by MAX/MIN in some cases:
SELECT coercibility(max(s1)) from t1;

Subselect collation and derivation was not processed
correctly:

create table a select (select s1 from t1);
select * from t1 where s1 = (select s2 from t1);

  Revise 4.1 Windows-install instructions.


Docs/README.1st:
  Revise 4.1 Windows-install instructions.

CREATE TABLE t1 (s1 CHAR(5) COLLATE latin1_german1_ci,
                 s2 CHAR(5) COLLATE latin1_swedish_ci);
SELECT * FROM t1 WHERE s1 = s2;

into deer.(none):/home/hf/work/mysql-4.1.clean

  Logging to logging@openlogging.org accepted
README.1st:
  new file - Readme for Windows distribution


BitKeeper/etc/logging_ok:
  Logging to logging@openlogging.org accepted

into deer.(none):/home/hf/work/mysql-4.1.clean

sql-common/client.c:
  SOCKET_EINTR is handled differently working in replication

bugs #701 and #783 - old .frm files were read in incorrectly in 4.1 (each field was assigned a charset, BLOBs were converted to TEXT)

into eagle.mysql.r18.ru:/home/vva/work/BUG_663/mysql-4.1


sql/sql_base.cc:
  Auto merged

mysql-test/r/auto_increment.result:
  added test for AUTO_VALUE_ON_ZERO
mysql-test/t/auto_increment.test:
  added test for AUTO_VALUE_ON_ZERO

into serg.mylan:/usr/home/serg/Abk/mysql-4.1


sql/sql_acl.cc:
  Auto merged
sql/sql_base.cc:
  Auto merged

into mysql.com:/home/mysql_src/mysql-4.1

have in fact no effect, because in 4.0 the slave always runs with --skip-innodb.
In 4.1, mysqld accepts option --innodb which can override --skip-innodb.
So I add this option to the test.
When the updated test (using InnoDB tables) from 4.0 will be merged
into 4.1, with this good .opt file, we'll have a meaningful test in 4.1.
Until then, the test in 4.1 will pass but be useless.
And in 4.0, it will stay useless forever.
("it" = the part of the test which tests replication of foreign_key_checks)


BitKeeper/deleted/.del-rpl_insert_id-slave.opt~ba10e66ea50b556:
  Delete: mysql-test/rpl_insert_id-slave.opt

into teton.kitebird.com:/home/paul/mysql-4.1

Proposed fix for bug#909

The latter is misleading because differs from the syntax used for GRANT/REVOKE.

Fix up test result affected by error message edits.


mysql-test/r/grant_cache.result:
  Fix up test result to match error message edit.
sql/share/czech/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/danish/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/dutch/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/english/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/estonian/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/french/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/german/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/greek/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/hungarian/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/italian/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/japanese/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/korean/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/norwegian-ny/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/norwegian/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/polish/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/portuguese/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/romanian/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/russian/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/serbian/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/slovak/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/spanish/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/swedish/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.
sql/share/ukrainian/errmsg.txt:
  Print user@host in syntax actually used for GRANT/REVOKE.

into mysql.com:/home/dlenev/src/mysql-4.1-local


BitKeeper/etc/logging_ok:
  auto-union

ignorant about ssl_type attribute


sql/sql_acl.cc:
  Now acl_getroot() honors ssl_type attribute even if we compile without openssl
BitKeeper/etc/logging_ok:
  Logging to logging@openlogging.org accepted

This is to avoid a test failure, which is fixed in 4.0 in
ChangeSet@1.1455.34.1, 2003-06-10 23:29:49+02:00, guilhem@mysql.com
by making RESET SLAVE reset the error.


mysql-test/r/rpl_loaddata.result:
  temporary modification until 4.0 is merged in 4.1
mysql-test/t/rpl_loaddata.test:
  temporary modification until 4.0 is merged in 4.1

into mysql.com:/home/mysql_src/mysql-4.1

and other replicate-*-table options in SHOW SLAVE STATUS.
Seems like it had not been done, so I push it now:
there's 4 new columns to SHOW SLAVE STATUS.


mysql-test/r/rpl000015.result:
  Result update (more columns)
mysql-test/r/rpl_empty_master_crash.result:
  Result update (more columns)
mysql-test/r/rpl_flush_log_loop.result:
  Result update (more columns)
mysql-test/r/rpl_log.result:
  Result update (more columns)
mysql-test/r/rpl_log_pos.result:
  Result update (more columns)
mysql-test/r/rpl_redirect.result:
  Result update (more columns)
mysql-test/r/rpl_replicate_do.result:
  Result update (more columns)
mysql-test/r/rpl_rotate_logs.result:
  Result update (more columns)
mysql-test/t/rpl_replicate_do.test:
  Result update (more columns)
sql/slave.cc:
  Print replicate_*_table options in SHOW SLAVE STATUS
sql/slave.h:
  two functions to make a string of replicate_*_table lists.

into deer.(none):/home/hf/work/mysql-4.1.2way

error handling fixed
fetch_lengths made to work differently in embedded and client cases


include/mysql.h:
  removed 'embedded' error containers - they're unnecessary now
  added declarations for fetch_lengths to be 'virtual'
include/sql_common.h:
  fetch_lengths declaration removed
libmysql/libmysql.c:
  implementations for fetch_lengths to be 'virtual' added
libmysqld/lib_sql.cc:
  error informations now is moved from thd->net to mysql-net
libmysqld/libmysqld.c:
  error data is in mysql->net now
sql-common/client.c:
  we have to return old fetch_lengths implementation for 'client' case
sql/protocol.cc:
  handling of sqlstate for embedded library added

into eagle.mysql.r18.ru:/home/vva/work/BUG_663/mysql-4.1


sql/field_conv.cc:
  Auto merged
sql/mysql_priv.h:
  Auto merged
sql/mysqld.cc:
  Auto merged
sql/sql_insert.cc:
  Auto merged

into deer.(none):/home/hf/work/mysql-4.1.2way

  tweak


INSTALL-WIN-SOURCE:
  tweak

sql/sql_insert.cc:
  Undo warnings for single row inserts
mysql-test/t/warnings.test:
  undo one row warning instead use sql_warnings=1
mysql-test/r/warnings.result:
  Fix result file

into deer.(none):/home/hf/work/mysql-4.1.2way

include/mysql.h:
  mysql_close isn't virtual nown

tests/client_test.c:
  test for timestamp bug (BR #819)
sql/sql_insert.cc:
  Enable warnings by default for single value list
  protocol
mysql-test/t/warnings.test:
  Updated test for single value list insert warning
mysql-test/r/warnings.result:
  Updated warnings results

into serg.mylan:/usr/home/serg/Abk/mysql-4.1


mysql-test/r/innodb.result:
  Auto merged