[ANNOUNCE] Stacked GIT 0.14.2

Stacked GIT 0.14.2 release is available from http://www.procode.org/stgit/.

StGIT is a Python application providing similar functionality to Quilt (i.e. pushing/popping patches to/from a stack) on top of GIT. These operations are performed using GIT commands and the patches are stored as GIT commit objects, allowing easy merging of the StGIT patches into other repositories using standard GIT functionality.

This is a maintenance release including bug-fixes and a few user interface enhancements. Acknowledgements:

Catalin Marinas (13):
      Don't set the default authdate if none specified
      Allow pick to import multiple patches
      Check for unnecessary push/pop in 'float'
      Allow the synchronisation of the unapplied patches
      Modify 'series' to use '#' instead of '|'
      Set umask to 0022 during the setup.py execution
      Refuse to send empty patches
      Add a boundary to parse_patches in pick.py
      Fix sync to push the popped patches back after sync'ing
      Remove the reordering side-effect of the latter sync changes
      Allow picking of one commit id
      Remove a newline from the e-mail template
      Release 0.14.2

Onno Kortmann (2):
      Simple rename of top-most patch
      Test the 'stg rename' command

Alex Chiang (1):
      Better "stg rebase" help text

Angus Salkeld (1):
      stg mail crashes when there is no patch description

Karl Hasselström (1):
      Make documentation less confusing

Peter Oberndorfer (1):
      replace "git repo-config" usage by "git config"

Toby Allsopp (1):
      Fix "stg branch --delete" on a nonexistent branch

On Monday 2008-03-24 20:59, Catalin Marinas wrote:

I always wondered what the difference between stgit and guilt is. Does anyone have a comparison up?

On 24-03-08 21:15, Jan Engelhardt wrote:

And I remember some mumblings about git growing quilt-like functionality itself. Anything on that?

Rene.

Rene Herman <rene.herman@keyaccess.nl> writes:

Not my mumbling, but I am quite open to slurp in guilt as a subdirectory in git.git at some point in the future just like we bundle git-gui and gitk if asked by the maintainer.

The same applies to StGIT for that matter, although I somehow feel that is much less likely to happen, because it lived long enough as a standalone project with enough following to achieve sustainable momentum by itself.

On 24-03-08 21:43, Junio C Hamano wrote:

Believe it was Linus. Seem to remember him saying something about possibly adding a native queues-like interface not too long ago but it's proving impossible to google for.

Rene.

On 24/03/2008, Rene Herman <rene.herman@keyaccess.nl> wrote:

It was Linus indeed, on the linux-arch mailing list:

http://www.mail-archive.com/linux-arch@vger.kernel.org/msg05012.html

On 24-03-08 23:50, Catalin Marinas wrote:

Probably read it on linux-kernel but yes, that was the one, thanks much. Stored it as an interesting something -- used to use quilt and although I didn't much care for its linear nature when managing an entire tree, it probably works out well for private topic branches. Not overly sure of added practical value over git rebase, but I did like the simple nature of shuffling things around by just editing a series file back when I was using quilt. Keeps one on top of things, so to speak (<-- amusing play on words).

Moreover not too keen on using many tools do do one thing, so getting this inside git itself sounded like a possibly nice notion.

Rene.

On 24/03/2008, Rene Herman <rene.herman@keyaccess.nl> wrote:

Git rebasing is useful but using a patch management tool is much easier, especially when you need to post a series of patches multiple times while implementing others' comments. And this is not only for private branches but public ones that are meant for upstream merging. The problem with these tools is that they rewrite the history of the repository but they are OK if you aren't in a position where people need to base their work on your tree.

I would much prefer to simply commit the additional comments at the top of a branch without changing the history but this makes patch posting and reviewing more difficult. Because of this requirement to keep clean patches (which I think is fair), people using maintainers' branches can only use rebasing (or some patch management tool).

I don't like using many tools either, hence I use mostly StGIT :-).

On Mon, Mar 24, 2008 at 01:43:00PM -0700, Junio C Hamano wrote:

 
Junio,

That'd be great! Is there anything special you want me to do? Well, I want to finish up a merge I've been working on first.

Josef 'Jeff' Sipek.

Hi,

[cutting down the Cc: list to the most-likely-interested parties]

On Mon, 24 Mar 2008, Josef Sipek wrote:

Would it not be more appropriate to slurp it in as a subproject? But then the same grace period as for git-gui and gitk would apply, since we still expect some users not to have a submodule capable git...

Ciao, Dscho

Johannes Schindelin <Johannes.Schindelin@gmx.de> writes:

All true. On the other hand we could start with a subtree merge and later switch.

In any case, the answer to Josef's question is to make sure I won't hear objections on the list that say "No, the stuff is still too premature to live in git.git itself and needs this and that fixups to be usable for the general public." ;-) IOW, keep up the good work. Gathering a bit more followers, preferably outside this mailing list (such as wine, x.org and/or the kernel community), would not hurt.

On Wed, Mar 26, 2008 at 02:24:22PM -0700, Junio C Hamano wrote: ...

Great...of course _I_ know all the things in guilt that _I_ would like to see changed (the maintainer syndrome?). :)

The git mailing list is interesting because it tends to have people from all the other projects as well. From the looks of things, Guilt seems to have a pretty strong following the Linux filesystem area.

If things go well, I'll be able to go to OLS and have a BoF session about Guilt (already got an Ok from the conference folks, it's just a matter of figuring out travel/hotel).

Overall, sounds like a plan.

Josef 'Jeff' Sipek.

On 24/03/2008, Junio C Hamano <gitster@pobox.com> wrote:

Indeed, StGIT has grown quite a lot (it now has ~45 commands) and it probably feels better as a tool on its own. However, I think it would be good for git to get such functionality (and it could even share some ideas StGIT and guilt). If this functionality will get fully integrated in git, StGIT could use it directly rather than the low-level git commands.

An example of additional functionality that might stay outside of git (at least for a while): since basic patch management support is pretty much complete in StGIT, a feature we are thinking about adding is better support for patch sharing between developers (or repositories/branches). This should allow easier cloning and merging with branches that have a volatile history at the top.

On 24/03/2008, Jan Engelhardt <jengelh@computergmbh.de> wrote:

There was a thread last year:

http://kerneltrap.org/mailarchive/git/2007/6/14/249310

I don't follow the guilt development to be able to comment. They are pretty similar regarding patch management but it's probably best to try both and see which tool you like. StGIT might have a few more features as it was around for longer (e-mail templates, patch synchronisation between branches etc.) but guilt seems actively developed as well.

I might be wrong here but I'm not sure whether guilt uses three-way merging when pushing a patch or just a two-way diff apply. The three-way merging has several advantages in dealing with conflicts.

On Mon, Mar 24, 2008 at 11:24:20PM +0000, Catalin Marinas wrote:

 
Right. Try both and use whatever you're happier with.

It's still using two-way diff apply, but I'd like to implement a three-way merge as a fallback.

Josef 'Jeff' Sipek.