[ANNOUNCE] GIT 1.5.3-rc7

[ANNOUNCE] GIT 1.5.3-rc7

[Junio C Hamano]

So I've tagged it, and this time have tarball and testing RPM at
the usual places

The latest feature release GIT 1.5.3-rc7 is available at the
usual places:

  http://www.kernel.org/pub/software/scm/git/

  git-1.5.3.rc7.tar.{gz,bz2}			(tarball)
  git-htmldocs-1.5.3.rc7.tar.{gz,bz2}		(preformatted docs)
  git-manpages-1.5.3.rc7.tar.{gz,bz2}		(preformatted docs)
  testing/git-*-1.5.3.                          rc7-1.$arch.rpm	(RPM)

[ANNOUNCE] GIT 1.5.3-rc7

[Junio C Hamano]

(oops, sent one out before finishing --- I am sitting in front
of a wrong machine and a wrong keyboard).

So I've tagged it, and this time have tarball and testing RPM at
the usual places.

  http://www.kernel.org/pub/software/scm/git/

  git-1.5.3.rc7.tar.{gz,bz2}			(tarball)
  git-htmldocs-1.5.3.rc7.tar.{gz,bz2}		(preformatted docs)
  git-manpages-1.5.3.rc7.tar.{gz,bz2}		(preformatted docs)
  testing/*-1.5.3.rc7-1.$arch.rpm		(RPM)

This time, hopefully it should be very close to final.  I still
have to help Jari whipping his last minute documentation patches
into shape, but other than that, I do not think there is
anything else that I want to delay the 1.5.3 final because of.

Please give it a gentle beating ;-).

Re: [ANNOUNCE] GIT 1.5.3-rc7

[Johannes Sixt]

The MinGW port is now at 1.5.3-rc7, too.

-- Hannes

Re: [ANNOUNCE] GIT 1.5.3-rc7

[Johannes Schindelin]

Hi,

On Wed, 29 Aug 2007, Johannes Sixt wrote:

> The MinGW port is now at 1.5.3-rc7, too.

Should we switch from  4msysgit.git's devel branch to your master branch?

Ciao,
Dscho

Re: [ANNOUNCE] GIT 1.5.3-rc7

[Johannes Sixt]

Johannes Schindelin schrieb:
> On Wed, 29 Aug 2007, Johannes Sixt wrote:
> 
>> The MinGW port is now at 1.5.3-rc7, too.
> 
> Should we switch from  4msysgit.git's devel branch to your master branch?

No. There are still too many differences.

But I think it is time to rebase 4msysgit on top of 'devel' - without 
the old devel commits, i.e.

	git rebase --onto mingw/devel 382210f715

There is a lot of stuff in 4msysgit.git that is no longer needed or 
superceded by mingw.git.

Side note: If you can live with a test suite does not complete without 
failures you can rebase on top of master and avoid the devel branch. 
(Think of broken repo when I rebase devel again.)

-- Hannes

Re: [ANNOUNCE] GIT 1.5.3-rc7

[Johannes Schindelin]

Hi,

On Wed, 29 Aug 2007, Johannes Sixt wrote:

> Johannes Schindelin schrieb:
> > On Wed, 29 Aug 2007, Johannes Sixt wrote:
> > 
> > > The MinGW port is now at 1.5.3-rc7, too.
> > 
> > Should we switch from  4msysgit.git's devel branch to your master branch?
> 
> No. There are still too many differences.

Right.  And I see that there are many more tests failing with mingw.git's 
master, too.  Should have tested that first.

> But I think it is time to rebase 4msysgit on top of 'devel' - without 
> the old devel commits, i.e.
> 
> 	git rebase --onto mingw/devel 382210f715
> 
> There is a lot of stuff in 4msysgit.git that is no longer needed or 
> superceded by mingw.git.
> 
> Side note: If you can live with a test suite does not complete without 
> failures you can rebase on top of master and avoid the devel branch. 
> (Think of broken repo when I rebase devel again.)

Hmm.  I really want to avoid that breakage...

To refresh my memory, could you please describe how mingw.git's master and 
devel branches are handled?

Ciao,
Dscho

Re: [ANNOUNCE] GIT 1.5.3-rc7

[Johannes Sixt]

Johannes Schindelin schrieb:
> To refresh my memory, could you please describe how mingw.git's master and 
> devel branches are handled?

'master' merges git.git's tagged releases, i.e. the 'maint' branch, but 
also the -rc series of a new feature release (like in the past weeks).

'devel' contains additional commits that make a "flawless" test suite, 
i.e. mostly disables tests that trigger bugs. This branch is always 
rebased onto of 'master'.

Originally 'devel' contained code changes without which MinGW git would 
have been unusable, but currently it really only disables tests that 
trigger exotic bugs[*] that I cannot fix right away and so is basically 
just for my own convenience. I could very well get rid of the branch - 
if only to raise the incentive for others to help fix the bugs. ;)

[*] bugs that I do not encounter in may daily workflow.

-- Hannes