[alsa-devel] GIT: something wrong with sound-2.6/devel?
Hi Takashi.
Is there something wrong with your devel branch or is it me?
rene@7ixe4:~/src/linux/7ixe4$ git fetch tiwai remote: Counting objects: 1695, done. remote: Compressing objects: 100% (418/418), done. remote: Total 1418 (delta 1182), reused 1222 (delta 1000) Receiving objects: 100% (1418/1418), 275.35 KiB, done. Resolving deltas: 100% (1182/1182), completed with 170 local objects. From git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6 ! [rejected] devel -> tiwai/devel (non fast forward) ! [rejected] dma-fix -> tiwai/dma-fix (non fast forward) 3e14b50..9432484 for-linus -> tiwai/for-linus ! [rejected] master -> tiwai/master (non fast forward) 481c534..50515af upstream -> tiwai/upstream rene@7ixe4:~/src/linux/7ixe4$ git checkout -b head v2.6.26 Switched to a new branch "head" rene@7ixe4:~/src/linux/7ixe4$ git merge tiwai/devel Auto-merged sound/pci/ac97/ac97_patch.c Auto-merged sound/pci/emu10k1/emu10k1_main.c Auto-merged sound/pci/hda/patch_analog.c Auto-merged sound/pci/hda/patch_realtek.c CONFLICT (content): Merge conflict in sound/pci/hda/patch_realtek.c Auto-merged sound/pci/hda/patch_sigmatel.c Automatic merge failed; fix conflicts and then commit the result.
It's no doubt trivial to fixup, but I believe this should've merged cleanly? (the master branch from where I'm branching off head is an undisturbed linus tree).
Rene.
On Sun, 20 Jul 2008, Rene Herman wrote:
Hi Takashi.
Is there something wrong with your devel branch or is it me?
I think that you have to recommit your patches on top of new Takashi's tree, because Takashi rebased his tree on top of Linus's tree.
Anyway, I just synced git.alsa-project.org to Linus's and Takashi's trees a few minutes ago, too.
Jaroslav
----- Jaroslav Kysela perex@perex.cz Linux Kernel Sound Maintainer ALSA Project, Red Hat, Inc.
On 20-07-08 21:08, Jaroslav Kysela wrote:
On Sun, 20 Jul 2008, Rene Herman wrote:
Hi Takashi.
Is there something wrong with your devel branch or is it me?
I think that you have to recommit your patches on top of new Takashi's tree, because Takashi rebased his tree on top of Linus's tree.
Thing is, I'm rebuilding my "head" branch here and there wasn't anything local in it yet. I know that his sound-2.6 is rebased at releases (the devel branch; the master branch more frequently) so I was trying to pull that into a clean linus 2.6.26:
(branch master is an undisturbied linus tree)
$ git checkout -b head v2.6.26 $ git fetch tiwai $ git merge tiwai/devel
As far as I'm aware, this should've merged cleanly.
(next step here is pulling in a few local branches into "head", such as Krzysztof's wss branch and other local stuff but that's after the above)
Anyway, I just synced git.alsa-project.org to Linus's and Takashi's trees a few minutes ago, too.
Yes, I tried that to
rene@7ixe4:~/src/linux/7ixe4$ git remote show alsa * remote alsa URL: git://git.alsa-project.org/alsa-kernel.git Tracked remote branches for-linus linux-2.6 master
However, while a "git pull alsa master" merges cleanly, it merges _all_ of current linus' GIT tree onto me!
This I definitely do not want at the moment. I want to be running 2.6.26, with just a selected set of changes, not just whatever changes Linus decided to merge. One of the changes I want is current ALSA, but as far as remote tracking branches go, nothing else...
Furthermore, given that alsa/master is rebased frequently, I cant use it anyway as a downstream person. If I'd want to stay current, I'd need to rebuild the entire head branch _every_ time I'd want current ALSA since I can't just pull alsa/master again -- after a rebase it clashes with what's already there in my tree.
That's basically the trouble with rebasing: your downstream can't sanely use your tree. Takashi said his devel branch would not be rebased during a single -rc cycle so I'm using that one...
Rene.
On 20-07-08 21:08, Jaroslav Kysela wrote:
On Sun, 20 Jul 2008, Rene Herman wrote:
Is there something wrong with your devel branch or is it me?
I think that you have to recommit your patches on top of new Takashi's tree, because Takashi rebased his tree on top of Linus's tree.
Hmm. If Takashi did not just now change something, then something that I don't understand is happening since right now it's working fine:
rene@7ixe4:~/src/linux/7ixe4$ git remote show tiwai * remote tiwai URL: git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6 Tracked remote branches devel dma-fix for-linus master upstream rene@7ixe4:~/src/linux/7ixe4$ git checkout -b alsa v2.6.26 Switched to a new branch "alsa" rene@7ixe4:~/src/linux/7ixe4$ git pull tiwai devel Updating bce7f79..e0bf09b Fast forward
[ ... ]
... and a clean merge. Takashi, if you did -- it's fixed.. :-/
Rene.
On 20-07-08 22:04, Rene Herman wrote:
On 20-07-08 21:08, Jaroslav Kysela wrote:
On Sun, 20 Jul 2008, Rene Herman wrote:
Is there something wrong with your devel branch or is it me?
I think that you have to recommit your patches on top of new Takashi's tree, because Takashi rebased his tree on top of Linus's tree.
Hmm. If Takashi did not just now change something, then something that I don't understand is happening since right now it's working fine:
rene@7ixe4:~/src/linux/7ixe4$ git remote show tiwai
- remote tiwai URL: git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6 Tracked remote branches devel dma-fix for-linus master upstream
rene@7ixe4:~/src/linux/7ixe4$ git checkout -b alsa v2.6.26 Switched to a new branch "alsa" rene@7ixe4:~/src/linux/7ixe4$ git pull tiwai devel Updating bce7f79..e0bf09b Fast forward
[ ... ]
... and a clean merge. Takashi, if you did -- it's fixed.. :-/
But you didn't... the "manual" git fetch tiwai/git merge tiwai/devel still fails.
It probably has something to do with this:
rene@7ixe4:~/src/linux/7ixe4$ git fetch tiwai From git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6 ! [rejected] devel -> tiwai/devel (non fast forward) ! [rejected] dma-fix -> tiwai/dma-fix (non fast forward) ! [rejected] master -> tiwai/master (non fast forward)
I'll go ask on the GIT list what this might be about...
Rene.
On Mon, Jul 21, 2008 at 10:45:29AM +0200, Rene Herman wrote:
rene@7ixe4:~/src/linux/7ixe4$ git fetch tiwai From git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6 ! [rejected] devel -> tiwai/devel (non fast forward) ! [rejected] dma-fix -> tiwai/dma-fix (non fast forward) ! [rejected] master -> tiwai/master (non fast forward)
I'll go ask on the GIT list what this might be about...
This means that the git remote is configured to only pull fast forwards from that repository. Add a + to the 'fetch =' for it in .git/config and it'll track non-linear updates like you get when the remote rebases.
On 21-07-08 11:51, Mark Brown wrote:
On Mon, Jul 21, 2008 at 10:45:29AM +0200, Rene Herman wrote:
rene@7ixe4:~/src/linux/7ixe4$ git fetch tiwai From git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6 ! [rejected] devel -> tiwai/devel (non fast forward) ! [rejected] dma-fix -> tiwai/dma-fix (non fast forward) ! [rejected] master -> tiwai/master (non fast forward)
I'll go ask on the GIT list what this might be about...
This means that the git remote is configured to only pull fast forwards from that repository. Add a + to the 'fetch =' for it in .git/config and it'll track non-linear updates like you get when the remote rebases.
I see.... thanks. I'll wait a bit with doing that then since it appears that this rejection might be the behaviour I want from it -- am using the devel branch specifically due to it only being rebased at releases.
Okay, thanks again, I get it. Things work again without anything special after a "git remote rm tiwai, git remote add tiwai <url>" cycle. That's ofcourse pretty horrible though. Really convinced by now that public repos that downstream users might have pulled should just not be rebased. Okay, the once per cycle as with that devel branch is probably okay but in essence any tree you rebase becomes a throw away tree that noone other than your upstream has any use for longer than once.
I'll keep the question on the GIT list outstanding due to the difference with "git pull" which I do still find unexpected.
Rene.
On Mon, Jul 21, 2008 at 12:39:05PM +0200, Rene Herman wrote:
On 21-07-08 11:51, Mark Brown wrote:
On Mon, Jul 21, 2008 at 10:45:29AM +0200, Rene Herman wrote:
This means that the git remote is configured to only pull fast forwards from that repository. Add a + to the 'fetch =' for it in .git/config and it'll track non-linear updates like you get when the remote rebases.
I see.... thanks. I'll wait a bit with doing that then since it appears that this rejection might be the behaviour I want from it -- am using the devel branch specifically due to it only being rebased at releases.
It probably doesn't hurt to track the current state of the remote anyway - even if the remote is rebased that won't affect the history of your development branch and I can't off-hand think of a reason why it'd be best to have the old status hanging around for a repo like this.
ofcourse pretty horrible though. Really convinced by now that public repos that downstream users might have pulled should just not be rebased. Okay,
Yes, that's the strongly expressed recommendation for things people might want to base work on.
participants (3)
-
Jaroslav Kysela -
Mark Brown -
Rene Herman