Some tests are skipped by checks in suite.pm. It is redundant to
have an sql-level run-time check in the .inc file itself.
In some cases it's not only redundant, but dangerous.
After one bug in 10.2 innodb.create_isl_with_direct failed
to start InnoDB, but the server started fine (just without InnoDB)
and instead of failing, the test was skipped by run-time check in
have_innodb.inc.
# Conflicts:
# mysql-test/include/not_embedded.inc
# mysql-test/r/change_user_notembedded.result
# mysql-test/suite.pm
# mysql-test/t/change_user_notembedded.test
and use it in mysql-test-run.pl
Use standard defines as last argument to 'fn_format'
mysql-test/r/not_ndb_default.require:
Rename: mysql-test/r/true.require -> mysql-test/r/not_ndb_default.require
client/mysqltest.c:
Use proper defines for last argument to fn_format(they are defined in my_sys.h)
Make abort_not_supported_test print info about in what file the test was
aborted and a reason for it. The reason is a best guess, using the filename of
the .require file to format it.
mysql-test/include/not_ndb_default.inc:
Use a more descriptive name of the .require file
mysql-test/include/not_windows.inc:
Use a more descriptive name of the .require file
mysql-test/include/windows.inc:
Use a more descriptive name of the .require file
mysql-test/lib/mtr_io.pl:
Add new function 'mtr_lastlinefromfile' that will return only the last line of a file.
mysql-test/mysql-test-run.pl:
Get reason why mysqltest decided to abort testcase from mysqltest.log
mysql-test/r/not_windows.require:
New BitKeeper file ``mysql-test/r/not_windows.require''
mysql-test/r/windows.require:
New BitKeeper file ``mysql-test/r/windows.require''
that platform, not real failures. (Bug #12328, Bug #11569)
mysql-test/t/packet.test:
Skip this test on Windows
mysql-test/t/rpl_flush_tables.test:
Skip this test on Windows
mysql-test/include/not_windows.inc:
Add include for skipping tests on Windows
