Re: cvs commit: src/share/mk

[YONETANI Tomokazu <qhwt+dfly@xxxxxxxxxx>]

On Mon, Nov 15, 2004 at 03:13:19PM +0900, YONETANI Tomokazu wrote:
> On Sat, Nov 13, 2004 at 08:52:57AM -0800, Joerg Sonnenberger wrote:
> > joerg       2004/11/13 08:52:36 PST
> > 
> > DragonFly src repository
> > 
> >   Modified files:
> >     share/mk             bsd.dep.mk 
> >   Log:
> >   Fix the condition under which yacc -o is used. This should fix the parallel
> >   build in gdb.
> >   
> >   Obtained-from: FreeBSD
> >   Problems-reported-by: various folks
> >   
> >   Revision  Changes    Path
> >   1.6       +2 -2      src/share/mk/bsd.dep.mk
> 
> Even after this fix(actuall 1.7), make -j3 depend in src/gnu/usr.bin/gdb/gdb
> fails(make -j2 depend succeeds BTW).

This change in revision 1.34 of gdb/Makefile in FreeBSD looks relavant
to me.

%%%
Index: Makefile
===================================================================
RCS file: /freebsd/cvs/src/gnu/usr.bin/gdb/gdb/Makefile,v
retrieving revision 1.33
retrieving revision 1.34
diff -u -r1.33 -r1.34
--- Makefile	1 May 1998 14:44:00 -0000	1.33
+++ Makefile	4 May 1998 17:56:22 -0000	1.34
@@ -1,4 +1,4 @@
-#	$Id: Makefile,v 1.32 1998/04/30 08:03:49 dfr Exp $
+#	$Id: Makefile,v 1.33 1998/05/01 14:44:00 bde Exp $
 
 PROG = gdb
 
@@ -32,7 +32,7 @@
 	typeprint.c utils.c valarith.c valops.c \
 	valprint.c values.c version.c \
 	serial.c ser-unix.c ser-tcp.c mdebugread.c \
-	c-exp.c f-exp.c m2-exp.c i387-tdep.c \
+	c-exp.y f-exp.y m2-exp.y i387-tdep.c \
 	kvm-fbsd.c bcache.c \
 	corefile.c ch-exp.c f-lang.c scm-exp.c scm-lang.c \
 	scm-valprint.c f-typeprint.c f-valprint.c nlmread.c \
@@ -45,8 +45,9 @@
 CFLAGS+= -DNO_MMALLOC
 # uncomment the next line if you want to debug gdb
 #CFLAGS+= -g
+YFLAGS=
 
-CLEANFILES+=	c-exp.c f-exp.c m2-exp.c init.c y.tab.h init.c-tmp
+CLEANFILES=	init.c init.c-tmp
 
 .if ${BINFORMAT} == elf
 
@@ -83,8 +84,6 @@
 DPADD+=	${LIBIBERTY}
 LDADD+=	${LIBIBERTY}
 
-.ORDER:	c-exp.c f-exp.c m2-exp.c
-
 # We do this by grepping through sources.  If that turns out to be too slow,
 # maybe we could just require every .o file to have an initialization routine
 # of a given name (remote-udi.o -> _initialize_remote_udi, etc.).
%%%


And the following smallest change to SRCS list seems to fix (not thoroughly
tested the compiled binary yet) the make -jN depend of gdb.

Index: Makefile
===================================================================
RCS file: /dragonfly/cvs/src/gnu/usr.bin/gdb/gdb/Makefile,v
retrieving revision 1.4
diff -u -r1.4 Makefile
--- Makefile	14 Nov 2004 15:16:59 -0000	1.4
+++ Makefile	15 Nov 2004 11:17:06 -0000
@@ -28,14 +28,14 @@
 # libgdb
 SRCS+=	annotate.c arch-utils.c auxv.c ax-gdb.c ax-general.c bcache.c
 SRCS+=	bfd-target.c block.c blockframe.c breakpoint.c bsd-kvm.c buildsym.c
-SRCS+=	c-exp.c c-lang.c c-typeprint.c c-valprint.c charset.c cli-cmds.c
+SRCS+=	c-exp.y c-lang.c c-typeprint.c c-valprint.c charset.c cli-cmds.c
 SRCS+=	cli-decode.c cli-dump.c cli-interp.c cli-logging.c cli-out.c
 SRCS+=	cli-script.c cli-setshow.c cli-utils.c coff-pe-read.c coffread.c
 SRCS+=	complaints.c completer.c copying.c corefile.c corelow.c cp-abi.c
 SRCS+=	cp-namespace.c cp-support.c cp-valprint.c dbxread.c dcache.c demangle.c
 SRCS+=	dictionary.c disasm.c doublest.c dummy-frame.c dwarf2-frame.c
 SRCS+=	dwarf2expr.c dwarf2loc.c dwarf2read.c dwarfread.c elfread.c environ.c
-SRCS+=	eval.c event-loop.c event-top.c exec.c expprint.c f-exp.c f-lang.c
+SRCS+=	eval.c event-loop.c event-top.c exec.c expprint.c f-exp.y f-lang.c
 SRCS+=	f-typeprint.c f-valprint.c fbsd-proc.c findvar.c fork-child.c
 SRCS+=	frame-base.c frame-unwind.c frame.c gcore.c gdb-events.c gdbarch.c
 SRCS+=	gdbtypes.c gnu-v2-abi.c gnu-v3-abi.c hpacc-abi.c inf-loop.c infcall.c