Regular Expressions Over Multiple Lines

rgraves

Hi,

I'm developing a php tool to convert thousands of pages at our company to a new structure. I want to match a piece of html code like this using regexp:

The problem I'm encountering is that some of the HTML code has line breaks in it and I can't be 100% certain where they will occur.

e.g.
<a
href="http://www.mydomain.com/">

e.g.
<a href=
"http://www.mydomain.com/">

I need a regular expression that can match the above code fragments and still be successfull if there is a line break anywhere within the code (i.e. I want it to ignore line breaks and focus on the code itself).

Any ideas?

For your reference, this is my current regexp:
"/<a(.)href=['|\"]{0,1}http:\/\/www.mydomain.com\/([^>])>/isU"

rgraves

I tried the /m modifier and it doesn't appear to work.

maxpup979

[[:space:]] matches any space, tab, newline, or carriage return

rgraves

Since there are so many pages developed by so many people in our organization (and in the past, by different WYSIWYG editors), I can't be certain where a new line may occur. What I'd like is an expression that will focus on the code itself and ignore any new line characters it encounters.

maxpup979

this is quite the stumper I think. the lack of any sort of standardization does not lend itself well to a regular expression. What you may want to do is instead, just look for the links, triggered by the <a (or <A I suppose), and run a couple of different things across it. like str_replace any line breaks with spaces and such. THEN run your regex, and see what happens. ugh. But then again, some regex wizard may have a better way...🙂

xblue

The m-modifier was the wrong one, but you seem to have the s modifier set already which makes the dot match linebreaks. However there are a few issues with your pattern that look like errors to me:

['|\"] <- this does not match apostroph or quote, it does match apostroph, vertical bar or quote
www.mydomain.com <- these dots need to be escaped or they match any character

And some pointers:
- short form of {0,1} is ?
- you don't need to use the slash as pattern delimiter. Using e.g. # instead saves you from needing to escape slashes within the pattern.

rgraves

Thanks for the tips xblue!

rgraves

Anyone have any ideas as to how I can have the regexp ignore line breaks anywhere in the code and just focus on the code itself? I don't want to have to test repeatedly for .* to handle this.

I had thought about removing line breaks via php from the string, however I need to output the revised string to a file, so doing that would remove all the formatting on the HTML page.

toplay

As xblue mentioned, the 's' modifier takes care of what you're asking about. Basically, you don't have the pattern correct and that's probably why you think want to blame the newlines.

I have to run, but follow xblue's advice and/or search the regexlib.com site for a pattern to use:
http://www.regexlib.com/Search.aspx?k=href

Look in my signature line for other links to help you understand regex patterns.

Good luck.

rgraves

Thanks toplay... some very handy regex links in your signature!