Re: Bug#109577: version 1.23 available
Hi release team,
What do you guys think of this ? It looks low-risk to attempt upgrading to
1.23 - maybe less risky than just taking the pieces of interest - but
hopfully bitkeeper should lower that risk anyway.
On Tue, Aug 28, 2001 at 01:11:21PM -0400, Theodore Tso wrote:
> On Tue, Aug 28, 2001 at 12:12:37PM +0200, ydirson@alcove.fr wrote:
> > Hi David,
> >
> > We've frozen the base system, as outlined in:
> > http://lists.debian.org/debian-devel-announce/2001/debian-devel-announce-200108/msg00011.html
> >
> > That means 1.23 can't go in, as it introduces many new features, and
> > thus lacks testing.
>
> Grumble. 1.23 was released on August 15th, and had I known that the
> base freeze was coming up on the 20th, I would have pushed you hard to
> get 1.23 into Woody.
I was away at the time...
> This is especially true since one of the more
> important improvements in 1.23 is the support for raw e2image files;
> this makes it much easier for folks who are reporting e2fsck bugs to
> be able to submit a meta-data only version of the filesystem for
> debugging purposes.
>
> The only other "risky" new features are the support for extended
> attributes and for external journals in e2fsck, but those are code
> paths which aren't even used if the filesystems don't have those
> features. So it would be very trivial to take 1.23 and disable
> support if you really think that would be a problem.
>
> > However it fixes a number of Debian-reported bugs, and I'll have a
> > look at backporting those simple enough to be integrated. Maybe
> > memory leaks as well if the fixes are localized enough.
>
> The fixes can be very easily localized; you can get the specific
> changes by viewing the source tree using bitkeeper:
>
> bk://thunk.org:5000
Cool. I have beta of bk 2.0 installed, I'll update it.
> If you're not willing to take 1.23 at this time, the following changes
> really ought to be merged in. (You can either get the patches from
> the BitKeeper tree, or I can mail them to you individually.) I really
> would strongly urge you to try find some way of moving to 1.23,
> though, since I'm not all that enthusiastic about supporting a 1.22
> "midway" codebase...
>
> - Ted
>
>
> ChangeSet@1.785, 2001-07-10 14:27:58-04:00, tytso@think.thunk.org
> Speed up journal recovery in userspace by avoiding a pointless
> unconditional ext2_flush() call. Also made ext2_flush() more
> efficient when the superblock is modified after the filesystem has
> been flushed, and make it clear the superblock dirty flag
> after doing the superblock/block group descriptor flush.
>
> ChangeSet@1.791, 2001-07-20 14:13:49-04:00, tytso@think.thunk.org
> unix.c (main): Add an explicit warning when the filesystem is left not
> completely fixed when e2fsck exits. (Addresses Debian bug #104502.)
>
> ChangeSet@1.793, 2001-07-20 14:25:23-04:00, tytso@think.thunk.org
> fsck.8.in: Add much more explicit language documenting how the
> fs_passno field in /etc/fstab is handled. (Addresses
> Debian bug #30833).
>
> ChangeSet@1.799, 2001-07-26 09:02:56-04:00, tytso@think.thunk.org
> pass1.c (e2fsck_pass1): Free ctx->block_ea_map at the end of pass 1.
> This avoids increasing the memory footprint of e2fsck by 10%!
>
> ChangeSet@1.834, 2001-08-09 05:41:29-04:00, tytso@think.thunk.org
> Add support for raw image files in e2image.
>
> ChangeSet@1.835, 2001-08-09 06:04:32-04:00, tytso@think.thunk.org
> Add support for specifiying - as the image file (to send the image
> file to standard out).
>
> Save the journal inode if it is internal to the raw image file.
>
> ChangeSet@1.838, 2001-08-10 19:13:11-04:00, tytso@think.thunk.org
> mke2fs.8.in: Slim down text describing the -O flag to remove
> text that erroneously implied that all features
> (including compatible flags) aren't compatible with older kernels.
> (This addresses some debian bug if I recall correctly, but I
> didn't remember to include the bug # in the message.)
>
> ChangeSet@1.842, 2001-08-13 10:58:41-04:00, tytso@think.thunk.org
> super.c (release_orphan_inodes): If the filesystem contains
> errors, don't run the orphan * list, since the orphan list
> can't be trusted.
> (Note this is critically importtant to avoid even worse filesystem
> corruption problems when you're using ext3 and the kernel
> finds that the filesystem contains inconsistencies.)
>
> ChangeSet@1.846, 2001-08-15 18:39:51-04:00, tytso@think.thunk.org
> fsck.c (check_all): Don't bother to interpret a device where the
> pass number is zero. (Addresses Debian bug #106696).
>
> ChangeSet@1.847, 2001-08-15 19:04:59-04:00, tytso@think.thunk.org
> fsck.c (main): Print a warning message if there are no devices to
> be checked. (Addresses Debian bug #107458.)
>
> ChangeSet@1.848, 2001-08-15 19:06:55-04:00, tytso@think.thunk.org
> fsck.8.in: Fixed error in synopsis of the man page.
>
> ChangeSet@1.849, 2001-08-15 19:17:37-04:00, tytso@think.thunk.org
> * tune2fs.c: Make sure that error messages are sent to stderr, and
> normal messages are sent to stdout. (Addresses Debian bug
> #108555.)
>
>
>
--
Yann Dirson <ydirson@altern.org> | Why make M$-Bill richer & richer ?
Debian-related: <dirson@debian.org> | Support Debian GNU/Linux:
Pro: <yann.dirson@fr.alcove.com> | Freedom, Power, Stability, Gratuity
http://ydirson.free.fr/ | Check <http://www.debian.org/>
Reply to: