On Sun, 3 Jun 2001, Greg Stein wrote:
> On Sun, Jun 03, 2001 at 11:08:55PM -0700, Mo DeJong wrote:
> > Hello.
> >
> > Here is a small little patch to get the toplevel Makefile
> > going when buildir != srcdir (this does not fix things,
> > it just gets the process started).
>
> It should be possible to use top_srcdir for this, rather than defining a new
> srcdir symbol. I agree with your change... checking in now (with some other
> stuff I'm checking in).
That sounds fine. It looks like these two changes will also be needed:
Index: Makefile.in
===================================================================
RCS file: /usr/local/tigris/data/helm/cvs/repository/subversion/Makefile.in,v
retrieving revision 1.10
diff -u -1 -r1.10 Makefile.in
--- Makefile.in 2001/06/04 06:40:32 1.10
+++ Makefile.in 2001/06/04 06:53:54
@@ -5,2 +6,3 @@
abs_srcdir = @abs_srcdir@
+VPATH = @top_srcdir@
@@ -124,3 +126,3 @@
# recursive targets to handle external projects (e.g. APR, Neon)
-external-all external-clean external-distclean external--install:
+external-all external-clean external-distclean external-install:
@target=`echo $@ | sed s/external-//`; \
Also, there seems to be a big problem in build.conf WRT libsvn_fs.
If the build is not configured with libsvn_fs then a `make`
works just fine but a `make install` will blow up since
the install-lib: rule wants to install everything.
Don't you think life would be a lot easier if folks just
assume that the fs will always get built and add
my patch to build berkeley db from the src tree?
Mo DeJong
Red Hat Inc
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Oct 21 14:36:31 2006