git-2.7.4 (was 2.7.2, 2.7.3)

[ken@…]

Newer version.

Still 159 test failures identified by 'not ok', all but one had '# TODO known breakage' - this output is painful to read. The failure is in [t7300-clean.sh] not ok 32 - should avoid cleaning possible submodules.

Looks as if the grep-2.23 test failures have been fixed, I'll need to review the commented code for using multiple jobs with make check.

From the release notes for 2.7.4:

• Bugfix patches were backported from the 'master' front to plug heap corruption holes, to catch integer overflow in the computation of pathname lengths, and to get rid of the name_path API. Both of these would have resulted in writing over an under-allocated buffer when formulating pathnames while tree traversal.

[Fernando de Oliveira]

First, thank you, Bruce and Pierre for some comments.

I doubt to be able to come back before 7.10 is released.

Now to topic.

All in 7.8, not 7.9, but some could help. No failure here.

The tests run in parallel.

But to run just one test, got to the directory (cd t) and execute:

sh t7300-clean.sh

Sorry to write obvious things, but sometimes we never think about them. I use:

make_test ()
{
  echo "make -k test START" &&
  { time \
    {
      MKFLGSTEST=$(($(getconf _NPROCESSORS_ONLN) / 2)) &&
      env LC_ALL=C make -j${MKFLGSTEST} -k test
      # The scripts in the test suite can be run with the "-x" option
      # to show a shell-trace of each command they run.
      #env LC_ALL=C make -j${MKFLGSTEST} -k -x test
      echo
    }
    echo &&
    grep 'not ok ' $LOGDIR/$PACKAGE-make-k-test-$DATE.log | \
    grep -v 'TODO known breakage' &&
    echo &&
    echo "Number of failed tests:" &&
    grep 'not ok ' $LOGDIR/$PACKAGE-make-k-test-$DATE.log | \
    grep -v 'TODO known breakage' | wc -l || or echo "0" &&
    echo
  } 2>&1 | tee $LOGDIR/$PACKAGE-make-k-test-$DATE.log &&
  xz -v $LOGDIR/$PACKAGE-make-k-test-$DATE.log &&
  echo "make -k test END"
} &&
make_test

[ken@…]

When I run the tests, as a user after upgrading it, they all pass. I did have a note that one failed if run as root [ and it did in my install ].

[ken@…]

Replying to ken@…:

When I run the tests, as a user after upgrading it, they all pass. I did have a note that one failed if run as root [ and it did in my install ].

Same for a regular user before upgrading, and now there is a nice summary at the end (but not when run as root!).

Fixed at r17128.