git pull regression?

git pull regression?

[Carlos R. Mafra]

Hi,

I tested git v1.6.0.3 today and noticed that git pull is
behaving strangely, compared to what I was used to with
v1.5.6.3

It no longer tells me "Already up-to-date." when it used
to, now it simply says nothing.

And when I pulled Linus tree for the kernel this morning 
it downloaded the new commits but git didn't show the 
diffstat.

Furthermore, it kept saying to me 
"Your branch is behind 'origin/master' by 31 commits, 
and can be fast-forwarded"
and git log was not showing the new commits.

At first I thought I had made some mistake with my repo,
but I searched the mailing list and there was a similar
case which was "fixed" after using 'git status' followed
by another 'git pull'. But in my case it did not help.

So I went back to v1.5.6.3 and everything was back
to normal.

After that I made some tests, and v1.6.0 was good
and v1.6.0.1 was bad, but I didn't bisect it
further.

It looks like a regression to me. I can finish
the bisection if people in the list say that
I am not making a mistake somewhere :-)

Re: git pull regression?

[Junio C Hamano]

"Carlos R. Mafra" <crmafra2@gmail.com> writes:

> It looks like a regression to me. I can finish
> the bisection if people in the list say that
> I am not making a mistake somewhere :-)

Interesting, and _sounds_ like a regression, but I do not think anybody
can tell if it is without looking at what .git/config and exact command
sequence you are using for this "git pull" and where you are starting
from.

Re: git pull regression?

[Carlos R. Mafra]

On Tue  4.Nov'08 at 15:37:03 -0800, Junio C Hamano wrote:
> "Carlos R. Mafra" <crmafra2@gmail.com> writes:
> 
> > It looks like a regression to me. I can finish
> > the bisection if people in the list say that
> > I am not making a mistake somewhere :-)
> 
> Interesting, and _sounds_ like a regression, but I do not think anybody
> can tell if it is without looking at what .git/config and exact command
> sequence you are using for this "git pull" and where you are starting
> from.

Ok, I am sorry for the lack of details, which I provide now.

It happened again with the git repo itself (Linus hasn't 
updated his yet).

[mafra@localhost:git]$ cat .git/config
[core]
        repositoryformatversion = 0
        filemode = true
        bare = false
        logallrefupdates = true
[remote "origin"]
        url = git://git.kernel.org/pub/scm/git/git.git
        fetch = +refs/heads/*:refs/remotes/origin/*
[branch "master"]
        remote = origin
        merge = refs/heads/master
	
And the commands were these (ok, I made a mistake that
I pulled without being first in the master branch,
but the problem I am describing here also happened
before, when I was already in the master branch)

[mafra@localhost:git]$ git --version
git version 1.6.0.3
[mafra@localhost:git]$ git log
[mafra@localhost:git]$ git pull
remote: Counting objects: 245, done.
remote: Compressing objects: 100% (49/49), done.
remote: Total 131 (delta 100), reused 109 (delta 82)
Receiving objects: 100% (131/131), 33.91 KiB, done.
Resolving deltas: 100% (100/100), completed with 43 local objects.
From git://git.kernel.org/pub/scm/git/git
   aa17c7c..826ea6f  html       -> origin/html
   a5a323f..16088d8  maint      -> origin/maint
   1655dcb..1e98488  man        -> origin/man
   9f8f132..5508064  master     -> origin/master
   58edc95..494ccbd  next       -> origin/next
 + fe6aa64...a5f8d19 pu         -> origin/pu  (forced update)
You asked me to pull without telling me which branch you
want to merge with, and 'branch..merge' in
your configuration file does not tell me either.  Please
name which branch you want to merge on the command line and
try again (e.g. 'git pull <repository> <refspec>').
See git-pull(1) for details on the refspec.

If you often merge with the same branch, you may want to
configure the following variables in your configuration
file:

    branch..remote = <nickname>
    branch..merge = <remote-ref>
    remote.<nickname>.url = <url>
    remote.<nickname>.fetch = <refspec>

See git-config(1) for details.
[mafra@localhost:git]$ git branch
* (no branch)
  master
[mafra@localhost:git]$ git checkout master
Previous HEAD position was 031e6c8... GIT 1.6.0.3
Switched to branch "master"
Your branch is behind 'origin/master' by 5 commits, and can be fast-forwarded.
[mafra@localhost:git]$ git pull
[mafra@localhost:git]$ git status
# On branch master
# Your branch is behind 'origin/master' by 5 commits, and can be fast-forwarded.
#
# Untracked files:
#   (use "git add <file>..." to include in what will be committed)
#
#       git-help--browse
#       git-merge-stupid
#       test-absolute-path
nothing added to commit but untracked files present (use "git add" to track)
[mafra@localhost:git]$ git pull
[mafra@localhost:git]$

Notice how "git pull" doesn't say "Already up-to-date" anymore.

And now 'git log' shows me the last commit being 
9f8f132621faedd instead of the really last one 550806439402

These are the same issues that I described in the first email
when I updated Linus' tree, so they are not due to the mistake
of not being in master branch.

And now if I go back to v1.6.0 I get this

[mafra@localhost:git]$ git checkout master
Switched to branch "master"
Your branch is behind 'origin/master' by 5 commits, and can be fast-forwarded.
[mafra@localhost:git]$ git pull
Updating 9f8f132..5508064
Fast forward
 Documentation/Makefile             |    8 ++++-
 Documentation/git-format-patch.txt |    3 +-
 INSTALL                            |    5 ++-
 Makefile                           |    6 ++++
 contrib/hooks/post-receive-email   |   57 ++++++++++++++++++++++++++++++------
 5 files changed, 65 insertions(+), 14 deletions(-)
[mafra@localhost:git]$

and everything is ok again.

Anything else I can do to help or is bisection the last resort? :-)

Thanks,
Carlos

Re: git pull regression?

[Carlos R. Mafra]

On Wed  5.Nov'08 at  9:38:10 +0100, Carlos R. Mafra wrote:
> On Tue  4.Nov'08 at 15:37:03 -0800, Junio C Hamano wrote:
> > "Carlos R. Mafra" <crmafra2@gmail.com> writes:
> > 
> > > It looks like a regression to me. I can finish
> > > the bisection if people in the list say that
> > > I am not making a mistake somewhere :-)
> > 
> > Interesting, and _sounds_ like a regression, but I do not think anybody
> > can tell if it is without looking at what .git/config and exact command
> > sequence you are using for this "git pull" and where you are starting
> > from.

Hm...that what I feared most has just happened.

I tried to bisect my problem when I realized that it wasn't
goint to take 2 hours and a lot of reboots like with the kernel :-)

And I soon realized that something was wrong when I was
running 'make install'. I had two terminals open, one with
the regular user with which I was compiling git, and the
other with the root user where I run 'make install'.

After compiling with the regular user and running
'make install' with the root user from the other terminal
the whole compilation was happening again and strangely
it was installing things in the prefix /root instead of
/usr/local

I simply don't know why this happened nor how. But 
somehow my git installation was screwed.

I have just compiled the latest git and it is ok now.

I am really sorry for all the noise, but I was 
sincerely fooled by the symptoms (as you can see
in the previous email).