This is the mail archive of the gdb-patches@sources.redhat.com mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[patch/rfc] Use i686-pc-linux-gnu as the releng target


Hello,

GDB, and I suspect binutils, and just possibly newlib (true?), as part of the release process does:
./configure sun4
(it is so that the .info et.al. files can be generated). The problem is that GDB no longer configures for the good old sun4.


This patch changes the system being used from "sun4" to "i686-pc-linux-gnu".

For binutils, can the release engineer give this a wirl.

For newlib, is this logic still used? Can that part of the file be removed?

Once we've sorted things here I'll see about pushing this upstream into GCC repository (gcc doesn't use this file).

Andrew
2004-01-12  Andrew Cagney  <cagney@redhat.com>

	* src-release: Update copyright year.
	(do-proto-toplev): Configure using i686-pc-linux-gnu.
	(NEWLIB_SUPPORT_DIRS): Update comments.

Index: src-release
===================================================================
RCS file: /cvs/src/src/src-release,v
retrieving revision 1.8
diff -u -r1.8 src-release
--- src-release	14 Sep 2003 23:16:45 -0000	1.8
+++ src-release	12 Jan 2004 15:51:13 -0000
@@ -1,5 +1,5 @@
 #   Copyright (C) 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
-#   1999, 2000, 2001, 2002, 2003 Free Software Foundation
+#   1999, 2000, 2001, 2002, 2003, 2004 Free Software Foundation
 #
 # This file is free software; you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
@@ -136,7 +136,7 @@
 	<Makefile.in >tmp
 	mv -f tmp Makefile.in
 	#
-	./configure sun4
+	./configure i686-pc-linux-gnu
 	$(MAKE) configure-host configure-target \
 	    ALL_GCC="" ALL_GCC_C="" ALL_GCC_CXX="" \
 	    CC_FOR_TARGET="$(CC)" CXX_FOR_TARGET="$(CXX)"
@@ -331,17 +331,18 @@
 
 .PHONY: newlib.tar.bz2
 NEWLIB_SUPPORT_DIRS=libgloss
-# taz configures for the sun4 target which won't configure newlib.
-# We need newlib configured so that the .info files are made.
-# Unfortunately, it is not enough to just configure newlib separately:
-# taz will build the .info files but since SUBDIRS won't contain newlib,
-# distclean won't be run (leaving Makefile, config.status, and the tmp files
-# used in building the .info files, eg: *.def, *.ref).
-# The problem isn't solvable however without a lot of extra work because
-# target libraries are built in subdir $(target_alias) which gets nuked during
-# the make distclean.  For now punt on the issue of shipping newlib info files
-# with newlib net releases and wait for a day when some native target (sun4?)
-# supports newlib (if only minimally).
+# taz configures for the i686-pc-linux-gnu (nee sun4) target which may
+# not configure newlib.  We need newlib configured so that the .info
+# files are made.  Unfortunately, it is not enough to just configure
+# newlib separately: taz will build the .info files but since SUBDIRS
+# won't contain newlib, distclean won't be run (leaving Makefile,
+# config.status, and the tmp files used in building the .info files,
+# eg: *.def, *.ref).  The problem isn't solvable however without a lot
+# of extra work because target libraries are built in subdir
+# $(target_alias) which gets nuked during the make distclean.  For now
+# punt on the issue of shipping newlib info files with newlib net
+# releases and wait for a day when some native target supports newlib
+# (if only minimally).
 newlib.tar.bz2: $(DIST_SUPPORT) $(NEWLIB_SUPPORT_DIRS) newlib
 	$(MAKE) -f $(SELF) taz TOOL=newlib \
 		MD5PROG="$(MD5PROG)" \

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]