Andriy Fedorov wrote:
> Ok, I seem to have the cvs problem resolved. As it turns out, the
> problem is that "cvs --help" returns 1, not 0 (at least on my
> platform).

Oooops.

> So when this happens, OpenGrok ScmChecker's test fails, and Grok
> thinks cvs is not found. Looks like a bug in either cvs or Executor. I
> am not sure what are the rules about return values.
> 
> So I did the following change in ScmChecker:
> 
>         Executor exec = new Executor(argv);
>         int ret = exec.exec(false);
>         available = false;
>         if(ret == 0 || ret == 1)
>           available = true;
> 
> It finds cvs, I re-run opengrok, but I still don't have history for
> svn, and neither history, nor annotations for svn.

Good catch!

Please file an issue here:
http://www.opensolaris.org/os/project/opengrok/Issues/

We'll have to re-think how we detect cvs.

> By the way, do I
> have to do something extra once I re-run the indexer? Do I have to
> re-deploy?

Don't know. Guess "No".

> I have a large project that includes a number of sub-projects, all of
> them located in the remote repositories managed by either CVS or SVN.

Ah, Trond tipses me that adding "-r on" on the command line might help?

Ex: "-r on" will allow retrival for remote SCM systems
"-r off" will ignore SCM for remote systems

-J


Reply via email to