On Sat, Oct 10, 2009 at 03:37:20PM +0200, Kovarththanan Rajaratnam wrote:
> > On Sat, Oct 10, 2009 at 11:19:46AM +0200, Kovarththanan Rajaratnam wrote:
> >>> What's the idea of crating the subdir CMakefiles and the static libs?
> >>> I was happy to get rid of them as they add an unnecessary linking step.
> >>> Also, are you sure that linking a static lib into a dynamic one is OK?
> >>> The same question holds for the other static libs added by other commits.
> >>>
> >> The idea was that instead of lumping it all together in one file, we
> >> could split it up to make it easier to understand/maintain in the
> >> future. This is also the approach used by our current build setup and I
> >> hadn't heard anyone complaining about that so I just went ahead and
> >> replicated the same structure.
> >
> > IMO, it was done because it was convenient and "easier" with the old build
> > system. I don't think there's any better reason for it. The -Werror stuff
> > also caused a split into a "clean" and a "non-clean" lib, which can be
> > handled much better with the CMake mechanism (and which is near the top of
> > my todo list).
> >
> > I'd really like to undo that change, because I don't see any advantage in
> > doing it the "old" way.
>
> Since I haven't heard any convicing arguments (from both sides ;) I
> think it would be great if other people could give their opnion on this
> subject.
OK, the (to me :-) convincing arguments were: More than double the amount
of disk space needed during build (less important) and the dissectors were
linked into the binaries statically instead of dynamically. Thinking a bit
more about it, I think that the libs built by the autofoo system are some
sort of "convenience" libs, they are not static libs.
Ciao
Joerg
--
Joerg Mayer <jmayer@xxxxxxxxx>
We are stuck with technology when what we really want is just stuff that
works. Some say that should read Microsoft instead of technology.