Re: Freenet 0.7.5 build 1397 oo at lkXpu0~CDV6dh0Idyw4MBwkUSgn~h~Bs3qqVXYOXSaY wrote : > toad-notrust at h2RzPS4fEzP0zU43GAfEgxqK2Y55~kEUNR01cWvYApI wrote: > >> oo at lkXpu0~CDV6dh0Idyw4MBwkUSgn~h~Bs3qqVXYOXSaY wrote : >>> oo at lkXpu0~CDV6dh0Idyw4MBwkUSgn~h~Bs3qqVXYOXSaY wrote: >>> >>>> toad-notrust at h2RzPS4fEzP0zU43GAfEgxqK2Y55~kEUNR01cWvYApI wrote: >>>> >>>>> oo at lkXpu0~CDV6dh0Idyw4MBwkUSgn~h~Bs3qqVXYOXSaY wrote : >>>>>> toad-notrust at h2RzPS4fEzP0zU43GAfEgxqK2Y55~kEUNR01cWvYApI wrote: >>>>>> >>>>>>> Freenet 0.7.5 build 1397 is now available. Please upgrade ASAP. This >>>>>>> build >>>>>>> enables new load management, and will be mandatory noon tomorrow. There >>>>>>> are also >>>>>>> a few minor bugfixes. There may be problems, please let me know how it >>>>>>> goes, I >>>>>>> will try to fix any remaining bugs (that months of testing on testnet >>>>>>> haven't >>>>>>> found) as quickly as possible without introducing lots of new bugs! >>>>>> >>>>>> Inserts are still broken, though not as bad as 1389+ >>>>>> About 30% of all downloads get stuck near the end. >>>>>> Often a lot of inserts of specific users are affected. >>>>> >>>>> Is this specifically a problem with SSKs? >>>> >>>> AFAICS there's no difference between SSK and CHK. >>> >>> Btw IMO 'recently failed' is very bad for download requests and should be >>> disabled for one of the upcoming builds to actually see its effects. >> >> And constantly spamming the network with SSKs is bad for the network too, so >> we >> should turn off Frost, FMS and Freetalk for one of the upcoming builds to see >> its effects ... oh, we can't. Fuck that then. > > It has worked very well for a long time without 'recently failed'.
That has nothing whatsoever to do with the recent problems, either the 1389 regression (which is now fixed), or the other problems with NLM. > As soon as it was introduced it reduced download speed to 50% of values > before and message insertion with Frost takes forever. > 'recently failed' was no step forward, toad. It has never been evaluated > whether there is any improvement. Just because it is supposed to work does > not mean it actually works. > If you introduce something it is necessary to verify that it works as > expected. If things turn bad it must be removed unless it can be fixed. And we never, ever, ever, have enough data to evaluate a single build, even on the simplest metrics (see the push-pull tests). I could write a plugin to get more data, but digger3 promises to do it eventually and anyway I don't have time given the remaining funding and unlikeliness of getting more. And it's always been this way! Our whole business model forces me to just do things and not evaluate them! Anyway you haven't addressed my point - what would you do about chat applications spamming the network with requests? Bear in mind that this will scale extremely badly. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 198 bytes Desc: This is a digitally signed message part. URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20110825/59e8394e/attachment.pgp>
