On Wed, 2004-02-11 at 14:13, Trevor Smith wrote:
> Man, I am completely at a loss how to use a regex to search for whitespace.
>
> I *want* to search and replace some complex strings in multiple HTML files but
> I can't even successfully SEARCH -- forget about the replace!!
>
> Do regular expressions not actually work to find whitespace?!? I can't make
> them work, although the regex editor in Kate seems to do the job, when I copy
> even a simple regex that it produces to the command line for use with grep or
> egrep, they match exactly nothing.
>
> For example, a test file named "test.file" containing:
>
> some words
> some other words
>
> produces matches with:
>
> grep 'e w' test.file
>
> but NO matches with:
>
> grep 'e\sw' test.file
> egrep 'e\sw' test.file
> grep 'e[\s]w' test.file
> etc.
>
> Now, that's fine if the file I'm searching has ONLY one space, but the WHOLE
> purpose of regexes is to allow me to search flexibly. I want to search for
> different lengths of whitespace, etc., but since I can't even search for ONE
> whitespace character successfully, I'm unable to expand my regex.
>
> Any help? (Eventually I was going to build the regex to look for
> space/tab/newline characters.)
{e,}grep regex'es are not perl regex'es: try 'e +w' to match multiple
spaces. There exists a -P switch to egrep that should make it understand
perl regex'es
btw if you are going to _replace_ text you should better look at sed or
perl one-liners
--
pc
