On Saturday 27 February 2010 23:55:38 Zoltan Padrah wrote:
> 2010/2/23 Julian Bäume <julian@...>
> > On Saturday 20 February 2010 20:49:11 P Zoltan wrote:
> > > The next thing i wanted to do is to merge the testing code in the
> > > kde4
> > >
> > > port tree. I have to learn more abot git to know how to do it; if
> > > anyone wants to do this, I'd happily copy the merged tree, and work on
> > > that.
> > done that. I also removed the Eigen sources and added everything to the
> > CMakeLists.txt that is necessary to find the eigen headers. This was a
> > little
> > more git work, than just merging, since I also did a rebase, so the Eigen
> > sources even don't show up in the history.
> Those junk-related commits could be also deleted (by mistake i've
> comitted the build binaries in the repository, then deleted it).
I thought about cleaning that up, too, but decided against it. This won't blow
up the history to much.
> Here is a problem: if cmake can't find eigen2, it still tries to build
> the tests, and fails. So a minor issue will turn to a build faliure. The
> tests should be only built when eigen2 is found.
Ah, I'm sorry about that. I just did a quick cleanup and integrated
everything. I thought you'd do the rest ;P Anyway, I put a check around the
mathtest, so it will only be built, if eigen2 is found.
> > I also moved around your files a bit. You can find your test in
> > tests/math. For now, there seems to be no output (at least for me), I'm
> > not sure, what has
> > to be done to get it back.
> Here is a useful blog entry on this:
> So in the build directory, you need to run "ctest" to run the tests, and
> "ctest -V" or "ctest -VV" to see all the output. This verbose mode should
> be somehow integrated in the cmake files.
Great, that did the trick ;) I didn't know about this ctest binary, I just ran
make test all the time. I think the ctest -V command is mostly useful to get
some output if a test fails. If everything passes, it's okay to have no output