> This is on 604, so there's no reason why it should be working > perfectly, but there's two obvious issues: > > 1. Trying to access a DBR site that DNFs, will display the > "The request followed a Date Based Redirect ... click here > for an earlier dated version" message - but when it says > "Technical details of this DBR: increment: 86400 offset:0" > wouldn't it be better (for the user) to say something like "This > earlier version is dated 00:00 October 30 2002". Actually > this highlights > another issue [1]
Good idea, I meant to remove this technical detail once it was working properly. > 2. The example I tried was > SSK at IuWJ~Zc7IZdfcCQHmHjp4-lBmksPAgM/RIAAsBane// > . Clicking on the "earlier dated version" link takes me to > "...?date=20021030-11:22:27. 11:22:27 is the local PC time, > rather than > anything to do with DBRs. - is there any reason it doesn't link to > "...?date=20021030-00:00:00" instead? (Or is that not how > the DBR offset > is used?) This is intended behaviour, it requests the page for the same date minus the page increment (normally 86400). This will have the intended behaviour, as if you were looking at the page at the exact same time, but on the previous day (or current time - increment if not a day). It also has the benefit of possibly highlighting any timezone problems as the calculation performed is obvious and will show any errors all the time rather than just on boundary cases. Once any timezone issues are checked-out, the time could quite easily be made sensible, but it needs a bit of fiddling with the offset to make sure it works for all sites. > oh, and > 3. Suppose the "earlier dated version" also DNFs. Another "cannot > find, try earlier dated version" page appears, but the link takes you > again to the SAME day. I.e. if it says "?date=20021030-11:27:30" and > you click on the earlier dated version link, it takes you to > "?date=20021030-11:27:32" (same date, current time) rather than > "?date=20021029-11:27:32" or "?date=20021029-00:00:00" or > //something// from the 29th. This is not the expected behaviour, it should do as you expect, drop back to a day earlier - I did test this and it worked for me. What timezone are you in? Its interesting that it changes time... it should keep the same time as that of the original request, so the bug's probably that its taking current time somewhere along the line. Will take a look and see what's up. > [1] - Are there any plans (current or proposal) for > localising FProxy? I expect this would increase the > massmarket appeal of Freenet significantly (we've already had > several support calls from non-English speakers requesting > help and configuration instructions in their own language). > Localising FProxy would not even be particularly difficult, > just a simple string table lookup on all strings and external > localised dictionary files, with presumably the American > English dictionary coded into the .jar. I'd do it if I knew > Java, but then it's so easy I doubt I'd even need to know > Java to do it. Obviously there's then the issue of providing > the translated strings. No biggie, that's what > babel.altavista.com is for. Hmm, interesting idea, not sure how much effort it would be (I've never looked at Java's internationalisation features), but I've put it on my page-long list. -- Mat Burnham _______________________________________________ devl mailing list devl at freenetproject.org http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/devl
