preg_replace oddness

darthmahon

Hey,

Ok using the following little snippet to replace <a href"www">www</a> links in html to plain text links - this is for a mailing list.

$mail_format = str_replace("\r", "", $message2);

$mail_format = stripslashes($mail_format);

$mail_format = preg_replace("/<a href=\"(.*)\">(.*)<\/a>/i", "\\2 (\\1)", $mail_format);

$main = strip_tags($mail_format);

It's the third bit of the code, the preg_replace function that seems to be doing some odd stuff.

For example, if I have the following code:

Links - <a href="http://www.google.com">Google</a> or <a href="http://www.yahoo.com">Yahoo</a>.

I get this in my plain text emails using the above PHP:

Links - Yahoo ([url]http://www.google.com[/url]">Google or

If I put a break after "or" I get perfect links:

Links - Google ([url]http://www.google.com[/url]) or 
Yahoo ([url]http://www.yahoo.com[/url]).

It seems as though when the two <a href> links are on the same line, it messes up the code?

Any ideas?

Cheers,
Chris

xblue

a) why it grabs both links: matching always is greedy unless you tell it not to be, so you might want to use .? instead of . for ungreedy matching.

b) why it did not with a line break: the . does match any character except line breaks by default. you'd use the s modifier to change this behaviour, or

<a href="">Multiline
link</a>

would fail to match, too.

Bloodwine

change /i to /gi for global replace (otherwise it stops after the first replace

darthmahon

/gi gave me errors, had to use the following in the end:

$mail_format = preg_replace("/<a href=\"([^\"]*)\">(.*?)<\/a>/i", "\\2 (\\1)", $mail_format);

Cheers guys,
Chris

Weedpacket

Originally posted by Bloodwine
change /i to /gi for global replace (otherwise it stops after the first replace

No it won't; [man]preg_replace[/man] is automatically global. You're thinking of another language.