On Mon, Nov 29, 2004 at 10:06:43AM -0600, John Goerzen wrote: > On Mon, Nov 29, 2004 at 03:39:32PM +0000, Steve McIntyre wrote: > > >In that case, we could have just kept the installer from woody and been > > >done with it, no? > > If necessary, I believe we could have done so, yes, _for some > > architectures_. In the end we didn't have to, as sufficient effort was > > found on all the different arches to make d-i happen everywhere. For > > several of the arches, we could have survived with woody boot floppies > > and let people boot/install from those and then upgrade straight to > I'd assume one could have kept the basic system and just replaced the > base tgz/packages with a sarge base. > > sarge. After all, that's one of the big attractions of debian - the > > upgradability. For active arches like i386 and powerpc, we needed > > desperately to move onto d-i to get newer kernels, better hardware > > support, etc. > I understand, but if it is the case that d-i is actually a blocker on > the sarge release (and I'm not sure that is true), then we could have at > least released sarge with the existing installer and support. We would > have been no worse off than with the current stable, and when d-i is > ready, we could make a point release of sarge that replaces the > installer only but changes nothing else. The "existing installer" used a kernel that is not being shipped in sarge, and would have installed this kernel, with all its vulnerabilities and lack of drivers, onto the installed system as well. As noted, no one could even be bothered to roll new b-f images to address the various kernel vulnerabilities that have been discovered over woody's lifetime. Perhaps you know differently, but I don't see any reason to believe getting b-f ready for sarge would have fared any better than keeping it up-to-date for woody did. -- Steve Langasek postmodern programmer
Attachment:
signature.asc
Description: Digital signature