On Wed, Dec 2, 2009 at 8:06 AM, Dan Pidcock <dan.pidcock@xxxxxxxxxxxxxx> wrote: > 2009/12/2 Alexander von Gluck <kallisti5@xxxxxxxxxxx> >> Just a quick FYI, as per Renes' comments the new nightly image no longer >> suffers from the memory explosions. > > It sounds like it would be useful to have a mechanism to mark a build as > bad/unusable to avoid other people also downloading the same build and > hitting the same problems when they are already known about. This would need > to show up in the relevant place in the haiku files list, rather than be at > some other location such as the wiki. Well, they are nightlies after all - there shouldn't really be much expectation from them at all IMO. Once you do this, then you have to define what "bad" means - for example, if a new nightly causes a regression that only affects nvidia chipsets, would that be "bad"? What if the same change improves intel chips as a result? It's just part of utilizing nightly builds - you get some good and some bad. It's too bad we don't yet have good a good release cycle for this kind of stuff. - Urias