Thank you for the reply, guys. I wonder if I could just write my own
stand-alone service (listening on port 8080) that could sit on the gitlab
box and just serve up the output of 'git show..." command.
On Tuesday, November 20, 2012 5:39:36 PM UTC-5, Matthew Woehlke wrote:
> On 2012-11-20 15:49, DXM wrote:
> > I'm trying to setup ReviewBoard to work with Gitlab repository. Since
> > Gitlab has a web front end API, From everything I've read so far, seems
> > most straightforward integration would be to specify a Raw File URL mask
> > the repository setup.
> > However, I've run into a bit of a problem. It looks like for <revision>
> > placeholder, RB server will pass in file blob's SHA1, but Gitlab API
> > the client to pass in SHA1 of the commit.
> Sorry to say, but if Gitlab cannot give you a blob by it's identifying
> SHA, you are out of luck. By the time a diff makes it to RB, the SHA of
> the commit against which the diff was generated is no longer known. (And
> I am not aware that it is possible to determine what commit "owns" a
> particular blob ID.)
> So... basically, see David's reply for your options. Getting the Gitlab
> server patched is probably best; being able to retrieve blobs by ID is a
> feature the server really ought to provide.
> (I had the same problem with an older gitorious server, and spent some
> time trying to determine if there was some other solution short of a
> complete mirror of the repository - e.g. a local web service that could
> fetch just the necessary blob - but wasn't able to come up with
> anything. In the end, I had to patch the gitorious server. Fortunately
> upstream gitorious has already been patched, so it was just a matter of
> grabbing the necessary patch from upstream and applying it.)
> p.s. If you end up patching Gitlab yourself, what you need is basically
> a URL that gives you the output of 'git show <blob-id>'. The path isn't
> actually needed for git.
Want to help the Review Board project? Donate today at
Happy user? Let us know at http://www.reviewboard.org/users/
To unsubscribe from this group, send email to
For more options, visit this group at