Pure LFS, gcc 3.2.2 and test_summary

S. Anthony Sequeira tony at sequestor.com
Fri Apr 18 07:34:51 PDT 2003


On Fri, 18 Apr 2003 13:39:07 +0000, Zack Winkles wrote:

> S. Anthony Sequeira wrote:
> <snip lots of stuff>
> 
> Are you sure you used 'make -k check' and not just regular old 'make
> check'? It seems like it detects that one failure in the testsuite and
> stops, which it shouldn't do because of the '-k' flag. Check and
> double-check that you're using the right command.

I cut and paste from the hints doc.

It didn't stop, it went on.  I am just concerned about the difference in
the test results.

Here are the *exact* commands I used (wrapped by pan):

    mkdir ../gcc-build &&     cd ../gcc-build &&     CFLAGS="-O2 -pipe"
    CXXFLAGS="-O2 -pipe" ../gcc-3*/configure
  --prefix=/stage1 --with-local-prefix=/stage1 --enable-languages=c,c++
      --enable-shared --enable-threads=posix --enable-__cxa_atexit
  --enable-clocale=gnu &&     make LDFLAGS="-s" &&     make -k check

followed by

make install

I will look through my log file again.  What you say about stopping seems
to make some kind of sense.  Though, as I said, I seem to have exactly the
same results for every PLFS build I have done, which caused me to restart
the first couple of times.

Here are the relevant bits from the log file:

FAIL: 26_numerics/c99_classification_macros_c.cc (test for excess errors)^M
^M
                === libstdc++-v3 Summary ===^M
^M
# of expected passes            432^M
# of unexpected failures        1^M
# of unexpected successes       26^M
make[4]: *** [check-DEJAGNU] Error 1^M
make[4]: Leaving directory `/mnt/lfs/stage1/gcc-build/i686-pc-linux-gnu/libstdc++-v3/testsuite'^M
make[3]: *** [check-am] Error 2^M
make[3]: Target `check' not remade because of errors.^M
make[3]: Leaving directory `/mnt/lfs/stage1/gcc-build/i686-pc-linux-gnu/libstdc++-v3/testsuite'^M
make[3]: Entering directory `/mnt/lfs/stage1/gcc-build/i686-pc-linux-gnu/libstdc++-v3'^M
make[3]: Nothing to be done for `check-am'.^M
make[3]: Leaving directory `/mnt/lfs/stage1/gcc-build/i686-pc-linux-gnu/libstdc++-v3'^M
make[2]: *** [check-recursive] Error 1^M
make[2]: Target `check' not remade because of errors.^M
make[2]: Leaving directory `/mnt/lfs/stage1/gcc-build/i686-pc-linux-gnu/libstdc++-v3'^M
make[1]: *** [check-target-libstdc++-v3] Error 2^M
make[2]: Entering directory `/mnt/lfs/stage1/gcc-build/i686-pc-linux-gnu/libiberty'^M
make[3]: Entering directory `/mnt/lfs/stage1/gcc-build/i686-pc-linux-gnu/libiberty/testsuite'^M
make[3]: Nothing to be done for `check'.^M
make[3]: Leaving directory `/mnt/lfs/stage1/gcc-build/i686-pc-linux-gnu/libiberty/testsuite'^M
make[2]: Leaving directory `/mnt/lfs/stage1/gcc-build/i686-pc-linux-gnu/libiberty'^M
make[2]: Entering directory `/mnt/lfs/stage1/gcc-build/gcc'^M
Making a new config file...^M
echo "set tmpdir /mnt/lfs/stage1/gcc-build/gcc/testsuite" >> ./tmp0^M
test -d testsuite || mkdir testsuite^M
rm -f testsuite/site.exp^M
sed '/set tmpdir/ s|testsuite|testsuite|' < site.exp > testsuite/site.exp^M
(rootme=`pwd`; export rootme; \^M
srcdir=`cd ../../gcc-3.2.2/gcc; pwd` ; export srcdir ; \^M
cd testsuite; \^M
EXPECT=expect ; export EXPECT ; \^M
if [ -f ${rootme}/../expect/expect ] ; then  \^M
   TCL_LIBRARY=`cd .. ; cd ../../gcc-3.2.2/gcc/../tcl/library ; pwd` ; \^M
   export TCL_LIBRARY ; fi ; \^M
runtest --tool gcc )^M
WARNING: Couldn't find the global config file.^M
WARNING: Couldn't find tool init file^M
Test Run By lfs on Fri Apr 18 10:08:43 2003^M
Native configuration is i686-pc-linux-gnu^M
^M
                === gcc tests ===^M
^M
Schedule of variations:^M
    unix^M
^M
Running target unix^M
Using /stage1/share/dejagnu/baseboards/unix.exp as board description file for target.^M
Using /stage1/share/dejagnu/config/unix.exp as generic interface file for target.^M
WARNING: Couldn't find tool config file for unix, using default.^M
^M
                === gcc Summary ===^M


I won't post any more, unless you need it.

Cheers
-- 
Tony

-- 
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe lfs-dev' in the subject header of the message



More information about the lfs-dev mailing list