new buildsystem merged into master
Dirk Hohndel
dirk at hohndel.org
Wed Oct 9 16:09:59 UTC 2013
On Wed, 2013-10-09 at 16:07 -0700, Dirk Hohndel wrote:
> On Wed, 2013-10-09 at 16:02 -0700, Linus Torvalds wrote:
> > On Wed, Oct 9, 2013 at 3:57 PM, Dirk Hohndel <dirk at hohndel.org> wrote:
> > >
> > > If it's correctly installed it links against the shared library.
> >
> > .. and then it doesn't actually RUN, because the shared library is in
> > /usr/local/lib, and the libdivecomputer install doesn't actually run
> > ldconfig on it.
> >
> > More importantly, the shared library is pure and utter *garbage*
> > technically, since it's basically unversioned and there are only
> > disadvantages from sharing.
> >
> > So please. Don't *ever* use the shared one. It's stupid. It's a
> > mistake. Didn't we go through this long ago? The Debian people can do
> > their own f*cking thing to satisfy their black little hearts that care
> > more about their idiotic policies than about actual technical issues.
> >
> > > In which scenario does this fail for you?
> >
> > The standard one "make ; make install", ie normal build rules.
> >
> > Linus
> diff --git a/subsurface-configure.pri b/subsurface-configure.pri
> index b70107f..045c9b1 100644
> --- a/subsurface-configure.pri
> +++ b/subsurface-configure.pri
> @@ -48,9 +48,9 @@ PKG_CONFIG_OUT = $$system(pkg-config --version 2>
> $$NUL)
> # about it if it doesn't.
> LIBS += $$system(pkg-config --libs libusb-1.0 2> /dev/null)
> } else:exists(/usr/local/lib/libdivecomputer.a) {
> - LIBS += -L/usr/local/lib -ldivecomputer
> + LIBS += /usr/local/lib/libdivecomputer.a
> } else:exists(/usr/local/lib64/libdivecomputer.a) {
> - LIBS += -L/usr/local/lib64 -ldivecomputer
> + LIBS += /usr/local/lib64/libdivecomputer.a
> } else:link_pkgconfig {
> # find it via pkg-config
> PKGCONFIG += libdivecomputer
Of course this does
a) break things on Windows again
b) create the insane situation that we link statically against
libdivecomputer but dynamically against libusb which libdivecomputer
needs
NUTS.
/D
More information about the subsurface
mailing list