0005220: url parser makes a few mistakes - MantisBT - Signup temporarily disabled due to spam

ID	Project	Category	View Status	Date Submitted	Last Update

0005220	mantisbt	bugtracker	public	2005-02-07 22:22	2009-06-26 11:59

Reporter	iamcal	Assigned To	grangeway
Priority	normal	Severity	minor	Reproducibility	always
Status	closed	Resolution	unable to reproduce
Product Version	0.19.2

Summary	0005220: url parser makes a few mistakes
Description	it doesn't allow [?=&] in urls, making urls with querystrings break it doesn't allow capital letters in urls it gets confused when urls have '@' symbols in them, after the first slash the attched patch fixes these issues
Tags	No tags attached.
Attached Files	core__string_api.php.diff (963 bytes) --- core/string_api.php.orig 2005-02-07 12:23:52.725125000 -0800 +++ core/string_api.php 2005-02-07 12:22:54.287625000 -0800 @@ -292,7 +292,7 @@ return $p_string; } # Find any URL in a string and replace it by a clickable link - $p_string = preg_replace( '/([http\|irc\|ftp\|https]{2,}:\/\/([a-z0-9_-]\|\/\|\@\|:{0,1}\.{0,1}){1,})/', + $p_string = preg_replace( '/([http\|irc\|ftp\|https]{2,}:\/\/([a-zA-Z0-9_-]\|\/\|\@\|[?=&]\|:{0,1}\.{0,1}){1,})/', '<a href="\1">\1</a> [<a href="\1" target="blank">^</a>]', $p_string); @@ -304,7 +304,7 @@ # when applied to data that has already had entities inserted. # # bpfennig: '@' doesn't accepted anymore - $t_atom = '[^\'@\'](?:[^()<>@,;:\\\".\[\]\000-\037\177 &]+)'; + $t_atom = '[^\'@\'](?:[^()<>@,;:\/\\\".\[\]\000-\037\177 &]+)'; # In order to avoid selecting URLs containing @ characters as email # addresses we limit our selection to addresses that are preceded by: core__string_api.php.diff (963 bytes)

grangeway 2005-02-08 18:28 reporter ~0009260	This issue has already been fixed I believe - can you try http://cvs.sourceforge.net/viewcvs.py/checkout/mantisbt/mantisbt/core/string_api.php?rev=1.64 Feel free to reopen if not. Paul

grangeway 2005-02-08 18:31 reporter ~0009262	@@ -304,7 +304,7 @@ when applied to data that has already had entities inserted. `# # bpfennig: '@' doesn't accepted anymore` $t_atom = '[^\'@\'](?:[^()<>@,;:\\".[]\000-\037\177 &]+)'; $t_atom = '[^\'@\'](?:[^()<>@,;:\/\\".[]\000-\037\177 &]+)'; And actually, this might still need doing.

jelv 2005-04-15 11:26 reporter ~0009820	I lifted the latest version of the change from the CVS and patched it in to 19.2 - it fixed a problem I had with a URL with ? and &. However still had a problem with a URL of http://usertools.plus.net/@gateway which gets changed to a mailto The change in note 9262 is not in CVS - when I applied that it fixed the problem.

jelv 2005-05-10 06:32 reporter ~0010058	I think there may still be a problem with email addresses. I have colon space preceeding the email address for example: abc@domainname.com. The href produced has the leading space included as part of the mailto e.g. mailto:%20abc@... If I am right it will happen in this note.

jelv 2005-05-10 06:35 reporter ~0010059	I was right you can see the problem in the last note. But there appears to be another problem as well - where did the "/>" come from?

stefanb 2005-05-25 09:06 reporter ~0010222	The "/>" is the rest of a line break tag. If you hover your mouse over the Link in 0005220:0010058, so can see the following link target in your status bar: mailto: abc@domainname.com.<br

jelv 2005-06-07 09:00 reporter ~0010386	This bug ought to have a relationship with 0004439 added.

grangeway 2009-05-23 09:47 reporter ~0021883	The url parsing functionality has been rewritten since 2005/6 ;/ Paul