On Monday 10 March 2008 06:24:44 Andrew Farris wrote: > > I don't know, thats why I said 'if'. It seems to me that most of the users > who posted here about the issue knew beforehand, or at least the tone of > their emails suggests that. > Actually, it is usually pretty difficult for me to figure out if an update will break something. For example, suspend seems to be eternally listed as "working," but several kernel updates have broken it. Suspend to disk is also listed as "working," but I have yet to actually observe this (I am told it is my graphics driver that is to blame; this is not documented, or wasn't the last time I checked). It is also unreasonable to expect everyone to know everything about every package on their system. My inbox is crowded with messages from ~3 active mailing lists; imagine if I tried to keep up with messages from the mailing lists for all of the software that I need working (~12 packages). It would be pandemonium. That's why we trust package maintainers to maintain our packages for us; that's also why we get unhappy when packages start regressing, and why we have updates-testing for packages that might break something or regress something, to help weed out bugs. My point is, putting a kernel with obvious and known problems into the non-testing updates may not be a wise idea. I understand that we are supposed to be bleeding edge, but that shouldn't require us to bleed (much). -- Benjamin Kreuter -- Message sent on: Mon Mar 10 09:51:02 EDT 2008
Description: This is a digitally signed message part.