On Mon, May 02, 2005 at 12:44:26AM -0400, Theodore Ts'o wrote: > On Mon, May 02, 2005 at 05:04:42AM +0200, Jeroen van Wolffelaar wrote: > > In order to get debian-installer and normal tools & sources in sync for > > sarge, which is a requirement, I'm going to upload 1.35-8 as > > 1.35-8sarge1 into testing. You don't need to do anyting. If this upload > > fails for whatever reason, we're going to reupload 1.35-8 as it was into > > testing. > > Sorry for the late warning. If there is any specific reason why the > > version currently used for debian-installer (1.35-8) is a bad thing to > > have in sarge in general, please contact debian-release@lists.debian.org > > asap. Currently 1.35-6 is in Sarge. > In previous discussions with Steve, I thought we were going to push > e2fsprogs 1.37-2 into Sarge, per discussions recorded debian bug > #295422? > If we the decision of the Debian release team is to use 1.35-8 > instead, that's fine. There are a number of bugs that are not fixed > in 1.35-8, including: > * E2fsck will now recover from a journal containing illegal blocks. > * Fix a double-free problem in resize2fs. (Red Hat Bugzilla #132707) > * Make sure e2fsck doesn't crash if /proc/acpi/ac_adapter does not > exist > * Make sure that we don't write garbage when writing a large inode. > * Add compatibility in e2fsck for filesystems created by Fedora Core 3 > The last one means if we don't use a newer version of e2fsprogs, > Debian stable will be incompatible with ext3 filesystems created by > newer distributions, including FC3, RHEL 4 and (probably) the upcoming > SLES 10. > But heck, sarge isn't even going to have the X.org server, so it might > as well live up to its billing of Debian obsolete from the moment it > was released. :-) Yeah, the bugs you mention above are pretty important to get in, so I'm going ahead and pushing in e2fsprogs 1.37-2. This version has been used in d-i daily builds for 22 days now, in addition to the testing it gets in unstable, so I think the risk of d-i breakage here is low enough to be justified; Joey, please smack me if you disagree. I've also personally verified that mke2fs doesn't segfault on 64-bit archs now, and creates usable filesystems instead; so I think it's best to have 1.37-2 in, I just apparently lost sight of the fixed 1.37-2 along the way. Sorry for putting you to the trouble of that upload to t-p-u, J. :) -- Steve Langasek postmodern programmer
Attachment:
signature.asc
Description: Digital signature