isValid return false for a valid URL

• Key: VALIDATOR-204
• Type: Bug
• Status: Closed
• Resolution: Fixed
• Priority: Major
• Assignee: Unassigned
• Reporter: Ben
• Affects Version/s: 1.3.0 Release
• Fix Version/s: 1.3.1 Release

Description

IsValid returns false for the following valid Url:

http://tech.yahoo.com/rc/desktops/102;_ylt=Ao8yevQHlZ4On0O3ZJGXLEQFLZA5

Stepping thru the code isValidPath return false for the above URL.

Test Code:

String[] schemes = {"http","https"}.
UrlValidator urlValidator = new UrlValidator(schemes) ;
urlValidator.isValid("http://tech.yahoo.com/rc/desktops/102;_ylt=Ao8yevQHlZ4On0O3ZJGXLEQFLZA5") ;

Comments

[Henri Yandell]

The pattern is:

/^(/[-\\w:@&?=+,.!/~*'%$]*)?$/

So the above url fails on the ; and the _. I don't see any reason why they're not being allowed. Looking at the ietf page, it seems to be fine there.

[Henri Yandell]

svn ci -m "Adding _ and ; to the list of allowed characters within the path component of a URL. _ is definitely allowed and ; is allowed but there are restrictions on its use and we'll need to refactor UrlValidator to properly understand how a ; may be used for path parameterization. This is with reference to #VALIDATOR-204."

Sending src/share/org/apache/commons/validator/UrlValidator.java
Transmitting file data .
Committed revision 453784.

[Henri Yandell]

This was done as a part of the day-job; so attaching a patch to signify that I passed it from my day-job role to my apache-role.