libjava make check, timing out again as of today

Gordon Sadler gbsadler1@lcisp.com
Fri Apr 13 09:59:00 GMT 2001


On Fri, Apr 13, 2001 at 09:43:47AM -0600, Tom Tromey wrote:
> >>>>> "Gordon" == Gordon Sadler <gbsadler1@lcisp.com> writes:
> 
> Gordon> Well, good job. This has been resolved. 
> 
> Hmm.  Did anybody actually do anything to fix it?  I know I didn't;
> your message is still on my to-look-at list.
> 
> Gordon> Last night's bootstrap/check did just fine. Libjava is down to
> Gordon> a handful of problem cases.
> 
> What platform?  How did you configure?
> I have 0 FAIL on x86 RH Linux 6.2.
>
I've been doing daily bootstraps for gcc-3_0-branch since the beginning
of Mar 2001 or so. I've sent results to gcc-testresults if it was
possible/built and I could get a sensible test.log.

uname -a
Linux home-desktop 2.4.3-ac4 #1 Wed Apr 11 17:39:50 CDT 2001 i686
unknown

Debian unstable. Bootstrapping with gcc 2.95.3.

Last night's testresult:
http://gcc.gnu.org/ml/gcc-testresults/2001-04/msg00239.html
Failing testresult(after I killed make check libjava and let it finish):
http://gcc.gnu.org/ml/gcc-testresults/2001-04/msg00223.html

My failure of the libjava testsuite is with the same kernel.. even
though I compile new ones regularly. It failed on Apr 12, worked fine
Apr 13. Same kernel/glibc/binutils/gcc.


I couldn't find anything specific... maybe

cvs diff -r1.129.2.6 -r1.129.2.7 libjava/Makefile.am

@@ -143,11 +140,10 @@ libgcj_la_DEPENDENCIES = libgcj.jar $(ja
-libgcj_la_LDFLAGS = -L$(here)/../libstdc++-v3/libsupc++ -lsupc++ \
-       -rpath $(toolexeclibdir) $(THREADLIBS) $(LIBLTDL) \
+libgcj_la_LDFLAGS = -rpath $(toolexeclibdir) $(THREADLIBS) $(LIBLTDL) \

@@ -157,7 +153,6 @@ EXTRA_libgcjx_la_SOURCES = $(x_java_sour
libgcjx_la_DEPENDENCIES = libgcj.jar $(x_javao_files)
libgcjx_la_LIBADD = $(x_javao_files)
libgcjx_la_LDFLAGS = @X_PRE_LIBS@ @X_LIBS@ -lX11 @X_EXTRA_LIBS@ \
-        -L$(here)/../libstdc++-v3/libsupc++ -lsupc++ \
         -rpath $(toolexeclibdir) \
			
@@ -267,13 +262,16 @@ $(javao_files) $(x_javao_files): %.lo: %
 libgcj.la: $(libgcj_la_OBJECTS) $(libgcj_la_DEPENDENCIES)
         @echo $(libgcj_la_OBJECTS) > libgcj.objectlist;
		 @echo $(libgcj_la_LIBADD) >> libgcj.objectlist;
-       $(libgcj_la_LINK) -objectlist libgcj.objectlist
		-rpath $(toolexeclibdir) $(libgcj_la_LDFLAGS) $(LIBS)
+       $(libgcj_la_LINK) -objectlist libgcj.objectlist \
+       ../libstdc++-v3/libsupc++/libsupc++convenience.la \
+       ../libffi/libfficonvenience.la \
+       -rpath $(toolexeclibdir) $(libgcj_la_LDFLAGS) $(LIBS)
						 

Only other changes inside libjava dir since my previous bootstrap:

java/io/File.java
testsuite/libjava.compile/PR207.xfail
testsuite/libjava.compile/PR208.xfail

Is it possible a change to gcj itself may have resolved this?

I have a cronjob that cvs updates every morning at 2:30 am CST then
configure/bootstrap/check.



More information about the Gcc mailing list