[ANNOUNCE] GIT 0.99.7b

[ANNOUNCE] GIT 0.99.7b

[Junio C Hamano]

GIT 0.99.7b

Contains the following post-0.99.7a fixes:

 - Commit walker performance fix, mostly during walking commits
   in a downloaded packfile, thanks to Sergey Vlasov.

 - Squelch unnecessarily alarming error message from fetch and
   clone over rsync transport, when remote repository does not
   borrow anything from other repositories.

   I've seen enough people got annoyed/worried/alarmed after
   seeing this one on both git and linux-kernel list.

 - Documentation was not rebuilt before installation, noticed by
   Randal L Schwartz.

 - Fetching of objects over http transport got a bit safer.


Tarballs, RPMs and Debs are available at

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

Or, if you use git already:

	{http,rsync}://kernel.org/pub/scm/git/git.git/

Re: [ANNOUNCE] GIT 0.99.7b

[walt]

Junio C Hamano wrote:
> GIT 0.99.7b
[...]
> Or, if you use git already:
> 
> 	{http,rsync}://kernel.org/pub/scm/git/git.git/

I did it that way and now I see this:

$git --version
git version 0.99.7

This is a bit confusing.  Did I upgrade or downgrade?
Maybe the version string should contain the same kind
of hexadecimal string that Linus is appending to his
kernel version?

Re: [ANNOUNCE] GIT 0.99.7b

[H. Peter Anvin]

walt wrote:
> Junio C Hamano wrote:
> 
>> GIT 0.99.7b
> 
> [...]
> 
>> Or, if you use git already:
>>
>>     {http,rsync}://kernel.org/pub/scm/git/git.git/
> 
> 
> I did it that way and now I see this:
> 
> $git --version
> git version 0.99.7
> 
> This is a bit confusing.  Did I upgrade or downgrade?
> Maybe the version string should contain the same kind
> of hexadecimal string that Linus is appending to his
> kernel version?
> 

Or just call it 0.99.7.1 and .2 or 0.99.8 and .9.

	-hpa

Re: [ANNOUNCE] GIT 0.99.7b

[Junio C Hamano]

walt <wa1ter@myrealbox.com> writes:

> Junio C Hamano wrote:
>> GIT 0.99.7b
> [...]
>> Or, if you use git already:
>> 	{http,rsync}://kernel.org/pub/scm/git/git.git/
>
> I did it that way and now I see this:

If you fetched v0.99.7b tag and built it, hopefully it should
say 0.99.7b otherwise you found a big bug.

If you pulled from "master" branch, then --version would still
say 0.99.7; I agree it is confusing.  On the other hand, I do
not think we would want to increment the version string with
every little changes, so...

Re: [ANNOUNCE] GIT 0.99.7b

[Joel Becker]

On Fri, Sep 23, 2005 at 01:46:14PM -0700, Junio C Hamano wrote:
> If you pulled from "master" branch, then --version would still
> say 0.99.7; I agree it is confusing.  On the other hand, I do
> not think we would want to increment the version string with
> every little changes, so...

	Every little change, no.  But ANNOUNCEd releases should be
considered 'major' from this point of view.  I don't think a little
extra version number incrementing is too big a deal.

Joel

-- 

Life's Little Instruction Book #451

	"Don't be afraid to say, 'I'm sorry.'"

Joel Becker
Principal Software Developer
Oracle
E-mail: joel.becker@oracle.com
Phone: (650) 506-8127

Re: [ANNOUNCE] GIT 0.99.7b

[Petr Baudis]

Dear diary, on Fri, Sep 23, 2005 at 10:46:14PM CEST, I got a letter
where Junio C Hamano <junkio@cox.net> told me that...
> If you pulled from "master" branch, then --version would still
> say 0.99.7; I agree it is confusing.  On the other hand, I do
> not think we would want to increment the version string with
> every little changes, so...

In ELinks, we pseudo-bump the version number from v1.2.3 to v1.2.3.CVS
(now v1.2.3.GIT) right after the release.

In Cogito, we append the commit ID current at the time of build to
cg-version output (if it was built from GIT-tracked working directory).

-- 
				Petr "Pasky" Baudis
Stuff: http://pasky.or.cz/
VI has two modes: the one in which it beeps and the one in which
it doesn't.

Re: [ANNOUNCE] GIT 0.99.7b

[Petr Baudis]

Dear diary, on Fri, Sep 23, 2005 at 10:52:42PM CEST, I got a letter
where Joel Becker <Joel.Becker@oracle.com> told me that...
> On Fri, Sep 23, 2005 at 01:46:14PM -0700, Junio C Hamano wrote:
> > If you pulled from "master" branch, then --version would still
> > say 0.99.7; I agree it is confusing.  On the other hand, I do
> > not think we would want to increment the version string with
> > every little changes, so...
> 
> 	Every little change, no.  But ANNOUNCEd releases should be
> considered 'major' from this point of view.  I don't think a little
> extra version number incrementing is too big a deal.

The point is that the bugfix releases aren't tagged on the master branch
but on the maint branch.

-- 
				Petr "Pasky" Baudis
Stuff: http://pasky.or.cz/
VI has two modes: the one in which it beeps and the one in which
it doesn't.

Re: [ANNOUNCE] GIT 0.99.7b

[Junio C Hamano]

Joel Becker <Joel.Becker@oracle.com> writes:

> On Fri, Sep 23, 2005 at 01:46:14PM -0700, Junio C Hamano wrote:
>> If you pulled from "master" branch, then --version would still
>> say 0.99.7; I agree it is confusing.  On the other hand, I do
>> not think we would want to increment the version string with
>> every little changes, so...
>
> 	Every little change, no.  But ANNOUNCEd releases should be
> considered 'major' from this point of view.  I don't think a little
> extra version number incrementing is too big a deal.

Yes, 'maint' has the GIT_VERSION bumped in the Makefile
(0.99.7a, 0.99.7b) and that is propagated to 'git --version'.

What I was having trouble was what to do with the stuff we put
in the 'master' between 0.99.7 and 0.99.8.  'maint' was *forked*
from 0.99.7, not just tagging some midpoint in 'master' line for
obvious reasons, so the 'git --version' on the 'master' side
does not have much to do with what happens on the 'maint' side.

We could say, immediately after 0.99.7 is released, bump the
GIT_VERSION in the 'master' branch to 0.99.8-pre or something
like that.

We need to pick something that sorts after '0.99.7a', '0.99.7b',
etc. and before '0.99.8'.  Somebody needs to come up with what
is appropriate for version number ordering scheme employed by
binary packaged distributions.  I know 0.99.8-pre does not sort
before 0.99.8 on Debian.  Probably '0.99.7zzz'?

Re: [ANNOUNCE] GIT 0.99.7b

[Junio C Hamano]

Petr Baudis <pasky@suse.cz> writes:

> Dear diary, on Fri, Sep 23, 2005 at 10:46:14PM CEST, I got a letter
> where Junio C Hamano <junkio@cox.net> told me that...
>> If you pulled from "master" branch, then --version would still
>> say 0.99.7; I agree it is confusing.  On the other hand, I do
>> not think we would want to increment the version string with
>> every little changes, so...
>
> In ELinks, we pseudo-bump the version number from v1.2.3 to v1.2.3.CVS
> (now v1.2.3.GIT) right after the release.

That's a nice way of doing it.  Maybe I should follow suit.