[thelist] Fwd: Regular Expression - Is it a bug or am I missing something very obvious?

VOLKAN ÖZÇELİK volkan.ozcelik at gmail.com
Thu Mar 10 01:29:10 CST 2005


I minimized the problem further.


/<\/(?!p)[\s\S]*?>/ig   matches "</p>"

but neither

/<\/?(?!p)[\s\S]*?>/ig

nor

/<\/{0,1}(?!p)[\s\S]*?>/ig


matches "</p>". Imho the three expressions above should match "</p>".

So it is repetition-related. (the ? or {0,1} makes it fail.

Is it a problem with javascript regexp's lookahead assertion?

Can I find any further documentation and examples pointing to such
peculiarities?

Any suggestions, ideas wil be welcome.


---------- Forwarded message ----------
From: VOLKAN ÖZÇELİK <volkan.ozcelik at gmail.com>
Date: Thu, 10 Mar 2005 06:46:14 +0200
Subject: Regular Expression - Is it a bug or am I missing something
very obvious?
To: "thelist at lists.evolt.org" <thelist at lists.evolt.org>


Hi Community, hi RegExp lovers!

Regarding the following js reg exp.

/<\/?(?!(p))/ig;   (It's a simplified version of what I'm working on)

I think what it is supposed to do is:

1. match character < literally
2. match craracter / literally (zero or one times)
3. If there is a p immediately after do not match anything, just
rollback (i.e. negative lookahead)

so the code below

var strTest1 = "</p>";
var regEx1 = /<\/?(?!(p))/ig;
strTest1 = strTest1.replace(regEx1,"")
alert(strTest1);

should alert "</p>".
Well it doesn't. It alerts "p/>". Which means that it has found a match.

I need to swap things around to make it work:

var strTest2 = "</p>";
regEx2 = /<(?!(\/p|p))/ig;
strTest2 = strTest2.replace(regEx2,"")
alert(strTest2);

but why on earth do I need to make things unnecessarily complicated ?!
The first one should work at least according to my humble logic!

Is it a bug? Or am I missing a really simple thing?

Please help me before I split this PC with an axe.

TIA,
Volkan.


More information about the thelist mailing list