This turned out to be due to a permissions error.  Some of the files
in the git repository were not readable by the webserver user.  Once I
fixed that, the upload to ReviewBoard worked fine.  Thank you for your
help, Christian!
   --Paul

On Sep 23, 6:38 pm, Paul Lynch <plynch...@gmail.com> wrote:
> It does seem that the error status is 0:> git --git-dir=/proj/git/def.git 
> cat-file -t febdf30
>
> fatal: Not a git repository (or any of the parent directories): .git
> blob> echo $?
>
> 0
>
> I just notice an odd thing, though.  I logged into two other systems
> that can see that repository (we have it mounted over NFS), one of
> which is the machine on which the repository is actually located and
> also is where Review Board runs.  On those two systems, I don't get
> the "fatal" error message.  I thought maybe it was because on those
> system I don't have a .gitconfig directory in my home directory but I
> tried my moving that out of the way on my local workstation and I
> still get the error, so I'm not sure what causes the difference.  In
> any event, it might be that Review Board isn't getting that fatal
> error message, so that might be a red-herring.
>
> Hmmm.  The two machines on which I don't get the error message are
> running git 1.7.0.4, and on my system it is git 1.7.4.1.  (Review
> Board is therefore getting 1.7.0.4.)
>
> On Sep 23, 6:24 pm, Christian Hammond <chip...@chipx86.com> wrote:
>
>
>
>
>
>
>
> > I wouldn't expect it to need to be in that directory. We do look for a
> > "fatal:" but only for "Not a valid object name." However, it's very likely
> > that the error code is when you execute that isn't 0, so we're failing.
>
> > So for that repo, if you're in /proj/git/def.git, it works?
>
> > What version of git is installed there?
>
> > Christian
>
> > --
> > Christian Hammond - chip...@chipx86.com
> > Review Board -http://www.reviewboard.org
> > VMware, Inc. -http://www.vmware.com
>
> > On Fri, Sep 23, 2011 at 2:59 PM, Paul Lynch <plynch...@gmail.com> wrote:
> > > It works, but if my current working directory is not in a git
> > > repository, I also get a "fatal" error, after which it outputs the
> > > type.  For example:
>
> > > > git --git-dir=/proj/git/def.git cat-file -t febdf30
> > > fatal: Not a git repository (or any of the parent directories): .git
> > > blob
>
> > > (But not that it does output "blob" after all.)  If I run this exact
> > > same command within my local repository, I just get "blob".  Could
> > > that "fatal" error message be messing things up?  Does Review Board's
> > > current working directory need to be inside the bare repository?
>
> > > On Sep 23, 5:50 pm, Christian Hammond <chip...@chipx86.com> wrote:
> > > > So off-hand, I don't know why you're hitting this, but let's just run
> > > > through a couple more things.
>
> > > > During upload time, when we go to verify a file's existence, we run:
>
> > > >     git --git-dir=<repo path> cat-file -t <revision>
>
> > > > If you run that on your bare repo, does that still work? What do you 
> > > > see?
>
> > > > If not, what happens when you clone the repo and try it there?
>
> > > > Christian
>
> > > > --
> > > > Christian Hammond - chip...@chipx86.com
> > > > Review Board -http://www.reviewboard.org
> > > > VMware, Inc. -http://www.vmware.com
>
> > > > On Fri, Sep 23, 2011 at 2:31 PM, Paul Lynch <plynch...@gmail.com> wrote:
> > > > > git diff FETCH_HEAD..HEAD
>
> > > > > (before the push).  After that a push was done, and then I tried to
> > > > > upload the file.  I have also done a pull in another clone of the
> > > > > repository, and can see the files there too.
>
> > > > > Uploading was done through the Review Board web site (now at 1.6.1)
> > > > > using the "New Review Request" link.  (We updated to 1.6.1  from an
> > > > > old beta version in hopes that it might fix this problem.)
>
> > > > > On Sep 23, 5:24 pm, Christian Hammond <chip...@chipx86.com> wrote:
> > > > > > Ah okay.
>
> > > > > > How did you generate/upload the diff?
>
> > > > > > Christian
>
> > > > > > --
> > > > > > Christian Hammond - chip...@chipx86.com
> > > > > > Review Board -http://www.reviewboard.org
> > > > > > VMware, Inc. -http://www.vmware.com
>
> > > > > > On Fri, Sep 23, 2011 at 2:22 PM, Paul Lynch <plynch...@gmail.com>
> > > wrote:
> > > > > > > The Review Board repository "path" setting is the (bare) git
> > > > > > > repository that is our upstream repository to which we push our
> > > > > > > changes.  I can see the sha1s that Review Board is complaining
> > > about
> > > > > > > in that repository, so they do exist there.
>
> > > > > > > On Sep 23, 5:19 pm, Christian Hammond <chip...@chipx86.com> wrote:
> > > > > > > > How is Review Board set up with that repository? Is it using a
> > > local
> > > > > > > clone?
> > > > > > > > If so, you have to have something set up so that the clone is
> > > always
> > > > > > > > up-to-date, or you'll hit errors like this.
>
> > > > > > > > I generally recommend setting up something like cgit on the
> > > > > repository
> > > > > > > and
> > > > > > > > then using that as the Raw File URL access. See our docs for how
> > > to
> > > > > set
> > > > > > > that
> > > > > > > > up.
>
> > > > > > > > Christian
>
> > > > > > > > --
> > > > > > > > Christian Hammond - chip...@chipx86.com
> > > > > > > > Review Board -http://www.reviewboard.org
> > > > > > > > VMware, Inc. -http://www.vmware.com
>
> > > > > > > > On Fri, Sep 23, 2011 at 1:56 PM, Paul Lynch <plynch...@gmail.com
>
> > > > > wrote:
> > > > > > > > > I recently switched from cvs to git.  Our procedure with
> > > > > ReviewBoard
> > > > > > > > > with git is now similar to what we were doing before:  commit
> > > > > changes,
> > > > > > > > > do a git pull, then create a diff of the unpushed changes with
> > > "git
> > > > > > > > > diff FETCH_HEAD..HEAD", then push the changes.  Very
> > > frequently,
> > > > > when
> > > > > > > > > I try to upload such a diff file to ReviewBoard, it complains
> > > about
> > > > > > > > > one or more (but not all) of the blob SHA1s in the file, with
> > > an
> > > > > error
> > > > > > > > > like:
>
> > > > > > > > >   error: unable to find
> > > 50e95d4c42cc5972e9a849502ec37ca6624379e5
> > > > > > > > > fatal: git cat-file 50e95d4: bad file
>
> > > > > > > > > If I remove the offending diffs, I can upload the remainder of
> > > the
> > > > > > > > > diff file (though the result obviously lacks the changes I
> > > > > removed).
> > > > > > > > > The sha1s which RB is complaining about are present in the
> > > > > repository,
> > > > > > > > > and I can see them.  In other words, when I run "git cat-file
> > > blob
> > > > > > > > > 50e95d4" I don't get an error, but see the file.
>
> > > > > > > > > Does anyone have an idea why ReviewBoard is having trouble 
> > > > > > > > > with
> > > > > those
> > > > > > > > > sha1s?
>
> > > > > > > > > Thanks,
> > > > > > > > >  --Paul
>
> > > > > > > > > --
> > > > > > > > > Want to help the Review Board project? Donate today at
> > > > > > > > >http://www.reviewboard.org/donate/
> > > > > > > > > Happy user? Let us know athttp://www.reviewboard.org/users/
> > > > > > > > > -~----------~----~----~----~------~----~------~--~---
> > > > > > > > > To unsubscribe from this group, send email to
> > > > > > > > > reviewboard+unsubscr...@googlegroups.com
> > > > > > > > > For more options, visit this group at
> > > > > > > > >http://groups.google.com/group/reviewboard?hl=en
>
> > > > > > > --
> > > > > > > Want to help the Review Board project? Donate today at
> > > > > > >http://www.reviewboard.org/donate/
> > > > > > > Happy user? Let us know athttp://www.reviewboard.org/users/
> > > > > > > -~----------~----~----~----~------~----~------~--~---
> > > > > > > To unsubscribe from this group, send email to
> > > > > > > reviewboard+unsubscr...@googlegroups.com
> > > > > > > For more options, visit this group at
> > > > > > >http://groups.google.com/group/reviewboard?hl=en
>
> > > > > --
> > > > > Want to help the Review Board project? Donate today at
> > > > >http://www.reviewboard.org/donate/
> > > > > Happy user? Let us know athttp://www.reviewboard.org/users/
> > > > > -~----------~----~----~----~------~----~------~--~---
> > > > > To unsubscribe from this group, send email to
> > > > > reviewboard+unsubscr...@googlegroups.com
> > > > > For more options, visit this group at
> > > > >http://groups.google.com/group/reviewboard?hl=en
>
> > > --
> > > Want to help the Review Board project? Donate today at
> > >http://www.reviewboard.org/donate/
> > > Happy user? Let us know athttp://www.reviewboard.org/users/
> > > -~----------~----~----~----~------~----~------~--~---
> > > To unsubscribe from this group, send email to
> > > reviewboard+unsubscr...@googlegroups.com
> > > For more options, visit this group at
> > >http://groups.google.com/group/reviewboard?hl=en

-- 
Want to help the Review Board project? Donate today at 
http://www.reviewboard.org/donate/
Happy user? Let us know at http://www.reviewboard.org/users/
-~----------~----~----~----~------~----~------~--~---
To unsubscribe from this group, send email to 
reviewboard+unsubscr...@googlegroups.com
For more options, visit this group at 
http://groups.google.com/group/reviewboard?hl=en

Reply via email to