From kde-commits Tue May 16 20:44:47 2000 From: Dawit Alemayehu Date: Tue, 16 May 2000 20:44:47 +0000 To: kde-commits Subject: Re: kdelibs/kio X-MARC-Message: https://marc.info/?l=kde-commits&m=95852131704374 On Tue, 16 May 2000, you wrote: > CVS by adawit wrote: > > - Added password caching support for all io-slave using the kdesu daemon. > > This support is provided through the ui_server and available to any > > application or io-slave. To make use of this caching mechanism simply > > change your call to openPassDlg( head, user, password ) to the new > > overloaded function openPassDlg( head, user, password, hostname ). Note > > the hostname is used as a key to cache the password. > > Only thing that immediately comes to mind is the "what if" scenario of > "what if a given host has more than one password protected realm for a > given protocol?" For example, what if.... Fixed. The hostname was misleading ; so I have renamed it to "key" and kio_http now passes the value of the "realm" as a key instead of the hostname. > http://www.somedomain.com/mystats/ > and > http://www.somedomain.com/yourstats/ > > are both password protected, but use different passwords? I know that in > HTTP the 'basic' authentication uses a "realm" to identify where within the > host/site the password is to be considered valid for, but don't know if > that concept applies well (or at all) to other protocols. > > Will admit that I have not looked at your patch to determine whether it > handles any of that (or whether its handled elsewhere by KIO), but from > reading the snippet you've outlined above it was one of the first things to > come to mind. Can't wait to try it....if only I had enough time to do a > full build right now..... The above fix should take care of the issues with "http" clients. For other situations it has to be upto the io-slave or app programmer that uses the it to specify a unique key for now ... Regards, Dawit A.