12:18 p.m.
query-replace-regexp behaves weird:
I type
abc<RET>
bcd<RET>
sar<RET>
C-P C-p C-p C-p <I am now above abc>
M-X ESC ESC
<in modeline I have:>
Redo: (query-replace-regexp "^[:space:]*\\(.*\\)" "\\$query .= \"\\1
\";" nil)
n <for first replace>
y <2nd>
y <3rd>
What I get:
$query .= "bc ";
$query .= "de ";
$query .= "r ";
What had happen to "sar" -> "r " ???
I can reproduce this with fresh started XEmacs (21.4.8) with modeline:
Query replace regexp ^[:space:]*\(.*\) with: \$query .= "\1 ";
I have has other weird regexp experience too, but I don't speak regexp, so
I has been thought that those has been my faults.
With this I really think there is something funny going on. Is there? =)
BR, Jani
--
Jani Averbach
12:43 p.m.
Jani Averbach <jaa(a)cc.jyu.fi> writes:
[...]
What I get:
$query .= "bc ";
$query .= "de ";
$query .= "r ";
What had happen to "sar" -> "r " ???
I can reproduce this with fresh started XEmacs (21.4.8) with modeline:
Query replace regexp ^[:space:]*\(.*\) with: \$query .= "\1 ";
I guess, [:space:] is a character class (correct wording?) and I
believe, XEmacs doesn't grok this. As an effect, the regexp matcher
sees a set of single characters to match against.
Have you tried using SPC and TAB in your regular expression? This
works for me.
norbert.
1:01 p.m.
New subject: forgot: (was: Re: query-replace-regexp behaves weird (bug?))
On Sat, 25 May 2002, Norbert Koch wrote:
I guess, [:space:] is a character class (correct wording?) and I
believe, XEmacs doesn't grok this. As an effect, the regexp matcher
sees a set of single characters to match against.
Ouh, of cource, I have to keep a break
[[:space:]]
BR, Jani, crawl back to hole, sorry for list-noise.
--
Jani Averbach
9:09 p.m.
New subject: forgot: (was: Re: query-replace-regexp behaves weird (bug?))
-- Jani Averbach <jaa(a)cc.jyu.fi> spake thusly:
On Sat, 25 May 2002, Norbert Koch wrote:
>
> I guess, [:space:] is a character class (correct wording?) and I
> believe, XEmacs doesn't grok this. As an effect, the regexp matcher
> sees a set of single characters to match against.
>
Ouh, of cource, I have to keep a break
[[:space:]]
XEmacs won't grok that, either. The character class for space is "\s-".
Take a look at the info page "(xemacs)Regexps"; do:
C-h i g (xemacs)Regexps
and search on that page for "\sCODE". Note that \s- will pick up
newlines also, which is often not what you want. You can also do:
[ <tab>]
where <tab> is the tab key and not the text "<tab>". This should
catch
all whitespace other than newlines (ignoring things like
carriage-return, vertical-tab and form-feed, which aren't likely to be
in your code).
9:33 p.m.
New subject: forgot: (was: Re: query-replace-regexp behaves weird (bug?))
>>>> "Matt" == Matt Tucker
<tuck(a)whistlingfish.net> writes:
Matt> -- Jani Averbach <jaa(a)cc.jyu.fi> spake thusly:
> On Sat, 25 May 2002, Norbert Koch wrote:
>
>>
>> I guess, [:space:] is a character class (correct wording?) and I
>> believe, XEmacs doesn't grok this. As an effect, the regexp matcher
>> sees a set of single characters to match against.
>>
>
> Ouh, of cource, I have to keep a break
>
> [[:space:]]
Matt> XEmacs won't grok that, either. The character class for
Matt> space is "\s-". Take a look at the info page
Matt> "(xemacs)Regexps"; do:
Hi Matt,
I think
C-h i g (xemacs)Syntax Entry
needs a patch.
It only lists:
` '
The class of whitespace characters.
even though
`-'
is supported also and has the benefit of being compatible with Emacs,
AFAIK.
How about this change, for which I could submit a patch:
`-'
The class of whitespace characters. Please don't use the
formerly advertised ` ', which is not supported by GNU Emacs.
Matt> C-h i g (xemacs)Regexps
Matt> and search on that page for "\sCODE". Note that \s- will
Matt> pick up newlines also, which is often not what you want. You
Matt> can also do:
Matt> [ <tab>]
Right, interactively one has to type
[ SPC C-q TAB ]
even though when written as a string in lisp it is generally
preferable to use "[ \t]"
Matt> where <tab> is the tab key and not the text "<tab>".
This
Matt> should catch all whitespace other than newlines (ignoring
Matt> things like carriage-return, vertical-tab and form-feed,
Matt> which aren't likely to be in your code).
--
Adrian Aichner
mailto:adrian＠xemacs.org
http://www.xemacs.org/
9:51 p.m.
New subject: forgot: (was: Re: query-replace-regexp behaves weird (bug?))
-- Adrian Aichner <adrian(a)xemacs.org> spake thusly:
I think
C-h i g (xemacs)Syntax Entry
needs a patch.
It only lists:
` '
The class of whitespace characters.
Good catch.
even though `-' is supported also and has the benefit of being
compatible with Emacs, AFAIK.
How about this change, for which I could submit a patch:
`-'
The class of whitespace characters. Please don't use the
formerly advertised ` ', which is not supported by GNU Emacs.
Seems reasonable to me.
Matt> You can also do:
Matt> [ <tab>]
Right, interactively one has to type [ SPC C-q TAB ] even though when
written as a string in lisp it is generally preferable to use "[ \t]"
For isearch-*-regexp and query-replace-regexp, when entering stuff on
the modeline the C-q is unnecessary. But it's probably a good habit
nonetheless, particularly when explaining to others.
8014
days inactive
8014
days old
5 comments
4 participants
participants (4)
-
Adrian Aichner -
Jani Averbach -
Matt Tucker -
Norbert Koch