Bug #778

Regular expression skips the first tag after the DOCTYPE in the DTD

Details

Submitted	2004-02-17 16:06 UTC
From	rshaull at simanage dot com
Assigned	schst
Status	Closed
Package	XML_DTD
PHP Version	4.3.4
OS	Windows 2000
Roadmaps	(Not assigned)

Comments

[2004-02-17 16:06 UTC] rshaull at simanage dot com

Description:
------------
The first <!ELEMENT is ignored in a DTD because the regular expression on line 127 of DTD.php pulls the contents of the first <!ELEMENT into the <!DOCTYPE [ element, so that the first match looks something like this:

DOCTYPE document [ <!ELEMENT document (page+)

I resolved it on my server by replacing '|<!([^>]+)>|s' with '/<!([^\[>]+)>/s'.

Please see http://dev.simanage.com/download/php-bugs/xml_dtd/001/ for a detailed report and tests.

Reproduce code:
---------------
Please visit this URL for full bug report:
http://dev.simanage.com/download/php-bugs/xml_dtd/001/

Expected result:
----------------
When I run the validator on pretty much an XML-DTD pair I get back 'invalid' because the first <!ELEMENT is not parsed into the XML tree that represents the DTD.

Actual result:
--------------
After applying my fix I get 'invalid' only for documents that are actually invalid and 'valid' for the rest.

[2004-05-17 23:32 UTC] rshaull at simanage dot com

Sorry about the unavailability. The page should be accessible now. I posted this report quite a while ago and had almost forgotten about it...

I think you are right (about "<!DOCTYPE [" not belonging in an external DTD). I don't write DTDs very often, and I guess I had the syntax confused with inline DTDs.

Well, thanks for checking into it.