgers

Re: Case-sensitive support with regexp specials ?

[Richard Stallman]

    Yuck: IMHO a balanced syntax like \(?:...\) would be better than a
	   state-switching construction.

For this operation, requiring "balance" might make it clumsy and could
also interfere with use.  I think the syntax of \c and \C fits in
better.

However, \c and \C already have meanings in Emacs regexps, for
matching in terms of character categories.  So we would need to choose
some other letter for this, or else change incompatibly to a different
letter for categories.  (Such an incompatible change would not be
tragic, since users probably don't use the syntax very much.)

Re: Case-sensitive support with regexp specials ?

[Richard Stallman]

    In vim regexp, `\c' means "ignore case diffs from here" and `\C' means
    "consider case diffs from here".  Thus, `\chello\CWorld' matches
    "HeLloWorld" but does not match "helloWoRLd".

I would not object if it were added.
However, it should wait till after the coming release.