Correctly handle ranges that don't terminate at the end of the file.
Also reject invalid ranges as described in RFC-2616.
Thanks to Kim Biesbjerg for the initial patch.
Fixes#3914
According to NANPA, there are area codes and exchanges that are invalid or reserved and not available for assignment. This would make those phone numbers invalid. For example:
- area-codes: 555, N11, N9X, 37X, 96X
- exchanges: N11
- numbers: 555-0100 thru 555-0199
You could also say 555-1212 is invalid as it is assigned as information for the US.
I changed the tests where 555 as an area code was being used as the assertion will give it a false on the area code and not the exchange as intended.
Trim off webroot/index.php when determining base and url.
Trimming off index.php from url and webroot/index.php from base url allows the correct values to be created when a path contains index.php in it.
Fixes#3318
Fixes https://cakephp.lighthouseapp.com/projects/42648-cakephp/tickets/3318
It seems fixing this in the htaccess file(s) isn't going to work even though a url rewriting based solution was more clean. On the plus side this works for any web server.
If a url is called with "index.php" in it then the CakeRequest swallows this part and fixes the path. Any linked url from the requested page will have a clean url. Thus after following one of these urls this problem is gone anyway.
Some code docblock improvements to CakeRequestTest.php
Added test case for fix
Also now you can call just index.php even if you have url rewriting enabled
When the first path segment matches the base path an incorrect URL was
generated. Trimming slashes off makes Router normalize the URL correctly
as the leading / implies that the base is already prepended.
Fixes#3897
The function allows ://example.com/file.ext but was treating //example.com as cake-relative URL. The updated regex matches URI schemes as defined in RFC2396. Will passthru any of these formats:
* Starts with a valid URI scheme (javascript:, https:, itunes:, ftp:)
* Starts with a '#'
* [NEW] Starts with a '?' which may be meaningless, but is as valid as starting with '#' (RFC1808)
* starts with //, or :// (:// is not technically valid, but included for compatibilty)
