Re: building subversion from source rpm - failures

--On Monday, February 07, 2005 6:14 PM -0600 Steve Cohen
<scohen@javactivity.org> wrote:

> my question is really more on the conceptual level. We are building a
> binary rpm from a source rpm. As far as rpmbuild knows, subversion has
> not even been deployed yet. So why would such tests be expected to work?

It looks like a server is set up to listen on a user port for purposes of
testing. This server need not be in the final location to work.

Do you have a firewall up on the build system? Perhaps blocking port 15835
on the loopback interface is preventing the tests from succeeding. My
firewall includes a rule that allows all connections on interface "lo".

Note that there's not really anything special here about the use of
rpmbuild. You'd run into the same issue with "make test" when building from
the tarball. The rpm build system just wraps all the unpacking, patching,
making, and building the binary container in a simple command. It captures
and documents all the steps the tarball user would have to perform "by
hand".

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Tue Feb 8 01:38:17 2005