We cann connect() in a non-blocking mode to be able to specify a
non-standard timeout.

The problem was that we did not fetch the status from the
non-blocking connect(). We assumed that poll() would not return
a POLLIN flag if the connect failed. But on some platforms this
is not true.

After a successful poll() we do now retrieve the status value
from connect() with getsockopt(...SO_ERROR...). Now we do know
if (and how) the connect failed.

The test case for my investigation was rpl.rlp_ssl1 on an
Ubuntu 9.04 x86_64 machine. Both, IPV4 and IPV6 were active.
'localhost' resolved first for IPV6 and then for IPV4. The
connection over IPV6 was blocked. rpl.rlp_ssl1 timed out
as it did not notice the failed connect(). The first read()
failed, which was interpreted as a master crash and the
connection was tried to reestablish with the same result
until the retry limit was reached.

With the fix, the connect() problem is immediately recognized,
and the connect() is retried on the second resolution for
'localhost', which is successful.


libmysqld/libmysqld.c:
  Bug#37267 - connect() EINPROGRESS failures mishandled in client library
  Changed a DBUG print string to distinguish the two mysql_real_connect()
  implementations in DBUG traces.
mysql-test/include/wait_for_slave_param.inc:
  Bug#37267 - connect() EINPROGRESS failures mishandled in client library
  Made timeout value available in error message.
mysql-test/suite/rpl/r/rpl_get_master_version_and_clock.result:
  Bug#37267 - connect() EINPROGRESS failures mishandled in client library
  Fixed test result. Connect error is now detected as CR_CONN_HOST_ERROR
  (2003) instead of CR_SERVER_LOST (2013).
sql-common/client.c:
  Bug#37267 - connect() EINPROGRESS failures mishandled in client library
  Added retrieval of the error code from the non-blocking connect().
  Added DBUG.
  Added comment.

DECIMAL and TIMESTAMP used to have NUM_FLAG, but NEWDECIMAL was forgotten.
It's correct that TIMESTAMP does not have the flag nowadays (manual will be updated, connectors
developers will be notified).

client/mysqldump.c:
  IS_NUM_FIELD(f) removed and replaced by its definition (f>flags & NUM_FLAG).
include/mysql.h:
  - IS_NUM_FIELD() is removed because name is too close to IS_NUM() and it is not used a lot
  - INTERNAL_NUM_FIELD() is removed:
    * it forgets to test NEWDECIMAL (when IS_NUM() was updated for NEWDECIMAL we forgot
    to update INTERNAL_NUM_FIELD()), that's why client didn't mark NEWDECIMAL with NUM_FLAG (a bug).
    * it has an obsolete test for length of the TIMESTAMP field: test became accidentally wrong when length
    of TIMESTAMP was changed to always be 19 (when the format was changed from YYYYMMDDhhmmss to
    YYYY-MM-DD hh:mm:ss), never 8 or 14 anymore. That obsolete test caused TIMESTAMP to lose NUM_FLAG,
    which was an accidental but good change (see below).
    * IS_NUM() should be used instead
  - IS_NUM(f) is changed: TIMESTAMP used to be parsable as a number without quotes (when it was formatted as
  "YYYYMMDDhhmmss"); but it is not anymore (now that it is "YYYY-MM-DD hh:mm:ss"), so it should not have NUM_FLAG
  (mysqldump needs to quote TIMESTAMP values), so IS_NUM() should return false for it.
libmysqld/lib_sql.cc:
  use IS_NUM() instead of INTERNAL_NUM_FIELD()
mysql-test/r/bigint.result:
  result change: NEWDECIMAL fields now have NUM_FLAG (32768)
mysql-test/r/metadata.result:
  result change: NEWDECIMAL fields now have NUM_FLAG (32768)
mysql-test/r/mysqldump.result:
  DECIMAL columns are not quoted anymore by mysqldump. Which is ok, the parser does not need '' for them
mysql-test/r/ps_2myisam.result:
  result change: NEWDECIMAL fields now have NUM_FLAG (32768)
mysql-test/r/ps_3innodb.result:
  result change: NEWDECIMAL fields now have NUM_FLAG (32768)
mysql-test/r/ps_4heap.result:
  result change: NEWDECIMAL fields now have NUM_FLAG (32768)
mysql-test/r/ps_5merge.result:
  result change: NEWDECIMAL fields now have NUM_FLAG (32768)
mysql-test/suite/ndb/r/ps_7ndb.result:
  result change: NEWDECIMAL fields now have NUM_FLAG (32768)
mysql-test/t/metadata.test:
  test for BUG#42980
sql-common/client.c:
  use IS_NUM() instead of INTERNAL_NUM_FIELD()

   Solaris binary packages should be compiled with '-g0', not '-g'

Update default.conf.