[VFS-325] Bad handling of hashs (#) in file names when walking a file tree using findFiles()

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.0, 2.0
• Fix Version/s: 2.1

Environment

Windows Seven, JDK 1.6 64 bit

• Assignee: Unassigned
• Reporter: Nicolas Guillaumin
• Created: 19/Oct/10 23:24
• Updated: 27/Jul/16 21:04
• Resolved: 24/Sep/11 08:08

Description

Consider a local directory tree containing files with hashs in their name, such as test-hash-#.txt.
When walking the tree using FileObject.findFiles(), the file is correctly found and returned, but it's URL is truncated to the #: test-hash-

• Calling file.getURL().toString() returns file://my/dir/test-hash-
• Calling file.toString() returns the correct URL file://my/dir/test-hash-#.txt
• For the sake of testing, calling new URL("http://my/file/with/hash-#.txt").toString() returns http://my/file/with/hash-#.txt (It's not an java.net.URL problem)

I think file.getURL().toString() should return test-hash-#.txt, otherwise caller have to rely on file.toString() to retrieve the URL of the file, which is probably bad.