need help for a regex

Anon

I'd like to write a regex which will convert server-relative urls to script calls:

a href="/directory/file.man"

should become:

a href="/basedir/script.php?lang=$current_lang&file=/basedir/directory/file.man"

More complete explanations, including the script which should contain the line:

http://www.tarastar.f2s.com/

Thanks for any help : )

Tara

Anon

Hey Tara:

Try the following.

$current_lang="en";
$message = "<A HREF=\"/directory/file.man\">This is my link</A>";

$newtext = preg_replace("/<A HREF=\"([^\"])\"/", "<A HREF=\"script.php?lang=$current_lang&file=/basedir\1\"", $message);

This should output:

<A HREF=\"script.php?lang=en&file=/basedir/directory/file.man\"> This is my link</A>";

Hope that helps!

Chris King

Anon

I'm not going to write "the line" for you. But, I'd break the problem out rather than look for a single, big-and-nasty line that does it all.

1) write a line or two that strips off the "<a href=" part.

2) A line or two that then strips off the ">" at the end.

3) A line or two that takes what's left and puts it into the final format.

Most problems are fairly easy if you break them up!

-Ben

PS: I like ereg_replace...

Anon

this isn't to be picky... ; ) - but my script has to parse a "big" file with a certain number of urls in it.

the file is read in a variable, and my regex (p or e, I don't care!) should work on that variable.

Are you suggesting that I chop my huge variable into little bits and then stick them back together? I'm not against the idea, but it sounds more complicated than writing a nasty line...

: )

thanks a lot for your time!!

Anon

Thanks Chris,

one question: I don't think the variable I'll be working on (ie, my $message - $output in my script) has the double quotes escaped.

The variable was generated by reading a file into it.

I unfortunately don't understand your expression enough to know what it would look like if the quotes aren't escaped. Could you please explain that bit to me?

: )

Tara

Anon

Of course 😉

Ok, lets look at the expression again and I'll walk you through it (actually, there was one minor mistake that I would like to correct):

$newtext = preg_replace("/<A HREF=\"([^\"]*)\"/", "<A HREF=\"script.php?lang=$current_lang&file=/basedir\1\"", $message)

Now, the actual expression of course is "/<A HREF=\"([^\")\"/" .. the reason that all of the quotes are escape is because the entire expression is enclosed in quotes, so to be correct in parsing any quote we use we have to escape. Now, the first / means the start of the expression. Then we match the text <A HREF=" (remember, we are escaping the quote, so when we are matching this is what it is actually looking for.) Next, we put the following if parentheses to pull it out. Now, we want to grab everything that is not a quote (") (get everything between the quotes) and pull it out. then, we say we have another quote (") and the last / ends the expression. the one minor fix is placing an i at the end to make it case insenitive (so we match <A HREF as well as <a href) ..

The same escaped quotes in the second part are true as well. So, it ending, the line you should use it:

$newtext = preg_replace("/<A HREF=\"([^\"]*)\"/i", "<A HREF=\"script.php?lang=$current_lang&file=/basedir\1\"", $message)

... I don't know how you plan on handling absolute vs. relative hrefs (i.e. /directory/file vs. subdir/file) but that would need a little different approach, but I'm not sure it's necessary. If you have any other questions, feel free to ask 🙂

Chris King

Anon

hmmmm ok : )

a) doesn't pregi_replace exist?

b) wouldn't preg_replace("/<a href=\".*?\"/") also match everything until the second quote? ("." being non-greedy....)

c) what does that \1\ do in the second part of the expression?

All the urls will be server-relative... : )

Anon

Ok, one by one 😉

a) no, the way I described is how do you case-insenitive perl regular expressions

b) that should work as well, but you must remember that . is ALWAYS greedy, the only reason it is inverted here is because you placed the question mark afterward. Just remember to put parentheses around it \"([.]?)\" .. that will be explained in

c) the \1 puts the url that we pulled out in the regular expression \"([^\"]*)\" .. you need to understand what the regular expression will return when you are matching. \0 will always be the entire matched string, and everything there on (\1, \2) will be anything you told it to capture by putting it in parentheses, in the order they occur in the regular expression. This is actually way out of the scope of this response, just trust me that it works 😉

Hope that Helps..

Chris King

Anon

"Are you suggesting that I chop my huge variable into little bits and then stick them back together?"

Oh, no no no! Don't break up the data, break up the PROCESS!

$input has the stuff.

1) do a single, simple regex against $input.

2) Do the next, simple regex against input.

3) Do the final regex against input.

return $input.

When you are climbing a cliff, you don't climb the whole cliff in a single step. Instead, inch forward, step by step, until you are there.

Anon

thanks!

www.tarastar.f2s.com/orange/script.txt (though the demo won't work, my server has php3...)