storage/maria/unittest/ma_pagecache_single.c:
  Check flush result.

Fix of unit tests.
2 new tests for new features.


storage/maria/ma_control_file.c:
  Getting the definition from the file for unit-test.
storage/maria/ma_loghandler.c:
  Fix bug with skipped syncing.
storage/maria/unittest/ma_control_file-t.c:
  Fix of unit test of control file.
  2 new tests for new features of control file.

Fixed bugs in undo logging
Fixed bug where head block was split before min_row_length (caused Maria to believe row was crashed on read)
Reserved place for reference-transid on key pages (for packing of transids)
ALTER TABLE and INSERT ... SELECT now uses fast creation of index
    
Known bugs:
ma_test_recovery fails because of a bug in redo handling when log is cut directly after a redo (Guilhem knows how to fix)
ma_test_recovery.excepted is not totally correct, because of the above bug
mysqld sometimes fails to restart; Fails with error "end_of_redo_phase: Assertion `long_trid != 0' failed"; Guilhem to investigate


include/maria.h:
  Prototype changes
  Added current_filepos to st_maria_sort_info
mysql-test/r/maria.result:
  Updated results that changes as alter table and insert ... select now uses fast creation of index
mysys/mf_iocache.c:
  Reset variable to gurard against double invocation
storage/maria/ma_bitmap.c:
  Added _ma_bitmap_reset_cache() (needed for repair)
storage/maria/ma_blockrec.c:
  Simplify code
  More initial allocations
  Fixed bug where head block was split before min_row_length (caused Maria to believe row was crashed on read)
storage/maria/ma_blockrec.h:
  Moved TRANSID_SIZE to maria_def.h
  Added prototype for new functions
storage/maria/ma_check.c:
  Simplicy code
  Fixed repair_by_sort to work with BLOCK_RECORD
  - When using BLOCK_RECORD or UNPACK create new Maria handle
  - Use common initializer function
  - Align code with maria_repair()
  
  Made some changes to maria_repair_parallel() to use common initializer function
  Removed ASK_MONTY section by fixing noted problem
storage/maria/ma_close.c:
  Moved check for readonly to _ma_state_info_write()
storage/maria/ma_key_recover.c:
  Use different log entries if key root changes or not.
  This fixed some bugs when tree grows
storage/maria/ma_key_recover.h:
  Added keynr to st_msg_to_write_hook_for_undo_key
storage/maria/ma_loghandler.c:
  Added INIT_LOGREC_UNDO_KEY_INSERT_WITH_ROOT
storage/maria/ma_loghandler.h:
  Added INIT_LOGREC_UNDO_KEY_INSERT_WITH_ROOT
storage/maria/ma_open.c:
  Added TRANSID to all key pages (for future compressing of trans id's)
  For compressed records, alloc a bit bigger buffer to avoid valgrind warnings
  If table is opened readonly, don't update state
storage/maria/ma_packrec.c:
  Allocate bigger array for bit unpacking to avoid valgrind errors
storage/maria/ma_recovery.c:
  Added UNDO_KEY_INSERT_WITH_ROOT & UNDO_KEY_DELETE_WITH_ROOT
storage/maria/ma_sort.c:
  More logging
storage/maria/ma_test_all.sh:
  More tests
storage/maria/ma_test_recovery.expected:
  Update results
  Note that this is not complete becasue of a bug in recovery
storage/maria/ma_test_recovery:
  Removed recreation of index (not needed when we have redo for index pages)
storage/maria/maria_chk.c:
  When using flag --read-only, don't update status for files
  When using --unpack, don't use REPAIR_BY_SORT if other repair option is given
  Enable repair_by_sort for BLOCK records
  Removed not needed newline at start of --describe
storage/maria/maria_def.h:
  Support for TRANSID_SIZE to key pages
storage/maria/maria_read_log.c:
  renamed --only-display to --display-only

Changed bitmaps to be written before unpinning of pages in write_block_record()
Log handler now assumes we never call it for not transactional tables
Fixed bug in ma_test_all that caused it to fail


sql/unireg.cc:
  Removed 'at', as this makes it hard to find valgrind errors from the debug log
storage/maria/ma_blockrec.c:
  Changed bzero() of blocks to get rid of a (non critical) valgrind error
  Changed bitmaps to be written before unpinning of pages in write_block_record()
  fixed that we don't log tails if table isn't transactional
storage/maria/ma_key_recover.c:
  Fixed wrong log_data[] that caused us to log uninitialized data
storage/maria/ma_loghandler.c:
  Replaced not needed test with DBUG_ASSERT()
storage/maria/ma_test_all.sh:
  Remove control file if block size changes

into  mysql.com:/home/my/mysql-maria


include/my_sys.h:
  Auto merged
sql/mysqld.cc:
  Auto merged
storage/maria/ma_checkpoint.c:
  Auto merged
storage/maria/ma_pagecache.c:
  Auto merged
storage/maria/ma_pagecache.h:
  Auto merged
storage/maria/maria_chk.c:
  Auto merged
storage/maria/ma_recovery.c:
  SCCS merged

New extendable format for maria_log_control file
Fixed some compiler warnings


include/maria.h:
  Added maria_disable_logging() and maria_enable_logging()
mysql-test/include/maria_verify_recovery.inc:
  Updated tests now when key redo/undo works
mysql-test/r/maria-recovery.result:
  Updated tests now when key redo/undo works
storage/maria/ma_blockrec.c:
  Use unified CLR code
  Added rec_lsn for full pages
  Moved clr write hook to ma_key_recover.c
  Changed REDO code to keep pages pinned until undo
  Mark page_link's as changed
storage/maria/ma_blockrec.h:
  Moved write_hook_for_clr_end() to ma_key_recover.c
storage/maria/ma_check.c:
  Changed key check code to use PAGECACHE_READ_UNKNOWN_PAGE
  Fixed wrong warning when checking files after maria_pack
  When unpacking files, we have to use new keypos_to_recpos method
  When doing repair, we can disregard index key file pages in page cache
storage/maria/ma_commit.c:
  Added simple enable/disable logging functions
  (Needed for recovery)
storage/maria/ma_control_file.c:
  Make maria control file extendable without having to make it incompatible for older versions
storage/maria/ma_control_file.h:
  New error messages
  Added CONTROL_FILE_VERSION
storage/maria/ma_delete.c:
  Added redo/undo for key pages
  change_length -> changed_length to make things similar
  More comments & more DBUG
storage/maria/ma_key_recover.c:
  Unified CLR method
  Moved here write_hook_for_clr_end() and common keypage log functions
  Changed REDO to keep pages pinned until undo
  Changed UNDO code to change key_root under log mutex
storage/maria/ma_key_recover.h:
  New structures and functions
storage/maria/ma_loghandler.c:
  Include needed files
storage/maria/ma_open.c:
  Change maria_open() to use pread() instead of read()
storage/maria/ma_page.c:
  Fixed bug in key_del handling
  Clear pages if IDENTICAL_PAGES_AFTER_RECOVERY is defined
storage/maria/ma_pagecache.c:
  Indentation and spelling fixes
  More DBUG
  Added helper function: pagecache_block_link_to_buffer()
storage/maria/ma_pagecache.h:
  Added pagecache_block_link_to_buffer()
storage/maria/ma_recovery.c:
  Fixed state.changed
  Fixed that REDO keeps pages pinned until UNDO
  Some bug fixes from previous commit
  Fixes for UNDO/REDO of key pages
storage/maria/ma_search.c:
  Fixed packing and storing of keys to provide more information to caller so
  that we can do efficent REDO logging of the changes.
storage/maria/ma_test1.c:
  Fixed bug with not initialized variable
storage/maria/ma_test2.c:
  Removed not used code
storage/maria/ma_test_all.res:
  Updated results
storage/maria/ma_test_all.sh:
  Changed one test to test more
  Removed timing tests as not relevant here
storage/maria/ma_test_recovery.expected:
  Updated test result after redo/undo if key pages works
storage/maria/ma_test_recovery:
  Updated test after redo/undo if key pages works
storage/maria/ma_write.c:
  Moved some general log functions to ma_key_recover.c
  Fixed some bugs in undo
  Moved ma_log_split() to _ma_split_page()
  Small changes in some function arguments to be able to do redo logging
storage/maria/maria_chk.c:
  disable logging while doing repair table
storage/maria/maria_def.h:
  New function prototypes
  Move some structs and functions to ma_key_recover.c
storage/maria/unittest/ma_control_file-t.c:
  Updated with patch from Sanja
  NOTE: This is not complete and need to be updated to new control file format
storage/maria/unittest/ma_test_loghandler-t.c:
  Fixed compiler warning

instead of fprintf(stderr) when a task (with no user connected) gets
an error, use my_printf_error(). Flags ME_JUST_WARNING and ME_JUST_INFO
added to my_error()/my_printf_error(), which pass it to
my_message_sql() which is modified to call the appropriate
sql_print_*(). This way recovery can signal its start and end with
[Note] and not [ERROR] (but failure with [ERROR]).
Recovery's detailed progress (percents etc) still uses stderr as they
have to stay on one single line.
sql_print_error() changed to use my_progname_short (nicer display).
mysql-test-run.pl --gdb/--ddd does not run mysqld, because
a breakpoint in mysql_parse is too late to debug startup problems;
instead, dev should set the breakpoints it wants and then "run" ("r").


include/my_sys.h:
  new flags to tell error_handler_hook that this is not an error
  but an information or warning
mysql-test/mysql-test-run.pl:
  when running with --gdb/--ddd to debug mysqld, breaking at mysql_parse
  is too late to debug startup problems; now, it does not run mysqld,
  does not set breakpoints, developer can set as early breakpoints
  as it wants and is responsible for typing "run" (or "r")
mysys/my_init.c:
  set my_progname_short
mysys/my_static.c:
  my_progname_short added
sql/mysqld.cc:
  * my_message_sql() can now receive info or warning, not only error;
  this allows mysys to tell the user (or the error log if no user)
  about an info or warning. Used from Maria.
  * plugins (or engines like Maria) may want to call my_error(), so
  set up the error handler hook (my_message_sql) before initializing
  plugins; otherwise they get my_message_no_curses which is less
  integrated into mysqld (is just fputs())
  * using my_progname_short instead of my_progname, in my_message_sql()
  (less space on screen)
storage/maria/ma_checkpoint.c:
  fprintf(stderr) -> ma_message_no_user()
storage/maria/ma_checkpoint.h:
  function for any Maria task, not connected to a user (example:
  checkpoint, recovery; soon could be deleted records purger)
  to report a message (calls my_printf_error() which, when inside ha_maria,
  leads to sql_print_*(), and when outside, leads to
  my_message_no_curses i.e. stderr).
storage/maria/ma_recovery.c:
  To tell that recovery starts and ends we use ma_message_no_user()
  (sql_print_*() in practice). Detailed progress info still uses
  stderr as sql_print() cannot put several messages on one line.
  071116 18:42:16 [Note] mysqld: Maria engine: starting recovery
  recovered pages: 0% 67% 100% (0.0 seconds); transactions to roll back: 1 0 (0.0
  seconds); tables to flush: 1 0 (0.0 seconds);
  071116 18:42:16 [Note] mysqld: Maria engine: recovery done
storage/maria/maria_chk.c:
  my_progname_short moved to mysys
storage/maria/maria_read_log.c:
  my_progname_short moved to mysys
storage/myisam/myisamchk.c:
  my_progname_short moved to mysys

background page flushing was using dfile even when it wanted to flush
the index file.


storage/maria/ma_checkpoint.c:
  * filter_flush_data_file* functions are in fact for the index file too,
  renaming them.
  * flush of index file was using dfile (bad copy-paste)

storage/maria/ma_checkpoint.c:
  React only on errors during the flush.

into  mysql.com:/home/my/mysql-maria


storage/maria/ha_maria.cc:
  Auto merged
storage/maria/ma_bitmap.c:
  Auto merged
storage/maria/ma_checkpoint.c:
  Auto merged
storage/maria/ma_close.c:
  Auto merged
storage/maria/ma_loghandler.c:
  Auto merged
storage/maria/ma_loghandler.h:
  Auto merged
storage/maria/ma_open.c:
  Auto merged
storage/maria/ma_pagecache.h:
  Auto merged
storage/maria/ma_write.c:
  Auto merged
storage/maria/maria_def.h:
  Auto merged
storage/maria/unittest/ma_pagecache_single.c:
  Auto merged
storage/maria/ma_blockrec.c:
  Manual merge
storage/maria/ma_page.c:
  Manual merge
storage/maria/ma_pagecache.c:
  Manual merge
storage/maria/ma_preload.c:
  Manual merge
storage/maria/ma_recovery.c:
  Manual merge
  Add _ma_unpin_all_pages() to all new UNDO redo_exec_hook's

Added key_nr to st_maria_keydef for faster keyinfo->keynr conversion
For transactional tables, shift record number in keys up with 1 bit to have place to indicate if transid follows
Checksum for MyISAM now ignores NULL and not used part of VARCHAR
Renamed some variables that caused shadow compiler warnings
Moved extra() call when waiting for tables to not be used to after tables are removed from cache.
Fixed crashing bugs when using Maria TEMPORARY tables with TRUNCATE. Removed 'hack' code in sql directory to go around this bug.
pagecache_unlock_by_ulink() now has extra argument to say if page was changed.
Give error message if we fail to open control file
Mark page cache variables as not flushable


include/maria.h:
  Made min page cache larger (needed for pinning key page)
  Added key_nr to st_maria_keydef for faster keyinfo->keynr conversion
  Added write_comp_flag to move some runtime code to maria_open()
include/my_base.h:
  Added new error message to be used when handler initialization failed
include/my_global.h:
  Renamed dummy to swap_dummy to avoid conflicts with local 'dummy' variables
include/my_handler.h:
  Added const to some parameters
mysys/array.c:
  More DBUG
mysys/my_error.c:
  Fixed indentation
mysys/my_handler.c:
  Added const to some parameters
  Added missing error messages
sql/field.h:
  Renamed variables to avoid variable shadowing
sql/handler.h:
  Renamed parameter to avoid variable name conflict
sql/item.h:
  Renamed variables to avoid variable shadowing
sql/log_event_old.h:
  Renamed variables to avoid variable shadowing
sql/set_var.h:
  Renamed variables to avoid variable shadowing
sql/sql_delete.cc:
  Removed maria hack for temporary tables
  Fixed indentation
sql/sql_table.cc:
  Moved extra() call when waiting for tables to not be used to after tables are removed from cache.
  This was needed to ensure we don't do a PREPARE_FOR_DROP or similar call while the table is still in use.
sql/table.cc:
  Copy page_checksum from share
  Removed Maria hack
storage/maria/Makefile.am:
  Added new files
storage/maria/ha_maria.cc:
  Renamed records -> record_count and info -> create_info to avoid variable name conflicts
  Mark page cache variables as not flushable
storage/maria/ma_blockrec.c:
  Moved _ma_unpin_all_pages() to ma_key_recover.c
  Moved init of info->pinned_pages to ma_open.c
  Moved _ma_finalize_row() to maria_key_recover.h
  Renamed some variables to avoid variable name conflicts
  Mark page_link.changed for blocks we change directly
  Simplify handling of undo link when writing LOGREC_UNDO_ROW_INSERT (old code crashed when having redo for index)
storage/maria/ma_blockrec.h:
  Removed extra empty line
storage/maria/ma_checkpoint.c:
  Remove not needed trnman.h
storage/maria/ma_close.c:
  Free pinned pages (which are now always allocated)
storage/maria/ma_control_file.c:
  Give error message if we fail to open control file
storage/maria/ma_delete.c:
  Changes for redo logging (first part, logging of underflow not yet done)
  - Log undo-key-delete
  - Log delete of key
  - Updated arguments to _ma_fetch_keypage(), _ma_dispose(), _ma_write_keypage(), _ma_insert()
  - Added new arguments to some functions to be able to write redo information
  - Mark key pages as changed when we write with PAGECACHE_LOCK_LEFT_WRITELOCKED
  
  Remove one not needed _ma_write_keypage() in d_search() when upper level will do the write anyway
  Changed 2 bmove_upp() to bmove() as this made code easer to understand
  More function comments
  Indentation fixes
storage/maria/ma_ft_update.c:
  New arguments to _ma_write_keypage()
storage/maria/ma_loghandler.c:
  Fixed some DBUG_PRINT messages
  Simplify code
  Added new log entrys for key page redo
  Renamed some variables to avoid variable name shadowing
storage/maria/ma_loghandler.h:
  Moved some defines here
  Added define for storing key number on key pages
  Added new translog record types
  Added enum for type of operations in LOGREC_REDO_INDEX
storage/maria/ma_open.c:
  Always allocate info.pinned_pages (we need now also for normal key page usage)
  Update keyinfo->key_nr
  Added virtual functions to convert record position o number to be stored on key pages
  Update keyinfo->write_comp_flag to value of search flag to be used when writing key
storage/maria/ma_page.c:
  Added redo for key pages
  - Extended _ma_fetch_keypage() with type of lock to put on page and address to used MARIA_PINNED_PAGE
  - _ma_fetch_keypage() now pin's pages if needed
  - Extended _ma_write_keypage() with type of locks to be used
  - ma_dispose() now locks info->s->state.key_del from other threads
  - ma_dispose() writes redo log record
  - ma_new() locks info->s->state.key_del from other threads if it was used
  - ma_new() now pins read page
  
  Other things:
  - Removed some not needed arguments from _ma_new() and _ma_dispose)
  - Added some new variables to simplify code
  - If EXTRA_DEBUG is used, do crc on full page to catch not unitialized bytes
storage/maria/ma_pagecache.h:
  Applied patch from Sanja to add extra argument to pagecache_unlock_by_ulink() to mark if page was changed
  Added some defines for pagecache priority levels that one can use
storage/maria/ma_range.c:
  Added new arguments for call to _ma_fetch_keypage()
storage/maria/ma_recovery.c:
  - Added hooks for new translog types:
    REDO_INDEX, REDO_INDEX_NEW_PAGE, REDO_INDEX_FREE_PAGE, UNDO_KEY_INSERT, UNDO_KEY_DELETE and
    UNDO_KEY_DELETE_WITH_ROOT.
  - Moved variable declarations to start of function (portability fixes)
  - Removed some not needed initializations
  - Set only relevant state changes for each redo/undo entry
storage/maria/lockman.c:
  Removed end space
storage/maria/ma_check.c:
  Removed end space
storage/maria/ma_create.c:
  Removed end space
storage/maria/ma_locking.c:
  Removed end space
storage/maria/ma_packrec.c:
  Removed end space
storage/maria/ma_pagecache.c:
  Removed end space
storage/maria/ma_panic.c:
  Removed end space
storage/maria/ma_rt_index.c:
  Added new arguments for call to _ma_fetch_keypage(), _ma_write_keypage(), _ma_dispose() and _ma_new()
  Fixed indentation
storage/maria/ma_rt_key.c:
  Added new arguments for call to _ma_fetch_keypage()
storage/maria/ma_rt_split.c:
  Added new arguments for call to _ma_new()
  Use new keypage header
  Added new arguments for call to _ma_write_keypage()
storage/maria/ma_search.c:
  Updated comments & indentation
  Added new arguments for call to _ma_fetch_keypage()
  Made some variables and arguments const
  Added virtual functions for converting row position to number to be stored in key
  use MARIA_RECORD_POS of record position instead of my_off_t
  Record in MARIA_KEY_PARAM how page was changed one key insert (needed for REDO)
storage/maria/ma_sort.c:
  Removed end space
storage/maria/ma_statrec.c:
  Updated arguments for call to _ma_rec_pos()
storage/maria/ma_test1.c:
  Fixed too small buffer to init_pagecache()
  Fixed bug when using insert_count and test_flag
storage/maria/ma_test2.c:
  Use more resonable pagecache size
  Remove not used code
  Reset blob_length to fix wrong output message
storage/maria/ma_test_all.sh:
  Fixed wrong test
storage/maria/ma_write.c:
  Lots of new code to handle REDO of key pages
  No logic changes because of REDO code, mostly adding new arguments and adding new code for logging 
  
  Added new arguments for calls to _ma_fetch_keypage(), _ma_write_keypage() and similar functions
  Move setting of comp_flag in ma_ck_wrte_btree() from runtime to maria_open()
  Zerofill new used pages for:
  - To remove possible sensitive data left in buffer
  - To get idenitical data on pages after running redo
  - Better compression of pages if archived
storage/maria/maria_chk.c:
  Added information if table is crash safe
storage/maria/maria_def.h:
  New virtual function to convert between record position on key and normal record position
  Aded mutex and extra variables to handle locking of share->state.key_del
  Moved some structure variables to get things more aligned
  Added extra arguments to MARIA_KEY_PARAM to be able to remember what was changed on key page on key insert
  Added argument to MARIA_PINNED_PAGE to indicate if page was changed
  Updated prototypes for functions
  Added some structures for signaling changes in REDO handling
storage/maria/unittest/ma_pagecache_single.c:
  Updated arguments for changed function calls
storage/myisam/mi_check.c:
  Made calc_check_checksum virtual
storage/myisam/mi_checksum.c:
  Update checksums to ignore null columns
storage/myisam/mi_create.c:
  Mark if table has null column (to know when we have to use mi_checksum())
storage/myisam/mi_open.c:
  Added virtual function for calculating checksum to be able to easily ignore NULL fields
storage/myisam/mi_test2.c:
  Fixed bug
storage/myisam/myisamdef.h:
  Added virtual function for calculating checksum during check table
  Removed ha_key_cmp() as this is in handler.h
storage/maria/ma_key_recover.c:
  New BitKeeper file ``storage/maria/ma_key_recover.c''
storage/maria/ma_key_recover.h:
  New BitKeeper file ``storage/maria/ma_key_recover.h''
storage/maria/ma_key_redo.c:
  New BitKeeper file ``storage/maria/ma_key_redo.c''

maria_read_log used to always print a warning message at startup
to say it is unsafe if ALTER TABLE was used. Now it prints it only
if the log does show the problem (=ALTER TABLE or CREATE SELECT, which
both disable logging of REDO_INSERT*).
For that, when ha_maria::external_lock() disables transactionality
it writes a LOGREC_INCOMPLETE_LOG to the log, which "maria_read_log -a"
picks up to write a warning.
REPAIR TABLE also disables those REDO_INSERT* but as maria_read_log
executes LOGREC_REDO_REPAIR no warning is needed.


storage/maria/ha_maria.cc:
  as we now log a record when disabling transactionility, we need the
  TRN to be set up first
storage/maria/ma_blockrec.c:
  comment
storage/maria/ma_loghandler.c:
  new type of log record
storage/maria/ma_loghandler.h:
  new type of log record
storage/maria/ma_recovery.c:
  * maria_apply_log() now returns a count of warnings. What currently
  produces warnings is:
  - skipping applying UNDOs though there are some (=> inconsistent table)
  - replaying log (in maria_read_log) though the log contains some
  ALTER TABLE or CREATE SELECT (log misses REDO_INSERT* for those
  and is so incomplete).
  Count of warnings affects the final message of maria_read_log and
  recovery (though in recovery none of the two conditions above should
  happen).
  * maria_read_log used to always print a warning message at startup
  to say it is unsafe if ALTER TABLE was used. Now it prints it only
  if the log does show the problem, i.e. ALTER TABLE or CREATE SELECT
  was used (both disable logging of REDO_INSERT* as those records are
  not needed for recovery; those missing records in turn make
  recreation-from-scratch, via maria_read_log, impossible). For that,
  when ha_maria::external_lock() disables transactionality,
  _ma_tmp_disable_logging_for_table() writes a LOGREC_INCOMPLETE_LOG to
  the log, which maria_apply_log() picks up to write a warning.
storage/maria/ma_recovery.h:
  maria_apply_log() returns a count of warnings
storage/maria/maria_def.h:
  _ma_tmp_disable_logging_for_table() grows so becomes a function
storage/maria/maria_read_log.c:
  maria_apply_log can now return a count of warnings, to temper the
  "SUCCESS" message printed in the end by maria_read_log.
  Advise users to make a backup first.

possibility, if requested in the test, to do byte-for-byte comparison
of data files after recovery. This makes sense only when testing
REDOs, as UNDO phase generates new log records thus data pages get
new LSNs. First test of maria-recovery.test is changed to test only
REDOs and do byte-for-byte comparison.


mysql-test/include/maria_make_snapshot.inc:
  possibility to compare physically the data files (makes sense
  only if testing REDO phase).
mysql-test/include/maria_verify_recovery.inc:
  allow physical comparison of data files after recovery if requested
mysql-test/r/maria-recovery.result:
  result update (first test tests REDO only now)
mysql-test/t/maria-recovery.test:
  first test tests REDO only now, and thus can do byte-for-byte
  comparison of data files.

See test maria-recovery.test for a model; all include scripts have
an "API" section at start if they do take parameters from outside.
* Fixing bug reported by Jani and Monty (when two REDOs about the same
page in one group, see ma_blockrec.c). 
* Fixing small bugs in recovery


mysql-test/include/wait_until_connected_again.inc:
  be sure to enter the loop (the previous query by the caller may not have
  failed: it could be
  query;
  mysqladmin shutdown;
  call this script).
mysql-test/lib/mtr_process.pl:
  * Through the "expect" file a test can tell mtr that a server crash
  is expected. What the file contains is irrelevant. Now if its last
  line starts with "wait", mtr will wait before restarting (it will
  wait for the last line to not start with "wait"). This is for
  tests which need to mangle files under the feet of a dead mysqld.
  * Remove "expect" file before restarting; otherwise there could be a
  race condition: tests sees server restarted, does something, writes
  an "expect" file, and then mtr removes that file, then
  test kills mysqld, and then mtr will never restart it.
storage/maria/ma_blockrec.c:
  - when applying a REDO in recovery, we don't anymore put UNDO's LSN on the page
  at once; indeed if in this REDO's group there comes another REDO
  for the same page it would be wrongly skipped. Instead, we keep
  pages pinned, don't change their LSN. When done with all REDOs
  of the group we unpin them and stamp them with UNDO's LSN.
  - fixing bug in applying of REDO_PURGE_BLOCKS in recovery: page_range
  sometimes has TAIL_BIT set, need to turn it down to know the real page
  range.
  - Both bugs are covered in maria-recovery.test
storage/maria/ma_checkpoint.c:
  Capability to, in debug builds only, do some special operations
  (flush all bitmap and data pages, flush state, flush log)
  and crash mysqld, to later test recovery.
   Driven by some --debug=d, symbols.
storage/maria/ma_open.c:
  debugging info
storage/maria/ma_pagecache.c:
  Now that we can _ma_unpin_all_pages() during the REDO phase
  to set page's LSN, the assertion needs to be relaxed.
storage/maria/ma_recovery.c:
  - open trace file in append mode (useful when a test triggers several
  recoveries, we see them all).
  - fixing wrong error detection, it's possible that during recovery
  we want to open an already open table.
  - when applying a REDO in recovery, we don't anymore put UNDO's LSN on the page
  at once; indeed if in this REDO's group there comes another REDO
  for the same page it would be wrongly skipped. Instead, we keep
  pages pinned, don't change their LSN. When done with all REDOs
  of the group we unpin them and stamp them with UNDO's LSN.
  - we verify that all log records of a group are about the same table,
  for debugging.
mysql-test/r/maria-recovery.result:
  result
mysql-test/t/maria-recovery-master.opt:
  crash is expected, core file would take room, stack trace would
  wake pushbuild up.
mysql-test/t/maria-recovery.test:
  Test of recovery from mysql-test (it is already tested as unit tests
  in ma_test_recovery) (WL#4137)
  - test that, if recovery is made to start on an empty table it can
  replay the effects of committed and uncommitted statements (having only
  the committed ones in the end result). This should be the first test
  for someone writing code of new REDOs.
  - test that, if mysqld is crashed and recovery runs we have only
  committed statements in the end result. Crashes are done in different
  ways: flush nothing (so, uncommitted statement is often missing
  from the log => no rollback to do); flush pagecache (implicitely flushes
  log (WAL)) and flush log, both causes rollbacks; flush log can also
  flush state (state.records etc) to test recovery of the state
  (not tested well now as we repair the index anyway).
  - test of bug found by Jani and Monty in recovery (two REDO about
  the same page in one group).
mysql-test/include/maria_empty_logs.inc:
  removes logs, to have a clean sheet for testing recovery.
mysql-test/include/maria_make_snapshot.inc:
  copies a table to another directory, or back, or compares both
  (comparison is not implemented as physical comparison is impossible
  if an UNDO phase happened).
mysql-test/include/maria_make_snapshot_for_comparison.inc:
  copies tables to another directory so that they can later
  serve as a comparison reference (they are the good tables,
  recovery should produce similar ones).
mysql-test/include/maria_make_snapshot_for_feeding_recovery.inc:
  When we want to force recovery to start on old tables, we prepare
  old tables with this script: we put them in a spare directory.
  They are later copied back over mysqltest tables while mysqld is dead.
  We also need to copy back the control file, otherwise mysqld,
  in recovery, would start from the latest checkpoint: latest
  checkpoint plus old tables is not a recovery-possible scenario of course.
mysql-test/include/maria_verify_recovery.inc:
  causes mysqld to crash, restores old tables if requested,
  lets recovery run, compares resulting tables with reference tables
  by using CHECKSUM TABLE.
  We don't do any sanity checks on page's LSN in resulting tables, yet.

Fixing bad comments (I remember my maths' teacher "one late night you'll
obey to the simplifications made by your tired neurons"; exactly
what happened here). In Checkpoint, when we flush a table's state
we must flush all log records (WAL), not only those before checkpoint
started.


storage/maria/ma_bitmap.c:
  there was a flaw in reasoning, bug does exist.
storage/maria/ma_blockrec.c:
  moving piece of comment to ma_checkpoint.c
storage/maria/ma_checkpoint.c:
  Comments.
  When checkpoint flushes a state, WAL imposes that all records up
  to this state have been flushed, not only up to checkpoint_start_log_horizon.
storage/maria/ma_recovery.c:
  finishing comment.

Observe WAL for the table's state: all log records needed for undoing
uncommitted state must be in the log before we flush state.


storage/maria/ha_maria.cc:
  comments
storage/maria/ma_bitmap.c:
  Comment for why there is no bug
storage/maria/ma_blockrec.c:
  comment for why there is no bug
storage/maria/ma_checkpoint.c:
  Observe WAL for the table's state: all log records needed for undoing
  uncommitted state must be in the log before we flush state. I tested
  by hand that the bug existed (create table, insert one row into it
  but let that insert pause after increasing data_file_length, let
  checkpoint start but kill it after it has flushed state).
  Log contains nothing, table is not recovered though it has
  a too big data_file_length. With this bugfix, the log contains
  REDO so table is opened so data_file_length is corrected.
storage/maria/ma_close.c:
  If table is read-only we must never write to it.
  Should be a no-change in fact, as if read-only, share->changed is
  normally always false.
storage/maria/ma_recovery.c:
  documenting bug found by Monty. Print when fixing data_file_length.

storage/maria/ma_loghandler.c:
  Infinite loop in case of given address more then last LSN fixed
  (now it means just flush all log).
  Removed unneeded ASSERT.
storage/maria/unittest/ma_test_loghandler-t.c:
  The test case of flushing all log added.

storage/maria/ma_pagecache.c:
  Support of rec_lsn added to pagecache_write call.
  A function for setting rec_lsn made.

into  mysql.com:/home/my/mysql-maria


include/my_sys.h:
  Auto merged
mysql-test/r/maria.result:
  Auto merged
mysql-test/t/maria.test:
  Auto merged
sql/handler.h:
  Auto merged
sql/mysqld.cc:
  Auto merged
storage/maria/ha_maria.cc:
  Auto merged
storage/maria/ma_bitmap.c:
  Auto merged
storage/maria/ma_blockrec.c:
  Auto merged
storage/maria/ma_loghandler.c:
  Auto merged
storage/maria/ma_pagecache.c:
  Auto merged
storage/maria/ma_test1.c:
  Auto merged
storage/maria/ma_test_recovery.expected:
  Auto merged
storage/maria/ma_test_recovery:
  Auto merged
sql/mysql_priv.h:
  manual merge
storage/maria/ma_recovery.c:
  manual merge
storage/maria/ma_test2.c:
  manual merge

Changed format for REDO_INSERT_ROWS_BLOBS
Fixed several bugs in handling of big blobs
Added redo_free_head_or_tail() & redo_insert_row_blobs()
Added uuid to control file
maria_checks now verifies that not used part of bitmap is 0
REDO_PURGE_BLOCKS -> REDO_FREE_BLOCKS
Added REDO_FREE_HEAD_OR_TAIL
Fixes problem when trying to read block outside of file during REDO


include/my_global.h:
  STACK_DIRECTION is already set by configure
mysql-test/r/maria.result:
  Updated results
mysql-test/t/maria.test:
  Test shrinking of VARCHAR
mysys/my_realloc.c:
  Fixed indentation
mysys/safemalloc.c:
  Fixed indentation
sql/filesort.cc:
  Removed some casts
sql/mysqld.cc:
  Added missing setting of myisam_stats_method_str
sql/uniques.cc:
  Removed some casts
storage/maria/ma_bitmap.c:
  Added printing of bitmap (for debugging)
  Renamed _ma_print_bitmap() -> _ma_print_bitmap_changes()
  Added _ma_set_full_page_bits()
  Fixed bug in ma_bitmap_find_new_place() (affecting updates) when using big files
storage/maria/ma_blockrec.c:
  Changed format for REDO_INSERT_ROWS_BLOBS
  Fixed several bugs in handling of big blobs
  Added code to fix some cases where redo when using blobs didn't produce idenital .MAD files as normal usage
  REDO_FREE_ROW_BLOCKS doesn't anymore change pages; We only mark things free in bitmap
  Remove TAIL and filler extents from REDO_FREE_BLOCKS log entry. (Fixed some asserts)
  REDO_PURGE_BLOCKS -> REDO_FREE_BLOCKS
  Delete tails in update. (Fixed bug when doing update that shrinks blob/varchar length)
  Fixed bug when doing insert in block outside of file size.
  Added redo_free_head_or_tail() & redo_insert_row_blobs()
  Added pagecache_unlock_by_link() when read fails.
  Much more comments, DBUG and ASSERT entries
storage/maria/ma_blockrec.h:
  Prototypes of new functions
  Define of SUB_RANGE_SIZE & BLOCK_FILLER_SIZE
storage/maria/ma_check.c:
  Verify that not used part of bitmap is 0
storage/maria/ma_control_file.c:
  Added uuid to control file
storage/maria/ma_loghandler.c:
  REDO_PURGE_BLOCKS -> REDO_FREE_BLOCKS
  Added REDO_FREE_HEAD_OR_TAIL
storage/maria/ma_loghandler.h:
  REDO_PURGE_BLOCKS -> REDO_FREE_BLOCKS
  Added REDO_FREE_HEAD_OR_TAIL
storage/maria/ma_pagecache.c:
  If we write full block, remove error flag for block.
  (Fixes problem when trying to read block outside of file)
storage/maria/ma_recovery.c:
  REDO_PURGE_BLOCKS -> REDO_FREE_BLOCKS
  Added REDO_FREE_HEAD_OR_TAIL
storage/maria/ma_test1.c:
  Allow option after 'b' to be compatible with ma_test2
  (This is just to simplify test scripts like ma_test_recovery)
storage/maria/ma_test2.c:
  Default size of blob is now 1000 instead of 1
storage/maria/ma_test_all.sh:
  Added test for bigger blobs
storage/maria/ma_test_recovery.expected:
  Updated results
storage/maria/ma_test_recovery:
  Added test for bigger blobs

don't take a checkpoint at end of REDO phase because blocks' rec_lsn
are not set. Fixes some assertion failures during tests of killing
the UNDO phase (recovery then started from the end-of-REDO-phase
checkpoint, wrongly ignoring REDOs before that record).


storage/maria/ma_blockrec.c:
  note for a bug (mail sent to colleagues, put on todo)
storage/maria/ma_recovery.c:
  this checkpoint at end of REDO phase is incorrect, disabling it.
  The one at end of UNDO phase remains.

into  gbichot4.local:/home/mysql_src/mysql-maria-for-undo-phase

- serializing calls to flush_pagecache_blocks_int() on the same file
to avoid known concurrency bugs
- having that, we can now enable the background thread, as the
flushes it does are now supposedly safe in concurrent situations.
- new type of flush FLUSH_KEEP_LAZY: when the background checkpoint
thread is flushing a packet of dirty pages between two checkpoints,
it uses this flush type, indeed if a file is already being flushed
by another thread it's smarter to move on to the next file than wait.
- maria_checkpoint_frequency renamed to maria_checkpoint_interval.


include/my_sys.h:
  new type of flushing for the page cache: FLUSH_KEEP_LAZY
mysql-test/r/maria.result:
  result update
mysys/mf_keycache.c:
  indentation. No FLUSH_KEEP_LAZY support in key cache.
storage/maria/ha_maria.cc:
  maria_checkpoint_frequency was somehow a hidden part of the
  Checkpoint API and that was not good. Now we have checkpoint_interval,
  local to ha_maria.cc, which serves as container for the user-visible
  maria_checkpoint_interval global variable; setting it calls
  update_checkpoint_interval which passes the new value to
  ma_checkpoint_init(). There is no hiding anymore.
  By default, enable background thread which does checkpoints
  every 30 seconds, and dirty page flush in between. That thread takes
  a checkpoint when it ends, so no need for maria_hton_panic to take one.
  The | is | and not ||, because maria_panic() must always be called.
  frequency->interval.
storage/maria/ma_checkpoint.c:
  Use FLUSH_KEEP_LAZY for background thread when it flushes packets of
  dirty pages between two checkpoints: it is smarter to move on to
  the next file than wait for it to have been completely flushed, which
  may take long.
  Comments about flush concurrency bugs moved from ma_pagecache.c.
  Removing out-of-date comment.
  frequency->interval.
  create_background_thread -> (interval>0).
  In ma_checkpoint_background(), some variables need to be preserved
  between iterations.
storage/maria/ma_checkpoint.h:
  new prototype
storage/maria/ma_pagecache.c:
  - concurrent calls of flush_pagecache_blocks_int() on the same file
  cause bugs (see @note in that function); we fix them by serializing
  in this situation. For that we use a global hash of (file, wqueue).
  When flush_pagecache_blocks_int() starts it looks into the hash,
  using the file as key. If not found, it inserts (file,wqueue) into the
  hash, flushes the file, and finally removes itself from the hash and
  wakes up any waiter in the queue. If found, it adds itself to the
  wqueue and waits.
  - As a by-product, we can remove changed_blocks_is_incomplete
  and replace it by scanning the hash, replace the sleep() by a queue wait.
  - new type of flush FLUSH_KEEP_LAZY: when flushing a file, if it's
  already being flushed by another thread (even partially), return
  immediately.
storage/maria/ma_pagecache.h:
  In pagecache, a hash of files currently being flushed (i.e. there
  is a call to flush_pagecache_blocks_int() for them).
storage/maria/ma_recovery.c:
  new prototype
storage/maria/ma_test1.c:
  new prototype
storage/maria/ma_test2.c:
  new prototype

    always process uint/ulong using ulonglong (unsigned) code
    dbug printout for adjusted option values
strings/llstr.c
    ullstr() - the unsigned brother of llstr()


include/m_string.h:
  ullstr() - the unsigned brother of llstr()
mysql-test/t/variables.test:
  test adjusted for 32bit
mysys/my_getopt.c:
  always process uint/ulong using ulonglong (unsigned) code
  dbug printout for adjusted option values
strings/llstr.c:
  ullstr() - the unsigned brother of llstr()

into  gbichot4.local:/home/mysql_src/mysql-maria-tmp

storage/maria/ma_loghandler.c:
  applying patch for bugfix, by Sanja.

into  gbichot4.local:/home/mysql_src/mysql-maria-for-undo-phase


storage/maria/ma_recovery.c:
  Auto merged

Ability for flush_pagecache_blocks() to flush only certain pages of
a file, as instructed by an option "filter" pointer-to-function argument;
Checkpoint and background dirty page flushing use that to flush only
pages which have been dirty for long enough and bitmap pages.
Fix for a bug in flush_cached_blocks() (no idea if it could produce
a bug in real life, but theoretically it is).
Testing checkpoint in ma_test_recovery via ma_test1 and ma_test2.
Background checkpoint & dirty pages flush thread is still disabled
by default in ha_maria.


mysql-test/r/maria.result:
  result update
storage/maria/ha_maria.cc:
  blank after function comment
storage/maria/ma_checkpoint.c:
  Using an enum instead of 0/1/2 (applying Sanja's review comments).
  The comment about "this is an horizon" can be removed as Sanja
  created translog_next_LSN() which parse_checkpoint_record() uses.
  Variables in ma_checkpoint_background() cannot be declared in the
  for() as their value must not be reset at each iteration!
storage/maria/ma_pagecache.c:
  adding to flush_pagecache_blocks() optional arguments 'filter'
  (pointer to function) and 'filter_arg'; if filter!=NULL this function
  will be called for each block of the file and will reply if this
  block and following ones should be flushed or not (3 possible
  replies).
  Fixing a bug when flush_cached_blocks() skips a pinned page: it has
  to unset PCBLOCK_IN_FLUSH set by flush_pagecache_blocks_int().
storage/maria/ma_pagecache.h:
  flush_pagecache_blocks() is changed to take "filter" and "filter_arg"
  arguments. "filter", if it is not NULL, may return one value
  among enum pagecache_flush_filter_result.
storage/maria/ma_recovery.c:
  open_count=0 when closing tables at the end of recovery.
storage/maria/ma_test1.c:
  Optional checkpoints (-H#) at various stages (stages similar
  to --testflag), for testing of checkpoints.
storage/maria/ma_test2.c:
  Optional checkpoints (-H#) at various stages (stages similar
  to -t), for testing of checkpoints.
storage/maria/ma_test_recovery.expected:
  Result update: the results of the additional test run with -H#
  (checkpoints) are added here. They are exactly identical to without
  checkpoints except that the index's Root (printed by maria_chk)
  is more correct when using checkpoints. This is because checkpoint
  flushed the state, so it happens to be correct, while no-checkpoint
  does not flush the state, and recovery does not recover indexes
  so Root is never fixed. When we recover indices, this will go away.
storage/maria/ma_test_recovery:
  We duplicate the loop of tests to add an additional run with
  checkpoints at various stages, to see if maria_read_log
  uses them fine.

into  desktop.sanja.is.com.ua:/home/bell/mysql/bk/work-maria-bug


storage/maria/ma_recovery.c:
  Auto merged

  on the loghandler start.
Variable definition moved because it is C programm.


storage/maria/ma_loghandler.c:
  Checking that the very last record is fully written
    on the loghandler start.
storage/maria/ma_recovery.c:
  Variable definition moved because it is C programm.

storage/maria/ma_delete_all.c:
  comment (@todo gone, nothing we can do)
storage/maria/ma_recovery.c:
  - fail if LOGREC_CREATE_TABLE contains some symlink info (that
  will be handled in a future version)
  - don't do buffer overrun when parsing a REDO record
  - out-of-date @todo

into  gbichot4.local:/home/mysql_src/mysql-maria-tmp

- fixes (in recovery, checkpoint, log handler) of bugs found
during testing.
- new option --check for maria_read_log: with --only-display (which only
reads the header) it reads the full record, for debugging.


storage/maria/ma_loghandler.c:
  importing patch from Sanja for bug of translog_next_LSN() found
  during recovery
storage/maria/ma_loghandler_lsn.h:
  better types (0L is 4 bytes on some platforms, it causes problems
  when used into lsn_store(): right shift >= width of type.
storage/maria/ma_pagecache.c:
  work around infamous "PAGECACHE_PLAIN_PAGE used for transactional
  tables in specialm case"; REDO phase disables logging and this causes
  pages to be PAGECACHE_PLAIN_PAGE, thus ignored wrongly by the
  checkpoint taken at the end of the REDO phase.
storage/maria/ma_recovery.c:
  - a #ifdef which broke maria_read_log in non-debug builds (no output!)
  - support for maria_read_log --check
  - detect record corruption before opening the table
  - updating is_of_horizon requires writing the state
  - fix for wrong parsing of checkpoint record by recovery
storage/maria/ma_recovery.h:
  support for maria_read_log --check
storage/maria/maria_read_log.c:
  Option --check: --only-display only looks at the header;
  adding --check tries a translog_read_record() to see if record can
  be fully read (this is to find bugs).

  bugfixes


dbug/dbug_add_tags.pl:
  bugfixes

client/mysqltest.c:
  fix my_option's with incorrect defaults
mysql-test/r/maria.result:
  update results
mysql-test/t/variables.test:
  update results
sql/mysqld.cc:
  fix my_option's with incorrect defaults

1. available_stack_size() was getting the direction wrong,
   so alloca() was never used
2. (char*) casts added to kill "break strict-aliasing rules" warnings
3. s/node/node=0/ to kill "pointer casted to integer" warning
4. added volatiles as appropriate to prevent gcc from moving
   assignment out of the loop


mysys/lf_alloc-pin.c:
  bugs/warnings fixed:
  1. available_stack_size() was getting the direction wrong,
     so alloca() was never used
  2. (char*) casts added to kill "break strict-aliasing rules" warnings
  3. s/node/node=0/ to kill "pointer casted to integer" warning
  4. added volatiles as appropriate to prevent gcc from moving
     assignment out of the loop