From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-return-50536-listarch-gcc=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 28350 invoked by alias); 23 Apr 2002 18:42:03 -0000
Mailing-List: contact gcc-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Archive: <http://gcc.gnu.org/ml/gcc/>
List-Post: <mailto:gcc@gcc.gnu.org>
List-Help: <http://gcc.gnu.org/ml/>
Sender: gcc-owner@gcc.gnu.org
Received: (qmail 28336 invoked from network); 23 Apr 2002 18:42:02 -0000
Received: from unknown (HELO dair.pair.com) (209.68.1.49)
  by sources.redhat.com with SMTP; 23 Apr 2002 18:42:02 -0000
Received: (qmail 17256 invoked by uid 20157); 23 Apr 2002 18:42:01 -0000
Received: from localhost (sendmail-bs@127.0.0.1)
  by localhost with SMTP; 23 Apr 2002 18:42:01 -0000
Date: Tue, 23 Apr 2002 11:53:00 -0000
From: Hans-Peter Nilsson <hp@bitrange.com>
X-Sender:  <hp@dair.pair.com>
To:  <bug-automake@gnu.org>,  <gcc@gcc.gnu.org>,  <libstdc++@gcc.gnu.org>
Subject: Cross-testing: check-target-libstdc++-v3 with dejagnu in build tree.
Message-ID: <Pine.BSF.4.30.0204231436330.11541-100000@dair.pair.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
X-SW-Source: 2002-04/txt/msg01188.txt.bz2

I use this patch locally so I can make check-target-libstdc++-v3
with dejagnu in the build tree, using the dejagnu at sourceware
(cf. <URL:http://gcc.gnu.org/simtest-howto.html>) which was
recently synched with FSF dejagnu (thanks to Nick Clifton).

The patch is no good as is, since Makefile.in in
libstdc++-v3/testsuite is generated from Makefile.am, and this
check-DEJAGNU rule is an automake-generated thingy.  I think it
exposes a problem with automake, which seems to by default
assume that libraries are only tested as host libraries, in
which case the library dir is one .. closer to the host expect
build dir.

IIRC the automake documented remedy is to enter the
check-DEJAGNU rule manually in Makefile.am, but I think the
automake-generated rule needs fixing too.  It seems "automake
1.4-p5" was used to generate this Makefile.in (please forgive me
for not digging deeper, particularly into newer automake
versions).  I don't really know how or if the test should be
tweaked to deal with the case where libstdc++-v3 is a multilib.

Thoughts?

Index: Makefile.in
===================================================================
RCS file: /cvs/gcc/gcc/libstdc++-v3/testsuite/Makefile.in,v
retrieving revision 1.20.2.2
diff -p -c -r1.20.2.2 Makefile.in
*** Makefile.in	2 Apr 2002 02:47:07 -0000	1.20.2.2
--- Makefile.in	23 Apr 2002 18:16:04 -0000
*************** RUNTESTDEFAULTFLAGS = --tool $(DEJATOOL)
*** 212,218 ****
  check-DEJAGNU: site.exp
  	srcdir=`cd $(srcdir) && pwd`; export srcdir; \
  	EXPECT=$(EXPECT); export EXPECT; \
! 	if [ -f $(top_builddir)/../expect/expect ]; then \
  	  TCL_LIBRARY=`cd $(top_srcdir)/../tcl/library && pwd`; \
  	  export TCL_LIBRARY; \
  	fi; \
--- 212,219 ----
  check-DEJAGNU: site.exp
  	srcdir=`cd $(srcdir) && pwd`; export srcdir; \
  	EXPECT=$(EXPECT); export EXPECT; \
! 	if [ -f $(top_builddir)/../expect/expect ] \
! 	   || [ -f $(top_builddir)/../../expect/expect ]; then \
  	  TCL_LIBRARY=`cd $(top_srcdir)/../tcl/library && pwd`; \
  	  export TCL_LIBRARY; \
  	fi; \

brgds, H-P