Re: Git conversion status

On Wed, Sep 22, 2010 at 4:04 PM, Alvaro Herrera
<alvherre@commandprompt.com> wrote:
> Excerpts from Robert Haas's message of mar sep 21 13:43:28 -0400 2010:
>
>> Oops.  I left out a step.  Fixed.
>
> Hmm, it still doesn't work.  I can pull into the normal clone and it
> works, but when I run a "git pull" in one of the local clones, it says
> it's up to date even when there are commits that I know should be pulled
> in.
>
> For example:
>
> $ cd postgresql/
> $ git fetch
> remote: Counting objects: 219, done.
> remote: Compressing objects: 100% (88/88), done.
> remote: Total 89 (delta 77), reused 0 (delta 0)
> Unpacking objects: 100% (89/89), done.
> From ssh://gitmaster.postgresql.org/postgresql
>   6b4453f..da907fd  REL7_4_STABLE -> origin/REL7_4_STABLE
>   92458c2..da49d16  REL8_0_STABLE -> origin/REL8_0_STABLE
>   3fb50a7..706a580  REL8_1_STABLE -> origin/REL8_1_STABLE
>   adbe80f..c206794  REL8_2_STABLE -> origin/REL8_2_STABLE
>   c39a381..60591cd  REL8_3_STABLE -> origin/REL8_3_STABLE
>   35b2f93..2792c82  REL8_4_STABLE -> origin/REL8_4_STABLE
>   bbf84ac..f23bc1e  REL9_0_STABLE -> origin/REL9_0_STABLE
>   726f9dd..6c137da  master     -> origin/master
>
> $ cd ../REL9_0_STABLE
> $ git pull
> Current branch REL9_0_STABLE is up to date.
>
> But I know this is wrong:
>
> $ git log -1
> commit a6923594114601b4aaaf0cfd82eb5088af837664
> Author: Magnus Hagander <magnus@hagander.net>
> Date:   Wed Sep 22 12:57:06 2010 +0200
>
>    Convert cvsignore to gitignore, and add .gitignore for build targets.
>
> $ cd ../postgresql
> $ git checkout REL9_0_STABLE
> Checking out files: 100% (2415/2415), done.
> Switched to branch 'REL9_0_STABLE'
> Your branch is behind 'origin/REL9_0_STABLE' by 2 commits, and can be fast-forwarded.
> $ git pull
>  [ lotsa output ]
>
> $ git log -1
> commit f23bc1e8a42cab50c204bbab837f95cbc2353311
> Author: Magnus Hagander <magnus@hagander.net>
> Date:   Wed Sep 22 21:49:07 2010 +0200
>
>    Add gitignore files for ecpg regression tests.
>
>    Backpatch to 8.2 as that's how far the structure looks the same.
>
>
>
> As far as I can see, I need to go to the master clone, run a checkout
> and pull on each branch, and *then* a pull on the local clone updates to
> the latest head on that branch.  It is not enough to pull when the
> master branch is checked out.

Ah, crap.  You're right.  Sucktastic.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise Postgres Company