Re: building libgtop

On Tue, 2004-02-17 at 07:29, Eugenia Loli-Queru wrote:
> > This is a classic aclocal/automake version issue that should be resolved
> > by demanding for specific aclocal/automake versions which is IMO planned
> > for gnome-common, but won't be in GNOME 2.6.
> > The problem here is that aclocal-1.4 doesn't complain about possible
> > CFLAGS redifinitions in while 1.7 requires you to do this
> > using a special variable (AM_CFLAGS). Unfortunately ac 1.4 doesn't know
> > the AM_PROG_AS macro that 1.7 needs for inodedb_SRCLIST / inodedb_DEFS
> > which seems to be treated as asm code.
> > I don't know what to do about it. For the moment, you could manually re-
> > run the 1.4 tools in this dir and manually execute autoconf.
> I manage to build Gnome from CVS after staying up to 5 AM yesterday, thanks
> :)
> (however there were a few real bugs on the way, like gnome-applets not
> building due to some glib unresolved symbols, gtkhtml not generating a
> "libtool" file, gnome-panel's libecal dependency on
> evolution-data-server/libsoup was not addressed by jhbuild etc...)

Have you filed bugs about all of these (at least for the ones that are
real bugs and not just you deciding to change the automake requirements
yourself)? For the record, I also built GNOME from CVS yesterday, into a
clean directory from clean sources and saw none of these problems. The
gnome-panel one, for example, is an optional dependency that only occurs
if eds is installed.

> I hope a solution will be found to fix these issues (including the jhbuild
> issues)  to be able to build Gnome, so we beta testers can have an easier
> ride with it, because I ain't going back to 1.4.x versions, in fact
> Slackware just updated today from 1.7 to 1.8.x.

Then you are doomed because there are some packages that *require*
automake-1.4 and that we are not really pushing to change this at such a
late date in the release process. Destabilising the tree at this point
for the benefit of people building from CVS who could fix their problems
by installing one tarball is not a good risk-reward equation.

We have plans to fix this in the 2.7 cycle.


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]