I hate using regexp...

[CrazyCat]

Hi there, I've a little trouble using regexp.
Here's my code.

Code: Select all

proc bash:random {nick uhost handle chan args} {
	set key [http::geturl $::rand_url]
	set line [http::data $key]
	set quote [regexp -inline -all {<p class="qt">[^\f\t\v]{1,}</p>} $line]
	foreach qt [split $quote "\n"] {
		if {[string first $quote "<b>#"] > 0} {
			return
		}
		set show_quote [uncode $qt]
		if {$show_quote != ""} {
			putserv "PRIVMSG $chan :$show_quote"
		}
	}
	putserv "PRIVMSG $chan :\002End of quote\002"
}

the source code I try to parse is like this:

Code: Select all

<p class="something"><b>#1</b></p><p class="qt">here is a line<br />
and another one</p>
<p class="something"><b>#2</b></p><p class="qt">oh, a line<br />
and another one</p>

I want to obtain:

<egg> here is a line
<egg> and another one
<egg> --- Next ---
<egg> oh, a line
<egg> and another one
<egg> End of quote

and I can't because the regexp returns:

<egg> here is a line
<egg> and another one
<egg> #2 oh, a line
<egg>and another one
<egg> End of quote

I can't succeed in finding the good way to have the regexp splitting my source in an array, or other format allowing me to have each part separated.
I know I can split on '<p class="something">' but I'm sure I can find a beautifull way with regexp...

[De Kus]

if the number of lines are fix, I would suggest using neither -inline neither -all and include all matching lines into the regexp. I have regular expressions that are over 800 byte, so dont worry about it as long this is nothing called to often ^-^.

Another suggestion would be to use regsub to kill the code you dont need and split it on \n then or so.

[CrazyCat]

Thanks, but I'd resolved it using another way (have a look to the tcl bash.tcl).
In fact, my big trouble is that I'm too used to work with PHP and it's hard to forget some tools