On Jun 26, 2013, at 9:52, Fischlin Andreas wrote: > > > On 25/06/2013, at 19:26 , Christiaan Hofman wrote: > > > On Jun 25, 2013, at 19:19, Colin A. Smith wrote: > > On Jun 25, 2013, at 19:04, Christiaan Hofman > <cmhof...@gmail.com<mailto:cmhof...@gmail.com>> wrote: > > On Jun 25, 2013, at 17:17, Fischlin Andreas wrote: > > And the mode is still activated via having the user ID pref either the empty > string (IP# authentication mode) or the user ID as offered in the usual > credential dialog as default (user ID/password authentication mode)? > > Regards, Andreas > > Not sure what you mean. Which mode, which credential dialog, and is there a > difference between no user ID and an empty string? > > I guess so. If plist contains <UserID></UserID>, then there is an empty > string. If no tag <UserID> is present in the plist file, there is no user ID > defined. > > > Christiaan > > In the code I wrote, password authentication is only done if both the > username and password are both not nil. IP authentication is done otherwise. > I don't know what Christiaan's code hands to my code under different > circumstances. > > Colin > > My code is doing the exact same thing, there is no difference in how that > works. The only difference is that they are set in the UI rather than a > hidden pref. > > ISI WOS offers two methods of authentication: 1) via IP# of the user's Mac > (default, and until recently the only method) or 2) via UserID and password > (new mode). > http://search.isiknowledge.com/esti/cxf-wsclient-demo/docs/soap/WOKMWSAuthenticate/index.html > > So, Christiaan, what triggers the setting via the UI? If you ignore > completely those hidden preferences, as your statements may imply, then the > UI is never going to be used, and only authentication method 1 is going to be > used. If, however, Colin's algorithm is still going to see whether the two > hidden preferences are set and only then switches to method 2, then the two > hidden preferences need nevertheless to be set. For security reasons perhaps > with something silly like "UserID" and "Password". But, I have another > suggestion... > > Given all this I suggest the following solution: > > 1) The hidden pref BDSKISIPassword is removed > 2) The presence of the hidden pref BDSKISIUsername is the only condition > needed to trigger method 2 of authentication (I would prefer empty string to > be supported) > 3) The UI dialog asking for the password uses for the UserID the string from > the pref BDSKISIUsername > > This would resolve all issues and I guess would make the use of this new > authentication method maximally convenient. > > Cheers, > Andreas
This does not make sense to me, at least in the context of the actual code. What UI and UI dialog are you talking about? The only UI I see in the code is the one I added, i.e. the search group setup sheet. There you can enter a password and user name. This is just a *replacement* of the hidden prefs, and they are used in *exactly* the same way, only more transparently (this also means that method #2 will be used when you set them). Hidden prefs are generally bad, I would like to avoid them if I can, and here we can. Christiaan ------------------------------------------------------------------------------ This SF.net email is sponsored by Windows: Build for Windows Store. http://p.sf.net/sfu/windows-dev2dev _______________________________________________ Bibdesk-develop mailing list Bibdesk-develop@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bibdesk-develop
