Re: Build System Changes

Hi Robert,

> -----Urspr=FCngliche Nachricht-----
> Von: Robert Schwebel <[email protected]>
> Gesendet: 07.02.07 22:47:05
> An: [email protected]
> Betreff: [DotGNU] Build System Changes


> Hi,=20
>=20
> Just some short notes about today's irc. I'm currently offline during
> the day because of workshop for a customer, so offline communication is
> better than irc atm :-)
>=20
> > [08:49:05]<klausT|work> replacing ranlib with libtool in pnet somehow
> > didn't work on my box
>=20
> What happened=3F

libtool complained about file not found.
I haven't looked at this further because i wanted first to get a working v=
ersion on my boxes.
We can look at this issue later.

>=20
> > [09:01:36]<Brubbel> running autoheader, automake, autoconf
>=20
> We should re-add the AM=5FMAINTAINER=5FMODE. In the meantime I've found out =
that,
> if you have it in, you can enable the auto-generation of the generated f=
iles
> (that's what you want as a maintainer) by adding --enable-maintainer-mod=
e;
> whereas normal users aren't bothered by magically starting autotools (us=
ers are
> not supposed to have autotools installed).
>=20

Ok. I'll re-add the AM=5FMAINTAINER=5FMODE to the configure.in of treecc.

> > [09:02:38]<Brubbel> what does make distcheck do=3F
>=20
> It runs "make dist", which packages the tarball according to the project=

> rules. It then extracts the packaged tarball to "pnet-<version>",
> creates a =5Fbuild/ directory in that dir and makes all sources + dirs of
> the source tree read-only. Then it makes an out-of-tree build in
> =5Fbuild/, runs the testsuite and, if everything went right, makes
> distclean and checks of some files are left in the build dir, which
> would mean that the clean rules are incorrect. If everything went right,=

> you have a tarball (tar.gz or tar.bz2) which is ready for release.
>=20
> Short - it does what you want to do if you want to make a release :-)
>=20
> > [09:03:18]<Brubbel> ditcheck error: treecc is required to build and
> > can be obtained from ...
>=20
> treecc has to be in your path. If you don't have it there regularly,
> please just run "PATH=3D/path/to/treenet:$PATH ./configure ...".
>=20
> > [09:04:07]<Brubbel> where is the tarball stored =3F
>=20
> Directly in the toplevel directory, where you started configure.
>=20
> > [09:05:44]<Brubbel> I'M configuring with --prefix=3Dxxx
>=20
> --prefix specifies the path in your "target directory". Note that this
> isn't necessarily identical with where it is really installed; for
> example, you may want to have the stuff installed on your embedded
> system in the user hierarchy, so you would --prefix=3D/usr. Nevertheless,
> if you have for example your nfsroot dir for the devel board in
> $HOME/root/, you would run 'make install DESTDIR=3D$HOME/root' and it
> would magically install into that dir. Nevertheless, if some of the
> project files had paths compiled in, they would corectly refer to the /
> as seen by embedded system, not by the host.
>=20
> > [09:06:46]<klausT|work> treecc is required to be somewhere in the $PAT=
H
> > [09:07:05]<Brubbel> uuu that's bad
>=20
> It wasn't different before the changes.
>=20
> > [09:09:41]<Brubbel> configure: error: C compiler cannot create executa=
bles
> > [09:10:13]<Brubbel> error: ../../libffi/configure failed for libffi
>=20

The strange thing here was an added -m64 to the call of gcc. I dunno where=
 that comes from.

> That's pretty strange. Did you find out why=3F If not, please send me your=

> config.log. I'll retest with today's CVS.
>=20
> > [09:31:05]<Brubbel> configure: error: ../../libffi/configure failed fo=
r libffi
>=20
> ../.. looks strange. Why does something try to use that path here=3F
>=20
> Robert
> --=20

Klaus
=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F
Viren-Scan f=FCr Ihren PC! Jetzt f=FCr jeden. Sofort, online und kostenlos.
Gleich testen! http://www.pc-sicherheit.web.de/freescan/=3Fmc=3D022222