On Tue, 17 May 2005, Jeff Garzik wrote:
>
> Please pull the master (HEAD) branch of
> rsync://rsync.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-2.6.git
Pulled.
> Review contents (diffstat/changelog/patch) attached to this email. I'm
> still new to git, so pull carefully. :)
My pull gets the same diffstat as you claimed, and everything looks good.
> Three git-related comments:
>
> 1) James Bottomley's git-changes-script is darned useful, for this
> ex-BitKeeper user. I've attached it.
It should _really_ be updated to do a better job, modern git can do much
better.
See the answer to the next question, since that's really the same thing:
> 2) What is the preferred way to generate a 'for Linus' diff? I used to
> BitKeeper's "repogca" feature to find the GCA for the diff.
You can use
"git-merge-base HEAD OTHER_HEAD"
to get the global common ancestor. However, that's pretty pointless, since
what you're after is really "what are the differences", and that's what
"git-rev-tree" gives you for any arbitrary points.
With modern git, what you do is something like this (totally untested,
but you hopefully get the idea):
local=.
remote=.
localhead=HEAD
remotehead=HEAD
#
# default to silent mode (changelog only),
# use "-p" to enable full patch information.
#
diff-tree-arg=-s
#
# Get the arguments
#
while true; do
case "$1" in
-R) shift;
remote="$1"
shift;;
-L) shift;
local="$1"
shift;;
-r) shift;
remotehead="$1"
shift;;
-p) diff-tree-args=-p
shift;;
*) localhead="$1"
break;;
esac
done
#
# Tell git about where it can find all the objects
#
export GIT_OBJECT_DIRECTORY=$local/.git/objects
export GIT_ALTERNATE_OBJECT_DIRECTORIES=$remote/.git/objects
# ..and get the heads
head=$(cat $local/.git/$localhead)
other_head=$(cat $remote/.git/$remotehead)
# What exists in remote-head but not in the local head?
git-rev-tree $other_head ^$head |
cut -d' ' -f2 |
git-diff-tree --stdin -v $diff-tree-args
and it should just DoTheRightThing(tm) modulo the inevitable
ObviousBugs(tm).
> 3) Note that my object database is not pruned. When I used
> git-pull-script to locally merge my libata-dev.git#misc-fixes branch
> into libata-2.6.git, it pulled all the objects in libata-dev. I was too
> slack to bother with pruning libata-2.6.git, knowing that eventually the
> other changesets will make their way upstream.
I do keep my trees pruned and I still run fsck religiously, so I just
force a prune after a pull. I _prefer_ to see clean trees, though - not
because it matters from a techncial perspective, but because it tells me
that the other side didn't have any left-over strange objects that might
have been intended to be included.
But your tree looked fine.
Linus
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[Index of Archives]
[Kernel Newbies]
[Netfilter]
[Bugtraq]
[Photo]
[Stuff]
[Gimp]
[Yosemite News]
[MIPS Linux]
[ARM Linux]
[Linux Security]
[Linux RAID]
[Video 4 Linux]
[Linux for the blind]
[Linux Resources]