- From: <bugzilla@wiggum.w3.org>
- Date: Tue, 09 Aug 2005 20:10:47 +0000
- To: www-validator-cvs@w3.org
- Cc:
http://www.w3.org/Bugs/Public/show_bug.cgi?id=1835 ville.skytta@iki.fi changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |WONTFIX Version|3.9.2 |4.2.1 ------- Additional Comments From ville.skytta@iki.fi 2005-08-09 20:10 ------- I'm inclined to reject the suggestion of making it possible to ignore robots.txt. If the server admin wants to allow the link checker, documentation how to achieve that exists. http://search.cpan.org/dist/W3C-LinkChecker/docs/checklink.html#bot Regarding username/password input, I don't see it feasible to implement much more than what's already in (ie. Basic (+maybe Digest, I don't remember now) auth proxying for the initial URL's realm). No matter what authentication method is used, it needs to be controlled where to send the authentication info (at the very least, based on URL(s)) during a link check, it cannot be sent everywhere due to security considerations. Regarding form-based login, the form field names where to add the username/password would be need to be made configurable. Also, the URL where to send the form (it might be different than the resource that one wants to check), whether to use GET or POST, whether the login form needs more parameters than just the user/pass, etc would need to be configured by the poor user. This is all theoretically doable, but I don't personally think we should be in that business at all, there are just too many "moving parts" in it.
Received on Tuesday, 9 August 2005 20:10:52 UTC