[PATCH xorg-gtest 0/2] Import googletest into xorg-gtest

[Peter Hutterer]

On Tue, Oct 09, 2012 at 08:21:19AM -0700, Chase Douglas wrote:
> On Wed, Oct 3, 2012 at 11:44 PM, Peter Hutterer
> <peter.hutterer at who-t.net> wrote:
> >
> > We need to build googletest and xorg-gtest from source for users of
> > xorg-gtest because we cannot guarantee a stable ABI (see googletest
> > documentation/C++ ODR for reasons).
> >
> > This means we need googletest source, xorg-gtest source and actual test
> > sources around to compile. To further entertain developers, one shouldn't
> > make install googletest (it will complain about it), one must provide an
> > absolute path to googletest's sources for xorg-gtest, one must make install
> > xorg-gtest, and then, in the client, provide the path to the googletest
> > sources and set PKG_CONFIG_PATH etc. to the xorg-gtest installed bits.
> > This is madness, or at least inconvenient, whichever comes first.
> >
> > Extra fun is to be had if one tries to build against a googletest that's in
> > the package repo and may be out of date.
> >
> > These patches import googletest into xorg-gtest and thus allow xorg-gtest to
> > be built with normal dependencies only, and likewise the clients to use
> > xorg-gtest without having to worry about googletest. Clients must update to
> > the new Makefile-xorg-gtest.am though to avoid build errors once this is
> > merged.
> 
> I'm ok with the change. It would be nice if there were a mechanism for
> detecting when the gtest sources are out of date, but I don't know of
> any prebuilt solution for that off-hand. We could have a script that
> checks the latest version available online whenever make distcheck is
> run.

there's an RSS feed available for googletest releases.
http://code.google.com/p/googletest/feeds
not as good as an email, but good enough to see when a new version comes
out.

having said that, unless we find some limitations or issues with v1.6
there's probably not a lot of reason to immediately update to the latest and
greatest. These tests don't really need to worry about security fixes, so
it's really only the API that would be the limiting factor.

> Reviewed-by: Chase Douglas <chase.douglas at canonical.com>

thanks.

Cheers,
   Peter