solved, Re: testsuite oddity

Marc Espie espie@quatramaran.ens.fr
Tue Nov 24 08:10:00 GMT 1998


In article < 199811241345.OAA08042@quatramaran.ens.fr > you write:
>There is an annoying `feature' of the testsuite. If by any chance you get
>a regression in the libio/libstdc++ tests, the corresponding call to runtest
>actually returns an error...  this does stop a make check dead in its tracks.

I must have been really tired... I was concentrating on the ${RUNTEST}
line, and not noticing that it was the end line of a small shell-script.
The libs testsuites Makefile.in differ from the other Makefiles:

Index: libio/testsuite/Makefile.in
===================================================================
RCS file: /egcs/carton/cvsfiles/egcs/libio/testsuite/Makefile.in,v
retrieving revision 1.3
diff -u -r1.3 Makefile.in
--- Makefile.in	1998/10/08 23:13:20	1.3
+++ Makefile.in	1998/11/24 16:08:47
@@ -80,7 +80,7 @@
 	-@rm -f ./tmp?
 
 just-check: site.exp
-	rootme=`pwd`; export rootme; \
+	-rootme=`pwd`; export rootme; \
 	srcdir=${srcdir} ; export srcdir ; \
 	EXPECT=${EXPECT} ; export EXPECT ; \
 	if [ -f $${rootme}/../../../expect/expect ] ; then  \
Index: libstdc++/testsuite/Makefile.in
===================================================================
RCS file: /egcs/carton/cvsfiles/egcs/libstdc++/testsuite/Makefile.in,v
retrieving revision 1.2
diff -u -r1.2 Makefile.in
--- Makefile.in	1997/08/26 18:28:26	1.2
+++ Makefile.in	1998/11/24 16:08:48
@@ -57,7 +57,7 @@
 	-@rm -f ./tmp?
 
 just-check: site.exp
-	rootme=`pwd`; export rootme; \
+	-rootme=`pwd`; export rootme; \
 	srcdir=${srcdir} ; export srcdir ; \
 	EXPECT=${EXPECT} ; export EXPECT ; \
 	if [ -f $${rootme}/../../expect/expect ] ; then  \



More information about the Gcc-bugs mailing list