[Resolved] preg_match_all - local vs. remote

xgab

A regex I have built is now working perfectly on Linux with PHP 4.3.3 when preg_match_all parses the html page locally, but does not work when I try to parse it remotely (in both cases using fopen).

My goal is to grab links from an online Web page.

<? 
$abrir = fopen("tribunam.html", "r"); /*This is working !!! */ 
/*$abrir = fopen("http://www.xxxxxxxxx.com", "r"); This is not working, I get an empty result*/ 

$leer = fread($abrir, 30000); 

preg_match_all("/(?i)<(H[1-6]|A)([^>]+HREF=\"(.*)\"[^>]*|[^>]*)>((.|\n)*)<\/\\1[^>]*>/U",$leer,$titular); ?>

PS: This message continues from my previous post.

Please help!

Natty_Dreadlock

hi,
maybe the php.ini settin allow_url_fopen is set to off (check usin phpinfo()) - u should set it to on.
hth

xgab

Thank you Natty_Dreadlock but... the fopen with http is working fine.

What I mean is that the regex within the preg_match_all function, behaves different when I get the web page locally than the web page remotely.

The Web

Did anybody experienced this before ???

Please help!

Natty_Dreadlock

u should compare the contents of $leer using local and remote fopen - they should be equal, since u r accessing the same html page, aren't u?

xgab

Yes, I am accesing the same html page.

I tell you that I compared the results of $leer using local and remote fopen and my surprise was that the fread does not read the links with the remote option!!!.

In both cases I can see text, IMG tags, etc., but the remote test has the links missing!!!. So preg_match_all would never get what I want, because the source info is missing!!.

With the local test I can see the entire page with links included.

This is very strange!!!, isn't it ???

Natty_Dreadlock

is your code available online somewhere so that i can see the results?? there has to be some obvious reason we have not taken into consideration yet.

xblue

I haven't checked it more carefully, but I suspect a simple \r\n-style line break might spoil everything, since it wouldn't be matched by
(.|\n)*

xgab

Bingo!!

A couple of hours ago I reviewed my information provider's online page once again, but this time in deep, byte by byte; and I discovered that the page had portions of XML that affected the fopen and the fread.

I requested my information provider to remove those XML portions and the page became a stardard readable HTML page.

So... thank God, this problem is resolved!

Thank you all for your very kind help!