mariadb/handler/handler0alter.cc

1215 lines
30 KiB
C++
Raw Normal View History

branches/innodb+: Merge revisions 4150:4528 from branches/zip: ------------------------------------------------------------------------ r4152 | marko | 2009-02-10 12:52:27 +0200 (Tue, 10 Feb 2009) | 12 lines branches/zip: When innodb_use_sys_malloc is set, ignore innodb_additional_mem_pool_size, because nothing will be allocated from mem_comm_pool. mem_pool_create(): Remove the assertion about size. The function will work with any size. However, an assertion would fail in ut_malloc_low() when size==0. mem_init(): When srv_use_sys_malloc is set, pass size=1 to mem_pool_create(). mem0mem.c: Add #include "srv0srv.h" that is needed by mem0dbg.c. ------------------------------------------------------------------------ r4153 | vasil | 2009-02-10 22:58:17 +0200 (Tue, 10 Feb 2009) | 14 lines branches/zip: (followup to r4145) Non-functional change: Change the os_atomic_increment() and os_compare_and_swap() functions to macros to avoid artificial limitations on the types of those functions' arguments. As a consequence typecasts from the source code can be removed. Also remove Google's copyright from os0sync.ic because that file no longer contains code from Google. Approved by: Marko (rb://88), also ok from Inaam via IM ------------------------------------------------------------------------ r4163 | marko | 2009-02-12 00:14:19 +0200 (Thu, 12 Feb 2009) | 4 lines branches/zip: Make innodb_thread_concurrency=0 the default. The old default was 8. ------------------------------------------------------------------------ r4169 | calvin | 2009-02-12 10:37:10 +0200 (Thu, 12 Feb 2009) | 3 lines branches/zip: Adjust the result file of innodb_thread_concurrency_basic test. The default value of innodb_thread_concurrency is changed to 0 (from 8) via r4163. ------------------------------------------------------------------------ r4174 | vasil | 2009-02-12 17:38:27 +0200 (Thu, 12 Feb 2009) | 4 lines branches/zip: Fix pathname of the file to patch. ------------------------------------------------------------------------ r4176 | vasil | 2009-02-13 10:06:31 +0200 (Fri, 13 Feb 2009) | 7 lines branches/zip: Fix the failing mysql-test partition_innodb, which failed only if run after innodb_trx_weight (or other test that would leave LATEST DEADLOCK ERROR into the output of SHOW ENGINE INNODB STATUS). Find further explanation for the failure at the top of the added patch partition_innodb.diff. ------------------------------------------------------------------------ r4198 | vasil | 2009-02-17 09:06:07 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: Add the full text of the GPLv2 license into the root directory of the plugin. In previous releases this file was copied from an external source (https://svn.innodb.com/svn/plugin/trunk/support/COPYING) "manually" when creating the source and binary archives. It is less confusing to have this present in the root directory of the SVN branch. ------------------------------------------------------------------------ r4199 | vasil | 2009-02-17 09:11:58 +0200 (Tue, 17 Feb 2009) | 4 lines branches/zip: Add Google's license into COPYING.Google. ------------------------------------------------------------------------ r4200 | vasil | 2009-02-17 09:56:33 +0200 (Tue, 17 Feb 2009) | 11 lines branches/zip: To the files touched by the Google patch from c4144 (excluding include/os0sync.ic because later we removed Google code from that file): * Remove the Google license * Remove old Innobase copyright lines * Add a reference to the Google license and to the GPLv2 license at the top, as recommended by the lawyers at Oracle Legal. ------------------------------------------------------------------------ r4201 | vasil | 2009-02-17 10:12:02 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 1/28] ------------------------------------------------------------------------ r4202 | vasil | 2009-02-17 10:15:06 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 2/28] ------------------------------------------------------------------------ r4203 | vasil | 2009-02-17 10:25:45 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 3/28] ------------------------------------------------------------------------ r4204 | vasil | 2009-02-17 10:55:41 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 4/28] ------------------------------------------------------------------------ r4205 | vasil | 2009-02-17 10:59:22 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 5/28] ------------------------------------------------------------------------ r4206 | vasil | 2009-02-17 11:02:27 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 6/28] ------------------------------------------------------------------------ r4207 | vasil | 2009-02-17 11:04:28 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 7/28] ------------------------------------------------------------------------ r4208 | vasil | 2009-02-17 11:06:49 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 8/28] ------------------------------------------------------------------------ r4209 | vasil | 2009-02-17 11:10:18 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 9/28] ------------------------------------------------------------------------ r4210 | vasil | 2009-02-17 11:12:41 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 10/28] ------------------------------------------------------------------------ r4211 | vasil | 2009-02-17 11:14:40 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 11/28] ------------------------------------------------------------------------ r4212 | vasil | 2009-02-17 11:18:35 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 12/28] ------------------------------------------------------------------------ r4213 | vasil | 2009-02-17 11:24:40 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 13/28] ------------------------------------------------------------------------ r4214 | vasil | 2009-02-17 11:27:31 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 13/28] ------------------------------------------------------------------------ r4215 | vasil | 2009-02-17 11:29:55 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 15/28] ------------------------------------------------------------------------ r4216 | vasil | 2009-02-17 11:33:38 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 16/28] ------------------------------------------------------------------------ r4217 | vasil | 2009-02-17 11:36:44 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 17/28] ------------------------------------------------------------------------ r4218 | vasil | 2009-02-17 11:39:11 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 18/28] ------------------------------------------------------------------------ r4219 | vasil | 2009-02-17 11:41:24 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 19/28] ------------------------------------------------------------------------ r4220 | vasil | 2009-02-17 11:43:50 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 20/28] ------------------------------------------------------------------------ r4221 | vasil | 2009-02-17 11:46:52 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 21/28] ------------------------------------------------------------------------ r4222 | vasil | 2009-02-17 11:50:12 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 22/28] ------------------------------------------------------------------------ r4223 | vasil | 2009-02-17 11:53:58 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 23/28] ------------------------------------------------------------------------ r4224 | vasil | 2009-02-17 12:01:41 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 24/28] ------------------------------------------------------------------------ r4225 | vasil | 2009-02-17 12:05:45 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 25/28] ------------------------------------------------------------------------ r4226 | vasil | 2009-02-17 12:09:16 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 26/28] ------------------------------------------------------------------------ r4227 | vasil | 2009-02-17 12:12:56 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 27/28] ------------------------------------------------------------------------ r4228 | vasil | 2009-02-17 12:14:04 +0200 (Tue, 17 Feb 2009) | 8 lines branches/zip: * Remove old Innobase copyright lines from C source files * Add a reference to the GPLv2 license as recommended by the lawyers at Oracle Legal [Step 28/28] ------------------------------------------------------------------------ r4229 | vasil | 2009-02-17 12:30:55 +0200 (Tue, 17 Feb 2009) | 4 lines branches/zip: Add the copyright notice to the non C files. ------------------------------------------------------------------------ r4231 | marko | 2009-02-17 14:26:53 +0200 (Tue, 17 Feb 2009) | 12 lines Minor cleanup of the Google SMP patch. sync_array_object_signalled(): Add a (void) cast to eliminate a gcc warning about the return value of os_atomic_increment() being ignored. rw_lock_create_func(): Properly indent the preprocessor directives. rw_lock_x_lock_low(), rw_lock_x_lock_func_nowait(): Split lines correctly. rw_lock_set_writer_id_and_recursion_flag(): Silence a Valgrind warning. Do not mix statements and variable declarations. ------------------------------------------------------------------------ r4232 | marko | 2009-02-17 14:59:54 +0200 (Tue, 17 Feb 2009) | 3 lines branches/zip: When assigning lock->recursive = FALSE, also flag lock->writer_thread invalid, so that Valgrind will catch more errors. This is related to Issue #175. ------------------------------------------------------------------------ r4242 | marko | 2009-02-18 17:01:09 +0200 (Wed, 18 Feb 2009) | 2 lines branches/zip: UT_DBG_STOP: Use do{} while(0) to silence a g++-4.3.2 warning about a while(0); statement. This should fix (part of) Issue #176. ------------------------------------------------------------------------ r4243 | marko | 2009-02-18 17:04:03 +0200 (Wed, 18 Feb 2009) | 3 lines branches/zip: buf_buddy_get_slot(): Fix a gcc 4.3.2 warning about an empty body of a "for" statement. This fixes part of Issue #176. ------------------------------------------------------------------------ r4244 | marko | 2009-02-18 17:25:45 +0200 (Wed, 18 Feb 2009) | 11 lines branches/zip: Protect ut_total_allocated_memory with ut_list_mutex. Unprotected updates to ut_total_allocated_memory in os_mem_alloc_large() and os_mem_free_large(), called during fast index creation, may corrupt the variable and cause assertion failures. Also, add UNIV_MEM_ALLOC() and UNIV_MEM_FREE() instrumentation around os_mem_alloc_large() and os_mem_free_large(), so that Valgrind can detect more errors. rb://90 approved by Heikki Tuuri. This addresses Issue #177. ------------------------------------------------------------------------ r4248 | marko | 2009-02-19 11:52:39 +0200 (Thu, 19 Feb 2009) | 2 lines branches/zip: page_zip_set_size(): Fix a g++ 4.3.2 warning about an empty body in a "for" statement. This closes Issue #176. ------------------------------------------------------------------------ r4251 | inaam | 2009-02-19 15:46:27 +0200 (Thu, 19 Feb 2009) | 8 lines branches/zip: Issue #178 rb://91 Change plug.in to have same CXXFLAGS as CFLAGS. This is to ensure that both .c and .cc files get compiled with same flags. To fix the issue where UNIV_LINUX was defined only in .c files. Approved by: Marko ------------------------------------------------------------------------ r4258 | vasil | 2009-02-20 11:52:19 +0200 (Fri, 20 Feb 2009) | 7 lines branches/zip: Cleanup in ChangeLog: * Wrap lines at 78 characters * Changed files are listed alphabetically * White-space cleanup ------------------------------------------------------------------------ r4259 | vasil | 2009-02-20 11:59:42 +0200 (Fri, 20 Feb 2009) | 6 lines branches/zip: ChangeLog: Remove include/os0sync.ic from the entry about the google patch, this file was modified later to not include Google's code. ------------------------------------------------------------------------ r4262 | vasil | 2009-02-20 14:56:59 +0200 (Fri, 20 Feb 2009) | 373 lines branches/zip: Merge revisions 4035:4261 from branches/5.1: ------------------------------------------------------------------------ r4065 | sunny | 2009-01-29 16:01:36 +0200 (Thu, 29 Jan 2009) | 8 lines Changed paths: M /branches/5.1/handler/ha_innodb.cc M /branches/5.1/mysql-test/innodb-autoinc.result M /branches/5.1/mysql-test/innodb-autoinc.test branches/5.1: In the last round of AUTOINC cleanup we assumed that AUTOINC is only defined for integer columns. This caused an assertion failure when we checked for the maximum value of a column type. We now calculate the max value for floating-point autoinc columns too. Fix Bug#42400 - InnoDB autoinc code can't handle floating-point columns rb://84 and Mantis issue://162 ------------------------------------------------------------------------ r4111 | sunny | 2009-02-03 22:06:52 +0200 (Tue, 03 Feb 2009) | 2 lines Changed paths: M /branches/5.1/handler/ha_innodb.cc branches/5.1: Add the ULL suffix otherwise there is an overflow. ------------------------------------------------------------------------ r4128 | vasil | 2009-02-08 21:36:45 +0200 (Sun, 08 Feb 2009) | 18 lines Changed paths: M /branches/5.1/mysql-test/innodb-autoinc.result M /branches/5.1/mysql-test/innodb-autoinc.test branches/5.1: Merge a change from MySQL: ------------------------------------------------------------ revno: 2709.20.31 committer: Timothy Smith <timothy.smith@sun.com> branch nick: 51 timestamp: Fri 2008-12-19 01:28:51 +0100 message: Disable part of innodb-autoinc.test, because the MySQL server asserts when compiled --with-debug, due to bug 39828, "autoinc wraps around when offset and increment > 1". This change should be reverted when that bug is fixed (and a a few other minor changes to the test as described in comments). modified: mysql-test/r/innodb-autoinc.result mysql-test/t/innodb-autoinc.test ------------------------------------------------------------------------ r4129 | vasil | 2009-02-08 21:54:25 +0200 (Sun, 08 Feb 2009) | 310 lines Changed paths: M /branches/5.1/mysql-test/innodb-autoinc.test branches/5.1: Merge a change from MySQL: [looks like the changes to innodb-autoinc.test were made as part of the following huge merge, but we are merging only changes to that file] ------------------------------------------------------------ revno: 2546.47.1 committer: Luis Soares <luis.soares@sun.com> branch nick: 5.1-rpl timestamp: Fri 2009-01-23 13:22:05 +0100 message: merge: 5.1 -> 5.1-rpl conflicts: Text conflict in client/mysqltest.cc Text conflict in mysql-test/include/wait_until_connected_again.inc Text conflict in mysql-test/lib/mtr_report.pm Text conflict in mysql-test/mysql-test-run.pl Text conflict in mysql-test/r/events_bugs.result Text conflict in mysql-test/r/log_state.result Text conflict in mysql-test/r/myisam_data_pointer_size_func.result Text conflict in mysql-test/r/mysqlcheck.result Text conflict in mysql-test/r/query_cache.result Text conflict in mysql-test/r/status.result Text conflict in mysql-test/suite/binlog/r/binlog_index.result Text conflict in mysql-test/suite/binlog/r/binlog_innodb.result Text conflict in mysql-test/suite/rpl/r/rpl_packet.result Text conflict in mysql-test/suite/rpl/t/rpl_packet.test Text conflict in mysql-test/t/disabled.def Text conflict in mysql-test/t/events_bugs.test Text conflict in mysql-test/t/log_state.test Text conflict in mysql-test/t/myisam_data_pointer_size_func.test Text conflict in mysql-test/t/mysqlcheck.test Text conflict in mysql-test/t/query_cache.test Text conflict in mysql-test/t/rpl_init_slave_func.test Text conflict in mysql-test/t/status.test removed: mysql-test/suite/parts/r/partition_bit_ndb.result mysql-test/suite/parts/t/partition_bit_ndb.test mysql-test/suite/parts/t/partition_sessions.test mysql-test/suite/sys_vars/inc/tmp_table_size_basic.inc mysql-test/suite/sys_vars/r/tmp_table_size_basic_32.result mysql-test/suite/sys_vars/r/tmp_table_size_basic_64.result mysql-test/suite/sys_vars/t/tmp_table_size_basic_32.test mysql-test/suite/sys_vars/t/tmp_table_size_basic_64.test mysql-test/t/log_bin_trust_function_creators_func-master.opt mysql-test/t/rpl_init_slave_func-slave.opt added: mysql-test/include/check_events_off.inc mysql-test/include/cleanup_fake_relay_log.inc mysql-test/include/have_simple_parser.inc mysql-test/include/no_running_event_scheduler.inc mysql-test/include/no_running_events.inc mysql-test/include/running_event_scheduler.inc mysql-test/include/setup_fake_relay_log.inc mysql-test/include/wait_condition_sp.inc mysql-test/r/fulltext_plugin.result mysql-test/r/have_simple_parser.require mysql-test/r/innodb_bug38231.result mysql-test/r/innodb_bug39438.result mysql-test/r/innodb_mysql_rbk.result mysql-test/r/partition_innodb_semi_consistent.result mysql-test/r/query_cache_28249.result mysql-test/r/status2.result mysql-test/std_data/bug40482-bin.000001 mysql-test/suite/binlog/r/binlog_innodb_row.result mysql-test/suite/binlog/t/binlog_innodb_row.test mysql-test/suite/rpl/r/rpl_binlog_corruption.result mysql-test/suite/rpl/t/rpl_binlog_corruption-master.opt mysql-test/suite/rpl/t/rpl_binlog_corruption.test mysql-test/suite/sys_vars/r/tmp_table_size_basic.result mysql-test/suite/sys_vars/t/tmp_table_size_basic.test mysql-test/t/fulltext_plugin-master.opt mysql-test/t/fulltext_plugin.test mysql-test/t/innodb_bug38231.test mysql-test/t/innodb_bug39438-master.opt mysql-test/t/innodb_bug39438.test mysql-test/t/innodb_mysql_rbk-master.opt mysql-test/t/innodb_mysql_rbk.test mysql-test/t/partition_innodb_semi_consistent-master.opt mysql-test/t/partition_innodb_semi_consistent.test mysql-test/t/query_cache_28249.test mysql-test/t/status2.test renamed: mysql-test/suite/funcs_1/r/is_collation_character_set_applicability.result => mysql-test/suite/funcs_1/r/is_coll_char_set_appl.result mysql-test/suite/funcs_1/t/is_collation_character_set_applicability.test => mysql-test/suite/funcs_1/t/is_coll_char_set_appl.test modified: .bzr-mysql/default.conf CMakeLists.txt client/mysql.cc client/mysql_upgrade.c client/mysqlcheck.c client/mysqltest.cc configure.in extra/resolve_stack_dump.c extra/yassl/include/openssl/ssl.h include/config-win.h include/m_ctype.h include/my_global.h mysql-test/extra/binlog_tests/database.test mysql-test/extra/rpl_tests/rpl_auto_increment.test mysql-test/include/commit.inc mysql-test/include/have_32bit.inc mysql-test/include/have_64bit.inc mysql-test/include/index_merge1.inc mysql-test/include/linux_sys_vars.inc mysql-test/include/windows_sys_vars.inc mysql-test/lib/mtr_report.pm mysql-test/mysql-test-run.pl mysql-test/r/alter_table.result mysql-test/r/commit_1innodb.result mysql-test/r/create.result mysql-test/r/csv.result mysql-test/r/ctype_ucs.result mysql-test/r/date_formats.result mysql-test/r/events_bugs.result mysql-test/r/events_scheduling.result mysql-test/r/fulltext.result mysql-test/r/func_if.result mysql-test/r/func_in.result mysql-test/r/func_str.result mysql-test/r/func_time.result mysql-test/r/grant.result mysql-test/r/index_merge_myisam.result mysql-test/r/information_schema.result mysql-test/r/innodb-autoinc.result mysql-test/r/innodb.result mysql-test/r/innodb_mysql.result mysql-test/r/log_bin_trust_function_creators_func.result mysql-test/r/log_state.result mysql-test/r/myisampack.result mysql-test/r/mysql.result mysql-test/r/mysqlcheck.result mysql-test/r/partition_datatype.result mysql-test/r/partition_mgm.result mysql-test/r/partition_pruning.result mysql-test/r/query_cache.result mysql-test/r/read_buffer_size_basic.result mysql-test/r/read_rnd_buffer_size_basic.result mysql-test/r/rpl_init_slave_func.result mysql-test/r/select.result mysql-test/r/status.result mysql-test/r/strict.result mysql-test/r/temp_table.result mysql-test/r/type_bit.result mysql-test/r/type_date.result mysql-test/r/type_float.result mysql-test/r/warnings_engine_disabled.result mysql-test/r/xml.result mysql-test/suite/binlog/r/binlog_database.result mysql-test/suite/binlog/r/binlog_index.result mysql-test/suite/binlog/r/binlog_innodb.result mysql-test/suite/binlog/r/binlog_row_mix_innodb_myisam.result mysql-test/suite/binlog/t/binlog_innodb.test mysql-test/suite/funcs_1/r/is_columns_is.result mysql-test/suite/funcs_1/r/is_engines.result mysql-test/suite/funcs_1/r/storedproc.result mysql-test/suite/funcs_1/storedproc/param_check.inc mysql-test/suite/funcs_2/t/disabled.def mysql-test/suite/ndb/t/disabled.def mysql-test/suite/parts/r/partition_bit_innodb.result mysql-test/suite/parts/r/partition_bit_myisam.result mysql-test/suite/parts/r/partition_special_innodb.result mysql-test/suite/parts/t/disabled.def mysql-test/suite/parts/t/partition_special_innodb.test mysql-test/suite/parts/t/partition_value_innodb.test mysql-test/suite/parts/t/partition_value_myisam.test mysql-test/suite/parts/t/partition_value_ndb.test mysql-test/suite/rpl/r/rpl_auto_increment.result mysql-test/suite/rpl/r/rpl_packet.result mysql-test/suite/rpl/r/rpl_row_create_table.result mysql-test/suite/rpl/r/rpl_slave_skip.result mysql-test/suite/rpl/r/rpl_trigger.result mysql-test/suite/rpl/t/disabled.def mysql-test/suite/rpl/t/rpl_packet.test mysql-test/suite/rpl/t/rpl_row_create_table.test mysql-test/suite/rpl/t/rpl_slave_skip.test mysql-test/suite/rpl/t/rpl_trigger.test mysql-test/suite/rpl_ndb/t/disabled.def mysql-test/suite/sys_vars/inc/key_buffer_size_basic.inc mysql-test/suite/sys_vars/inc/sort_buffer_size_basic.inc mysql-test/suite/sys_vars/r/key_buffer_size_basic_32.result mysql-test/suite/sys_vars/r/key_buffer_size_basic_64.result mysql-test/suite/sys_vars/r/sort_buffer_size_basic_32.result mysql-test/suite/sys_vars/r/sort_buffer_size_basic_64.result mysql-test/t/alter_table.test mysql-test/t/create.test mysql-test/t/csv.test mysql-test/t/ctype_ucs.test mysql-test/t/date_formats.test mysql-test/t/disabled.def mysql-test/t/events_bugs.test mysql-test/t/events_scheduling.test mysql-test/t/fulltext.test mysql-test/t/func_if.test mysql-test/t/func_in.test mysql-test/t/func_str.test mysql-test/t/func_time.test mysql-test/t/grant.test mysql-test/t/information_schema.test mysql-test/t/innodb-autoinc.test mysql-test/t/innodb.test mysql-test/t/innodb_mysql.test mysql-test/t/log_bin_trust_function_creators_func.test mysql-test/t/log_state.test mysql-test/t/myisam_data_pointer_size_func.test mysql-test/t/myisampack.test mysql-test/t/mysql.test mysql-test/t/mysqlcheck.test mysql-test/t/partition_innodb_stmt.test mysql-test/t/partition_mgm.test mysql-test/t/partition_pruning.test mysql-test/t/query_cache.test mysql-test/t/rpl_init_slave_func.test mysql-test/t/select.test mysql-test/t/status.test mysql-test/t/strict.test mysql-test/t/temp_table.test mysql-test/t/type_bit.test mysql-test/t/type_date.test mysql-test/t/type_float.test mysql-test/t/warnings_engine_disabled.test mysql-test/t/xml.test mysys/my_getopt.c mysys/my_init.c scripts/mysql_install_db.sh sql-common/my_time.c sql/field.cc sql/field.h sql/filesort.cc sql/ha_partition.cc sql/ha_partition.h sql/item.cc sql/item_cmpfunc.cc sql/item_func.h sql/item_strfunc.cc sql/item_sum.cc sql/item_timefunc.cc sql/item_timefunc.h sql/log.cc sql/log.h sql/log_event.cc sql/log_event.h sql/mysql_priv.h sql/mysqld.cc sql/opt_range.cc sql/partition_info.cc sql/repl_failsafe.cc sql/rpl_constants.h sql/set_var.cc sql/slave.cc sql/spatial.h sql/sql_acl.cc sql/sql_base.cc sql/sql_binlog.cc sql/sql_class.h sql/sql_cursor.cc sql/sql_delete.cc sql/sql_lex.cc sql/sql_lex.h sql/sql_locale.cc sql/sql_parse.cc sql/sql_partition.cc sql/sql_plugin.cc sql/sql_plugin.h sql/sql_profile.cc sql/sql_repl.cc sql/sql_select.cc sql/sql_select.h sql/sql_show.cc sql/sql_table.cc sql/sql_trigger.cc sql/sql_trigger.h sql/table.cc sql/table.h sql/unireg.cc storage/csv/ha_tina.cc storage/federated/ha_federated.cc storage/heap/ha_heap.cc storage/innobase/Makefile.am storage/innobase/btr/btr0sea.c storage/innobase/buf/buf0lru.c storage/innobase/dict/dict0dict.c storage/innobase/dict/dict0mem.c storage/innobase/handler/ha_innodb.cc storage/innobase/handler/ha_innodb.h storage/innobase/include/btr0sea.h storage/innobase/include/dict0dict.h storage/innobase/include/dict0mem.h storage/innobase/include/ha_prototypes.h storage/innobase/include/lock0lock.h storage/innobase/include/row0mysql.h storage/innobase/include/sync0sync.ic storage/innobase/include/ut0ut.h storage/innobase/lock/lock0lock.c storage/innobase/os/os0file.c storage/innobase/plug.in storage/innobase/row/row0mysql.c storage/innobase/row/row0sel.c storage/innobase/srv/srv0srv.c storage/innobase/srv/srv0start.c storage/innobase/ut/ut0ut.c storage/myisam/ft_boolean_search.c strings/ctype.c strings/xml.c tests/mysql_client_test.c win/configure.js mysql-test/suite/funcs_1/t/is_coll_char_set_appl.test ------------------------------------------------------------------------ r4165 | calvin | 2009-02-12 01:34:27 +0200 (Thu, 12 Feb 2009) | 1 line Changed paths: M /branches/5.1/handler/ha_innodb.cc branches/5.1: minor non-functional changes. ------------------------------------------------------------------------ ------------------------------------------------------------------------ r4263 | vasil | 2009-02-20 15:00:46 +0200 (Fri, 20 Feb 2009) | 4 lines branches/zip: Add a ChangeLog entry for a change in r4262. ------------------------------------------------------------------------ r4265 | marko | 2009-02-20 22:31:03 +0200 (Fri, 20 Feb 2009) | 5 lines branches/zip: Make innodb_use_sys_malloc=ON the default. Replace srv_use_sys_malloc with UNIV_LIKELY(srv_use_sys_malloc) to improve branch prediction in the default case. Approved by Ken over the IM. ------------------------------------------------------------------------ r4266 | vasil | 2009-02-20 23:29:32 +0200 (Fri, 20 Feb 2009) | 7 lines branches/zip: Add a sentence at the top of COPYING.Google to clarify that this license does not apply to the whole InnoDB. Suggested by: Ken ------------------------------------------------------------------------ r4268 | marko | 2009-02-23 12:43:51 +0200 (Mon, 23 Feb 2009) | 9 lines branches/zip: Initialize ut_list_mutex at startup. Without this fix, ut_list_mutex would be used uninitialized when innodb_use_sys_malloc=1. This fix addresses Issue #181. ut_mem_block_list_init(): Rename to ut_mem_init() and make public. ut_malloc_low(), ut_free_all_mem(): Add ut_a(ut_mem_block_list_inited). mem_init(): Call ut_mem_init(). ------------------------------------------------------------------------ r4269 | marko | 2009-02-23 15:09:49 +0200 (Mon, 23 Feb 2009) | 7 lines branches/zip: When freeing an uncompressed BLOB page, tolerate garbage in FIL_PAGE_TYPE. (Bug #43043, Issue #182) btr_check_blob_fil_page_type(): New function. btr_free_externally_stored_field(), btr_copy_blob_prefix(): Call btr_check_blob_fil_page_type() to check FIL_PAGE_TYPE. ------------------------------------------------------------------------ r4272 | marko | 2009-02-23 23:10:18 +0200 (Mon, 23 Feb 2009) | 8 lines branches/zip: Adjust the fix of Issue #182 in r4269 per Inaam's suggestion. btr_check_blob_fil_page_type(): Replace the parameter const char* op with ibool read. Do not print anything about page type mismatch when reading a BLOB page in Antelope format. Print space id before page number. ------------------------------------------------------------------------ r4273 | marko | 2009-02-24 00:11:11 +0200 (Tue, 24 Feb 2009) | 1 line branches/zip: ut_mem_init(): Add the assertion !ut_mem_block_list_inited. ------------------------------------------------------------------------ r4274 | marko | 2009-02-24 00:14:38 +0200 (Tue, 24 Feb 2009) | 12 lines branches/zip: Fix bugs in the fix of Issue #181. Tested inside and outside Valgrind, with innodb_use_sys_malloc set to 0 and 1. mem_init(): Invoke ut_mem_init() before mem_pool_create(), because the latter one will invoke ut_malloc(). srv_general_init(): Do not initialize the memory subsystem (mem_init()). innobase_init(): Initialize the memory subsystem (mem_init()) before calling srv_parse_data_file_paths_and_sizes(), which needs ut_malloc(). Call ut_free_all_mem() in error handling to clean up after the mem_init(). ------------------------------------------------------------------------ r4280 | marko | 2009-02-24 15:14:59 +0200 (Tue, 24 Feb 2009) | 1 line branches/zip: Remove unused function os_mem_alloc_nocache(). ------------------------------------------------------------------------ r4281 | marko | 2009-02-24 16:02:48 +0200 (Tue, 24 Feb 2009) | 1 line branches/zip: Remove the unused function dict_index_get_type(). ------------------------------------------------------------------------ r4283 | marko | 2009-02-24 23:06:56 +0200 (Tue, 24 Feb 2009) | 1 line branches/zip: srv0start.c: Remove unnecessary #include "mem0pool.h". ------------------------------------------------------------------------ r4284 | marko | 2009-02-24 23:26:38 +0200 (Tue, 24 Feb 2009) | 1 line branches/zip: mem0mem.c: Remove unnecessary #include "mach0data.h". ------------------------------------------------------------------------ r4288 | vasil | 2009-02-25 10:48:07 +0200 (Wed, 25 Feb 2009) | 21 lines branches/zip: Merge revisions 4261:4287 from branches/5.1: ------------------------------------------------------------------------ r4287 | sunny | 2009-02-25 05:32:01 +0200 (Wed, 25 Feb 2009) | 10 lines Changed paths: M /branches/5.1/handler/ha_innodb.cc M /branches/5.1/mysql-test/innodb-autoinc.result M /branches/5.1/mysql-test/innodb-autoinc.test branches/5.1: Fix Bug#42714 AUTO_INCREMENT errors in 5.1.31. There are two changes to the autoinc handling. 1. To fix the immediate problem from the bug report, we must ensure that the value written to the table is always less than the max value stored in dict_table_t. 2. The second related change is that according to MySQL documentation when the offset is greater than the increment, we should ignore the offset. ------------------------------------------------------------------------ ------------------------------------------------------------------------ r4289 | vasil | 2009-02-25 10:53:51 +0200 (Wed, 25 Feb 2009) | 4 lines branches/zip: Add ChangeLog entry for the fix in r4288. ------------------------------------------------------------------------ r4290 | vasil | 2009-02-25 11:05:44 +0200 (Wed, 25 Feb 2009) | 11 lines branches/zip: Make ChangeLog entries for bugs in bugs.mysql.com in the form: Fix Bug#12345 bug title (for bugs after 1.0.2 was released and the ChangeLog published) There is no need to bloat the ChangeLog with information that is available via bugs.mysql.com. Discussed with: Marko ------------------------------------------------------------------------ r4291 | vasil | 2009-02-25 11:08:32 +0200 (Wed, 25 Feb 2009) | 4 lines branches/zip: Fix Bug synopsis and remove explanation ------------------------------------------------------------------------ r4292 | marko | 2009-02-25 12:09:15 +0200 (Wed, 25 Feb 2009) | 25 lines branches/zip: Correct the initialization of the memory subsystem once again, to finally put Issue #181 to rest. Revert some parts of r4274. It is best not to call ut_malloc() before srv_general_init(). mem_init(): Do not call ut_mem_init(). srv_general_init(): Initialize the memory subsystem in two phases: first ut_mem_init(), then mem_init(). This is because os_sync_init() and sync_init() depend on ut_mem_init() and mem_init() depends on os_sync_init() or sync_init(). srv_parse_data_file_paths_and_sizes(), srv_parse_log_group_home_dirs(): Remove the output parameters. Assign to the global variables directly. Allocate memory with malloc() instead of ut_malloc(), because these functions will be called before srv_general_init(). srv_free_paths_and_sizes(): New function, for cleaning up after srv_parse_data_file_paths_and_sizes() and srv_parse_log_group_home_dirs(). rb://92 approved by Sunny Bains ------------------------------------------------------------------------ r4297 | vasil | 2009-02-25 17:19:19 +0200 (Wed, 25 Feb 2009) | 4 lines branches/zip: White-space cleanup in the ChangeLog ------------------------------------------------------------------------ r4301 | vasil | 2009-02-25 21:33:32 +0200 (Wed, 25 Feb 2009) | 5 lines branches/zip: Do not output the commands that restore the environment because they depend on the state of the environment before the test starts executing. ------------------------------------------------------------------------ r4315 | vasil | 2009-02-26 09:21:20 +0200 (Thu, 26 Feb 2009) | 5 lines branches/zip: Apply any necessary patches to the mysql tree at the end of setup.sh This step was previously done manually (and sometimes forgotten). ------------------------------------------------------------------------ r4319 | marko | 2009-02-26 23:27:51 +0200 (Thu, 26 Feb 2009) | 6 lines branches/zip: btr_check_blob_fil_page_type(): Do not report FIL_PAGE_TYPE mismatch even when purging a BLOB. Heavy users may have large data files created with MySQL 5.0 or earlier, and they don not want to have the error log flooded with such messages. This fixes Issue #182. ------------------------------------------------------------------------ r4320 | inaam | 2009-02-27 02:13:19 +0200 (Fri, 27 Feb 2009) | 8 lines branches/zip This is to revert the changes made to the plug.in (r4251) as a fix for issue# 178. Changes to plug.in will not propogate to a plugin installation unless autotools are rerun which is unacceptable. A fix for issue# 178 will be committed in a separate commit. ------------------------------------------------------------------------ r4321 | inaam | 2009-02-27 02:16:46 +0200 (Fri, 27 Feb 2009) | 6 lines branches/zip This is a fix for issue#178. Instead of using UNIV_LINUX which is defined through CFLAGS we use compiler generated define __linux__ that is effective for both .c and .cc files. ------------------------------------------------------------------------ r4324 | vasil | 2009-02-27 13:27:18 +0200 (Fri, 27 Feb 2009) | 39 lines branches/zip: Add FreeBSD to the list of the operating systems that have sizeof(pthread_t) == sizeof(void*) (i.e. word size). On FreeBSD pthread_t is defined like: /usr/include/sys/_pthreadtypes.h: typedef struct pthread *pthread_t; I did the following tests (per Inaam's recommendation): a) appropriate version of GCC is available on that platform (4.1.2 or higher for atomics to be available) On FreeBSD 6.x the default compiler is 3.4.6, on FreeBSD 7.x the default one is 4.2.1. One can always install the version of choice from the ports collection. If gcc 3.x is used then HAVE_GCC_ATOMIC_BUILTINS will not be defined and thus the change I am committing will make no difference. b) find out if sizeof(pthread_t) == sizeof(long) On 32 bit both are 4 bytes, on 64 bit both are 8 bytes. c) find out the compiler generated platform define (e.g.: __aix, __sunos__ etc.) The macro is __FreeBSD__. d) patch univ.i with the appropriate platform define e) build the mysql f) ensure it is using atomic builtins (look at the err.log message at system startup. It should say we are using atomics for both mutexes and rw-locks) g) do sanity testing (keeping in view the smp changes) I ran the mysql-test suite. All tests pass. ------------------------------------------------------------------------ r4353 | vasil | 2009-03-05 09:27:29 +0200 (Thu, 05 Mar 2009) | 6 lines branches/zip: As suggested by Ken, print a message that says that the Google SMP patch (GCC atomics) is disabled if it is. Also extend the message when the patch is partially enabled to make it clear that it is partially enabled. ------------------------------------------------------------------------ r4356 | vasil | 2009-03-05 13:49:51 +0200 (Thu, 05 Mar 2009) | 4 lines branches/zip: Fix typo made in r4353. ------------------------------------------------------------------------ r4357 | vasil | 2009-03-05 16:38:59 +0200 (Thu, 05 Mar 2009) | 23 lines branches/zip: Implement a check whether pthread_t objects can be used by GCC atomic builtin functions. This check is implemented in plug.in and defines the macro HAVE_ATOMIC_PTHREAD_T. This macro is checked in univ.i and the relevant part of the code enabled (the one that uses GCC atomics against pthread_t objects). In addition to this, the same program that is compiled as part of the plug.in check is added in ut/ut0auxconf.c. In the InnoDB Plugin source archives that are shipped to the users, a generated Makefile.in is added. That Makefile.in will be modified to compile ut/ut0auxconf.c and define the macro HAVE_ATOMIC_PTHREAD_T if the compilation succeeds. I.e. Makefile.in will emulate the work that is done by plug.in. This is done in order to make the check happen and HAVE_ATOMIC_PTHREAD_T eventually defined without regenerating MySQL's ./configure from ./storage/innobase/plug.in. The point is not to ask users to install the autotools and regenerate ./configure. rb://95 Approved by: Marko ------------------------------------------------------------------------ r4360 | vasil | 2009-03-05 22:23:17 +0200 (Thu, 05 Mar 2009) | 21 lines branches/zip: Merge revisions 4287:4357 from branches/5.1: ------------------------------------------------------------------------ r4325 | sunny | 2009-03-02 02:28:52 +0200 (Mon, 02 Mar 2009) | 10 lines Changed paths: M /branches/5.1/handler/ha_innodb.cc M /branches/5.1/mysql-test/innodb-autoinc.result M /branches/5.1/mysql-test/innodb-autoinc.test branches/5.1: Bug#43203: Overflow from auto incrementing causes server segv It was not a SIGSEGV but an assertion failure. The assertion was checking the invariant that *first_value passed in by MySQL doesn't contain a value that is greater than the max value for that type. The assertion has been changed to a check and if the value is greater than the max we report a generic AUTOINC failure. rb://93 Approved by Heikki ------------------------------------------------------------------------ ------------------------------------------------------------------------ r4361 | vasil | 2009-03-05 22:27:54 +0200 (Thu, 05 Mar 2009) | 30 lines branches/zip: Merge revision 4358 from branches/5.1 (resolving a conflict): ------------------------------------------------------------------------ r4358 | vasil | 2009-03-05 21:21:10 +0200 (Thu, 05 Mar 2009) | 21 lines Changed paths: M /branches/5.1/handler/ha_innodb.cc branches/5.1: Merge a change from MySQL: ------------------------------------------------------------ revno: 2728.19.1 committer: Alfranio Correia <alfranio.correia@sun.com> branch nick: mysql-5.1-bugteam timestamp: Tue 2009-02-03 11:36:46 +0000 message: BUG#42445 Warning messages in innobase/handler/ha_innodb.cc There was a type casting problem in the storage/innobase/handler/ha_innodb.cc, (int ha_innobase::write_row(...)). Innobase uses has an internal error variable of type 'ulint' while mysql uses an 'int'. To fix the problem the function manipulates an error variable of type 'ulint' and only casts it into 'int' when needs to return the value. modified: storage/innobase/handler/ha_innodb.cc ------------------------------------------------------------------------ ------------------------------------------------------------------------ r4362 | vasil | 2009-03-05 22:29:07 +0200 (Thu, 05 Mar 2009) | 23 lines branches/zip: Merge revision 4359 from branches/5.1: ------------------------------------------------------------------------ r4359 | vasil | 2009-03-05 21:42:01 +0200 (Thu, 05 Mar 2009) | 14 lines Changed paths: M /branches/5.1/handler/ha_innodb.cc branches/5.1: Merge a change from MySQL: ------------------------------------------------------------ revno: 2747 committer: Timothy Smith <timothy.smith@sun.com> branch nick: 51 timestamp: Fri 2009-01-16 17:49:07 +0100 message: Add another cast to ignore int/ulong difference in error types, silence warning on Win64 modified: storage/innobase/handler/ha_innodb.cc ------------------------------------------------------------------------ ------------------------------------------------------------------------ r4363 | vasil | 2009-03-05 22:31:37 +0200 (Thu, 05 Mar 2009) | 4 lines branches/zip: Add ChangeLog entry for the bugfix in c4360. ------------------------------------------------------------------------ r4378 | calvin | 2009-03-09 10:10:17 +0200 (Mon, 09 Mar 2009) | 7 lines branches/zip: remove compile flag MYSQL_SERVER for dynamic plugin The dynamic plugin on Windows used to be built with MYSQL_SERVER compile flag, while it is not the case for other platforms. r3797 assumed MYSQL_SERVER was not defined for dynamic plugin, which introduced the engine crash during dropping a database. ------------------------------------------------------------------------ r4396 | marko | 2009-03-12 09:22:27 +0200 (Thu, 12 Mar 2009) | 3 lines branches/zip: btr_store_big_rec_extern_fields(): Initialize FIL_PAGE_TYPE in a separate redo log entry. This will make ibbackup --apply-log debugging easier. ------------------------------------------------------------------------ r4397 | marko | 2009-03-12 09:26:11 +0200 (Thu, 12 Mar 2009) | 3 lines branches/zip: trx_sys_create_doublewrite_buf(): As the dummy change, initialize FIL_PAGE_TYPE. This will make it easier to write the debug assertions for ibbackup --apply-log. ------------------------------------------------------------------------ r4401 | marko | 2009-03-12 10:26:40 +0200 (Thu, 12 Mar 2009) | 19 lines branches/zip: Merge revisions 4359:4400 from branches/5.1: ------------------------------------------------------------------------ r4399 | marko | 2009-03-12 09:38:05 +0200 (Thu, 12 Mar 2009) | 2 lines branches/5.1: row_sel_get_clust_rec_for_mysql(): Store the cursor position also for unlock_row(). (Bug #39320) ------------------------------------------------------------------------ r4400 | marko | 2009-03-12 10:06:44 +0200 (Thu, 12 Mar 2009) | 5 lines branches/5.1: Fix a bug in multi-table semi-consistent reads. Remember the acquired record locks per table handle (row_prebuilt_t) rather than per transaction (trx_t), so that unlock_row should successfully unlock all non-matching rows in multi-table operations. This deficiency was found while investigating Bug #39320. ------------------------------------------------------------------------ These were submitted as rb://94 and rb://96 and approved by Heikki Tuuri. ------------------------------------------------------------------------ r4455 | marko | 2009-03-16 11:43:34 +0200 (Mon, 16 Mar 2009) | 2 lines branches/zip: UT_LIST_VALIDATE(): Add the parameter ASSERTION and adjust all callers. ------------------------------------------------------------------------ r4456 | marko | 2009-03-16 12:59:25 +0200 (Mon, 16 Mar 2009) | 6 lines branches/zip: UT_LIST_VALIDATE(): Assert that the link is non-NULL before dereferencing it. In this way, ut_list_node_313 will be pointing to the last non-NULL list item at the time of the assertion failure. (gcc-4.3.2 -O3 seems to optimize the common subexpressions and make the variable NULL, though.) ------------------------------------------------------------------------ r4457 | marko | 2009-03-16 14:12:02 +0200 (Mon, 16 Mar 2009) | 2 lines branches/zip: sync_thread_add_level(): Make the assertions about level == SYNC_BUF_BLOCK more readable. ------------------------------------------------------------------------ r4461 | vasil | 2009-03-17 09:38:19 +0200 (Tue, 17 Mar 2009) | 6 lines branches/zip: Remove mysql-test/patches/bug32625.diff because that bug was fixed in the mysql repository (1 year and 4 months after sending them the simple patch!). See http://bugs.mysql.com/32625 ------------------------------------------------------------------------ r4465 | marko | 2009-03-17 12:34:19 +0200 (Tue, 17 Mar 2009) | 1 line branches/zip: buf0buddy.c: Add and adjust some debug assertions. ------------------------------------------------------------------------ r4473 | vasil | 2009-03-17 15:50:30 +0200 (Tue, 17 Mar 2009) | 5 lines branches/zip: Increment the InnoDB Plugin version from 1.0.3 to 1.0.4 now that 1.0.3 has been released. ------------------------------------------------------------------------ r4478 | vasil | 2009-03-18 11:53:53 +0200 (Wed, 18 Mar 2009) | 5 lines branches/zip: Remove mysql-test/patches/bug41893.diff because that bug has been fixed in the MySQL repository, see http://bugs.mysql.com/41893. ------------------------------------------------------------------------ r4479 | marko | 2009-03-18 12:43:54 +0200 (Wed, 18 Mar 2009) | 2 lines branches/zip: buf_LRU_block_remove_hashed_page(): Add some debug assertions. ------------------------------------------------------------------------ r4480 | marko | 2009-03-18 14:32:13 +0200 (Wed, 18 Mar 2009) | 1 line branches/zip: buf_buddy_free_low(): Correct the function comment. ------------------------------------------------------------------------ r4482 | marko | 2009-03-19 15:23:32 +0200 (Thu, 19 Mar 2009) | 12 lines branches/zip: Merge revisions 4400:4481 from branches/5.1: ------------------------------------------------------------------------ r4481 | marko | 2009-03-19 15:01:48 +0200 (Thu, 19 Mar 2009) | 6 lines branches/5.1: row_unlock_for_mysql(): Do not unlock records that were modified by the current transaction. This bug was introduced or unmasked in r4400. rb://97 approved by Heikki Tuuri ------------------------------------------------------------------------ ------------------------------------------------------------------------ r4490 | marko | 2009-03-20 12:33:33 +0200 (Fri, 20 Mar 2009) | 4 lines branches/zip: Non-functional change for reducing dependencies in InnoDB Hot Backup: Replace srv_sys->dummy_ind1 and srv_sys->dummy_ind2 with dict_ind_redundant and dict_ind_compact, initialized in dict_init(). ------------------------------------------------------------------------ r4491 | marko | 2009-03-20 12:45:18 +0200 (Fri, 20 Mar 2009) | 2 lines branches/zip: Add const qualifiers or in/out comments to some function parameters in log0log. ------------------------------------------------------------------------ r4492 | marko | 2009-03-20 12:52:14 +0200 (Fri, 20 Mar 2009) | 5 lines branches/zip: page_validate(): Always report the space id and the name of the index. In Hot Backup, do not invoke comparison functions, as MySQL collations will be unavailable. ------------------------------------------------------------------------ r4493 | marko | 2009-03-20 13:24:06 +0200 (Fri, 20 Mar 2009) | 1 line branches/zip: Replace fil_get_space_for_id_low() with fil_space_get_by_id(). ------------------------------------------------------------------------ r4494 | marko | 2009-03-20 13:51:35 +0200 (Fri, 20 Mar 2009) | 3 lines branches/zip: fil0fil.c: Refer to fil_system directly, not via local vars. This eliminates some "unused variable" warnings when building InnoDB Hot Backup in such a way that all mutex operations are no-ops. ------------------------------------------------------------------------ r4495 | marko | 2009-03-20 14:15:52 +0200 (Fri, 20 Mar 2009) | 1 line branches/zip: innobase_get_at_most_n_mbchars(): Declare in ha_prototypes.h. ------------------------------------------------------------------------ r4496 | marko | 2009-03-20 14:48:26 +0200 (Fri, 20 Mar 2009) | 1 line branches/zip: recv_recover_page(): Remove compile-time constant parameters. ------------------------------------------------------------------------ r4497 | marko | 2009-03-20 14:56:19 +0200 (Fri, 20 Mar 2009) | 1 line branches/zip: recv_sys_init(): Remove a compile-time constant parameter. ------------------------------------------------------------------------ r4498 | marko | 2009-03-20 15:08:05 +0200 (Fri, 20 Mar 2009) | 4 lines branches/zip: Non-functional change: Add const qualifiers. log_block_checksum_is_ok_or_old_format(), recv_sys_add_to_parsing_buf(): The log block is read-only. Make it const. ------------------------------------------------------------------------ r4499 | marko | 2009-03-20 15:10:25 +0200 (Fri, 20 Mar 2009) | 1 line branches/zip: recv_scan_log_recs(): Remove a compile-time constant parameter. ------------------------------------------------------------------------ r4500 | marko | 2009-03-20 15:47:17 +0200 (Fri, 20 Mar 2009) | 1 line branches/zip: fil_init(): Add the parameter hash_size. ------------------------------------------------------------------------ r4501 | vasil | 2009-03-20 16:50:41 +0200 (Fri, 20 Mar 2009) | 4 lines branches/zip: Add any entry about the release of 1.0.3 in the ChangeLog. ------------------------------------------------------------------------ r4515 | marko | 2009-03-23 10:49:53 +0200 (Mon, 23 Mar 2009) | 1 line branches/zip: hash_table_t: adaptive: Remove from UNIV_HOTBACKUP builds. ------------------------------------------------------------------------ r4516 | marko | 2009-03-23 10:57:16 +0200 (Mon, 23 Mar 2009) | 2 lines branches/zip: Define and use ASSERT_HASH_MUTEX_OWN. Make it a no-op in UNIV_HOTBACKUP builds. ------------------------------------------------------------------------ r4517 | marko | 2009-03-23 11:07:20 +0200 (Mon, 23 Mar 2009) | 2 lines branches/zip: Define and use PAGE_ZIP_MATCH. In UNIV_HOTBACKUP builds, assume fixed allocation. ------------------------------------------------------------------------ r4521 | marko | 2009-03-23 12:05:47 +0200 (Mon, 23 Mar 2009) | 1 line branches/zip: buf_page_print(): Clean up the code #ifdef UNIV_HOTBACKUP. ------------------------------------------------------------------------ r4522 | marko | 2009-03-23 12:20:50 +0200 (Mon, 23 Mar 2009) | 2 lines branches/zip: Exclude some operating system interface code from UNIV_HOTBACKUP builds. ------------------------------------------------------------------------ r4523 | marko | 2009-03-23 13:00:43 +0200 (Mon, 23 Mar 2009) | 2 lines branches/zip: Remove the remaining references to hash_table_t::adapive from UNIV_HOTBACKUP builds. This should have been done in r4515. ------------------------------------------------------------------------ r4524 | marko | 2009-03-23 14:05:18 +0200 (Mon, 23 Mar 2009) | 2 lines branches/zip: Enclose recv_recovery_from_backup_on and recv_recovery_from_backup_is_on() in #ifdef UNIV_LOG_ARCHIVE. ------------------------------------------------------------------------ r4525 | marko | 2009-03-23 14:57:45 +0200 (Mon, 23 Mar 2009) | 2 lines branches/zip: recv_parse_or_apply_log_rec_body(): Add debug assertions ensuring that FIL_PAGE_TYPE makes sense when applying log records. ------------------------------------------------------------------------ r4526 | marko | 2009-03-23 16:21:34 +0200 (Mon, 23 Mar 2009) | 2 lines branches/zip: Remove unneeded definitions and dependencies from UNIV_HOTBACKUP builds. ------------------------------------------------------------------------ r4527 | calvin | 2009-03-23 23:15:33 +0200 (Mon, 23 Mar 2009) | 5 lines branches/zip: adjust build files on Windows Adjust the patch positions based on the latest MySQL source. Also add the patches to the .bat files for vs9. ------------------------------------------------------------------------
2009-03-24 09:32:21 +01:00
/*****************************************************************************
Copyright (c) 2005, 2009, Innobase Oy. All Rights Reserved.
This program is free software; you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free Software
Foundation; version 2 of the License.
This program is distributed in the hope that it will be useful, but WITHOUT
ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License along with
this program; if not, write to the Free Software Foundation, Inc., 59 Temple
Place, Suite 330, Boston, MA 02111-1307 USA
*****************************************************************************/
/******************************************************
Smart ALTER TABLE
*******************************************************/
#include <mysql_priv.h>
#include <mysqld_error.h>
extern "C" {
#include "log0log.h"
#include "row0merge.h"
#include "srv0srv.h"
#include "trx0trx.h"
#include "trx0roll.h"
#include "ha_prototypes.h"
#include "handler0alter.h"
}
#include "ha_innodb.h"
branches/innodb+: Merge revisions 2867:2986 from branches/zip: ------------------------------------------------------------------------ r2867 | marko | 2008-10-24 10:24:17 +0300 (Fri, 24 Oct 2008) | 2 lines branches/zip: ChangeLog: Document r2763, r2794, r2683, r2799, r2809, r2866. ------------------------------------------------------------------------ r2869 | vasil | 2008-10-24 11:14:16 +0300 (Fri, 24 Oct 2008) | 4 lines branches/zip: White space cleanup in ChangeLog ------------------------------------------------------------------------ r2870 | vasil | 2008-10-24 13:36:14 +0300 (Fri, 24 Oct 2008) | 8 lines branches/zip: Remove a statement that causes the innodb-index test to fail. The change in behavior was introduced in MySQL BZR-r2738. Suggested by: Marko ------------------------------------------------------------------------ r2871 | vasil | 2008-10-24 13:48:38 +0300 (Fri, 24 Oct 2008) | 5 lines branches/zip: Adjust mysql-test/patches/innodb-index.diff after the change to mysql-test/innodb-index.(test|result) in r2870. ------------------------------------------------------------------------ r2878 | calvin | 2008-10-27 11:05:42 +0200 (Mon, 27 Oct 2008) | 8 lines branches/zip: port the fix of Bug#19424 - InnoDB: Possibly a memory overrun of the buffer being freed with 64-bit Microsoft Visual C++. The changed file: CMakeLists.txt: Removing Win64 compiler optimizations for all innodb/mem/* files. ------------------------------------------------------------------------ r2884 | vasil | 2008-10-27 11:48:46 +0200 (Mon, 27 Oct 2008) | 7 lines branches/zip: ChangeLog: Add entry for the fix of Bug#19424 InnoDB: Possibly a memory overrun of the buffer being freed (64-bit Visual C) ------------------------------------------------------------------------ r2886 | calvin | 2008-10-27 22:39:11 +0200 (Mon, 27 Oct 2008) | 8 lines branches/zip: This patch is to solve the issue that file handles can not cross DLL/EXE boundaries on Windows. In builtin InnoDB, it makes call to MySQL server for creating tmp files. innobase_mysql_tmpfile is now rewritten for the plugin. rb://5 Approved by: Marko ------------------------------------------------------------------------ r2887 | calvin | 2008-10-27 22:48:29 +0200 (Mon, 27 Oct 2008) | 44 lines branches/zip: implement the delayloading of externals for the plugin on Windows, which includes: * Load mysqld.map and insert all symbol/address pairs into hash for quick access * Resolves all external data variables. The delayloading mechanism in MSVC does not support automatic imports of data variables. A workaround is to explicitly handle the data import using the delay loader during the initialization of the plugin. * Resolves all external functions during run-time, by implementing the delayed loading helper function delayLoadHelper2, which is called by run-time as well as HrLoadAllImportsForDll. The delay loader reuses the hash implementation in InnoDB. The normal hash_create (in hash0hash.c) creates hash tables in buffer pool. But the delay loader is invoked before the engine is initialized, and buffer pools are not ready yet. Instead, the delay loader has its own implementation of hash_create() and hash_table_free(), called wdl_hash_create() and wdl_hash_table_free(). This patch should be used with other two patches in order to build a dynamically linked plugin on Windows: * patch for tmpfile functions (r2886) * patch for "build" files (to be committed) The list of file changed: handler/handler0vars.h: new file, defines a list of external data variables (no external functions). handler/win_delay_loader.cc: new file, the implementation of the delay loader for Windows plugin. handler/ha_innodb.cc: add a header file, and changes for copying the system variables. handler/handler0alter.cc: add a header file handler/i_s.cc: add a header file rb://27 Reviewed by: Sunny, Marko Approved by: Sunny ------------------------------------------------------------------------ r2888 | calvin | 2008-10-28 01:51:49 +0200 (Tue, 28 Oct 2008) | 25 lines branches/zip: for building dynamic plugin on Windows, ha_innodb.dll, when INNODB_DYNAMIC_PLUGIN is specified. The changes are: CMakeLists.txt: add project ha_innodb for dynamic plugin on Windows. ha_innodb depends on project mysqld. ha_innodb.def: a new file with standard exports for a dynamic plugin. Two new files will be added: * sql/mysqld.def: .def file for 32-bit compiler * sql/mysqld_x64.def: .def file for x64 compiler It is also required to apply a patch to the MySQL source tree. The patch is described in win-plugin/README: win-plugin/win-plugin.diff - a patch to be applied to MySQL source tree. When applied, the following files will be modified: * CMakeLists.txt: add INNODB_DYNAMIC_PLUGIN and _USE_32BIT_TIME_T * sql/CMakeLists.txt: add mysqld.def or mysqld_x64.def for mysqld * win/configure.js: add INNODB_DYNAMIC_PLUGIN * win/build-vs71.bat: provide an option to specify CMAKE_BUILD_TYPE * win/build-vs8.bat: provide an option to specify CMAKE_BUILD_TYPE * win/build-vs8_x64.bat: provide an option to specify CMAKE_BUILD_TYPE ------------------------------------------------------------------------ r2894 | marko | 2008-10-28 08:36:39 +0200 (Tue, 28 Oct 2008) | 4 lines branches/zip: dict_str_starts_with_keyword(): Removed this unused function. Spotted by Sunny. ------------------------------------------------------------------------ r2895 | vasil | 2008-10-28 08:40:45 +0200 (Tue, 28 Oct 2008) | 6 lines branches/zip: ChangeLog: add entry for the Windows plugin. ------------------------------------------------------------------------ r2917 | marko | 2008-10-28 23:53:23 +0200 (Tue, 28 Oct 2008) | 3 lines branches/zip: innodb_plugin_init(): Do not copy session variables, even when the variable is a global variable in the built-in InnoDB. ------------------------------------------------------------------------ r2918 | calvin | 2008-10-29 00:08:11 +0200 (Wed, 29 Oct 2008) | 2 lines branches/zip: fix a problem introduced in r2917 - dyn is not initialized. Move the check into for(). ------------------------------------------------------------------------ r2922 | calvin | 2008-10-29 08:29:01 +0200 (Wed, 29 Oct 2008) | 16 lines branches/zip: fix issue #102 - Windows plugin: resolve dbug functions during run-time. Implement wrapper functions in the plugin. The plugin will get the function entries from mysqld.exe during the init, and invoke the corresponding functions (in mysqld.exe). The list of functions are: _db_pargs_ _db_doprnt_ _db_enter_ _db_return_ _db_dump_ rb://38 Approved by: Marko ------------------------------------------------------------------------ r2923 | marko | 2008-10-29 09:52:30 +0200 (Wed, 29 Oct 2008) | 1 line branches/zip: ChangeLog: Mention Bug #27276. ------------------------------------------------------------------------ r2925 | calvin | 2008-10-29 10:09:41 +0200 (Wed, 29 Oct 2008) | 16 lines branches/zip: change function names in sql/mysqld.def in order to work with 5.1.29-rc. In 5.1.29, the following function names are changed: _hash_init hash_free hash_search hash_delete changed to _my_hash_init my_hash_free my_hash_search my_hash_delete Approved by: Marko (on IM) ------------------------------------------------------------------------ r2927 | marko | 2008-10-29 11:43:23 +0200 (Wed, 29 Oct 2008) | 4 lines branches/zip: ha_innodb.cc: Make some functions static, so that they will not be compiled as weak global symbols. These functions must not be redirected to the built-in InnoDB. ------------------------------------------------------------------------ r2928 | michael | 2008-10-29 19:20:10 +0200 (Wed, 29 Oct 2008) | 4 lines Remove unnecessary assert Approved by: Heikki, over IM ------------------------------------------------------------------------ r2930 | marko | 2008-10-29 21:39:24 +0200 (Wed, 29 Oct 2008) | 33 lines branches/zip: Merge revisions 2854:2929 from branches/5.1, except r2924, which was merged from branches/zip r2866 to branches/5.1 and except r2879 which was merged separately by Calvin: ------------------------------------------------------------------------ r2902 | vasil | 2008-10-28 12:10:25 +0200 (Tue, 28 Oct 2008) | 10 lines branches/5.1: Fix Bug#38189 innodb_stats_on_metadata missing Make the variable innodb_stats_on_metadata visible to the users and also settable at runtime. Previously it was only "visible" as a command line startup option to mysqld. Approved by: Marko (https://svn.innodb.com/rb/r/36) ------------------------------------------------------------------------ r2929 | marko | 2008-10-29 21:26:14 +0200 (Wed, 29 Oct 2008) | 13 lines branches/5.1: dtype_get_sql_null_size(): return the correct storage size of a SQL NULL column. (Bug #40369) When MySQL Bug #20877 was fixed in r834, this function was accidentally modified to return 0 or 1. Apparently, the only impact of this bug is that fixed-length columns cannot be updated in-place from or to SQL NULL, even in ROW_FORMAT=REDUNDANT. After this fix, fixed-length columns in ROW_FORMAT=REDUNDANT will have a constant storage size as they should, no matter if NULL or non-NULL. The bug caused fixed-length NULL columns to occupy 1 byte. rb://37 approved by Heikki over IM. ------------------------------------------------------------------------ ------------------------------------------------------------------------ r2931 | vasil | 2008-10-29 22:10:40 +0200 (Wed, 29 Oct 2008) | 4 lines branches/zip: Add 2 ChangeLog entries for the 2 bugfixes that were merged from branches/5.1. ------------------------------------------------------------------------ r2935 | vasil | 2008-10-30 12:17:23 +0200 (Thu, 30 Oct 2008) | 17 lines branches/zip: Fix "Bug#40360 Binlog related errors with binlog off" in InnoDB code in order to have a Bug#40360-free InnoDB Plugin 1.0.2. The fix does check whether binary logging is enabled in MySQL by accessing the opt_bin_log global variable that is defined in sql/mysqld.cc. In case MySQL does develop another solution to this via Bug#40360 then we can revert this patch (except the mysql-tests). The windows-plugin part of this fix will be committed as a separate commit to ease eventual merge into branches/5.1 [note from the future: the separate commit went into r2936]. Approved by: Marko (https://svn.innodb.com/rb/r/39) ------------------------------------------------------------------------ r2936 | vasil | 2008-10-30 12:24:09 +0200 (Thu, 30 Oct 2008) | 7 lines branches/zip: Followup to r2935: add the Windows Delay Loader stuff for the MySQL variable that we are accessing. If someday we have another solution for Bug#40360 Binlog related errors with binlog off then this should also be reverted. ------------------------------------------------------------------------ r2937 | vasil | 2008-10-30 12:28:47 +0200 (Thu, 30 Oct 2008) | 4 lines branches/zip: Add ChangeLog entry for Bug#40360 Binlog related errors with binlog off ------------------------------------------------------------------------ r2938 | vasil | 2008-10-30 12:33:28 +0200 (Thu, 30 Oct 2008) | 5 lines branches/zip: Non-functional change: convert handler/handler0vars.h and handler/win_delay_loader.cc from \r\n (dos) to \n (unix) line terminators. ------------------------------------------------------------------------ r2939 | marko | 2008-10-30 12:38:18 +0200 (Thu, 30 Oct 2008) | 2 lines branches/zip: Set svn:eol-style native on some recently added text files. ------------------------------------------------------------------------ r2940 | marko | 2008-10-30 12:46:21 +0200 (Thu, 30 Oct 2008) | 1 line branches/zip: ChangeLog, ha_innodb.def: Set svn:eol-style native ------------------------------------------------------------------------ r2941 | vasil | 2008-10-30 19:34:27 +0200 (Thu, 30 Oct 2008) | 4 lines branches/zip: Increment the InnoDB Plugin version from 1.0.1 to 1.0.2. ------------------------------------------------------------------------ r2943 | sunny | 2008-10-31 09:40:29 +0200 (Fri, 31 Oct 2008) | 15 lines branches/zip: 1. We add a vector of locks to trx_t. This array contains the autoinc locks granted to a transaction. There is one per table. 2. We enforce releasing of these locks in the reverse order from the one in which they are acquired. The assumption is that since the AUTOINC locks are statement level locks. Nested statements introduced by triggers are stacked it should hold. There was some cleanup done to the vector code too by adding const and some new functions. Rename dict_table_t::auto_inc_lock to autoinc_lock. Fix Bug#26316 Triggers create duplicate entries on auto-increment columns rb://22 ------------------------------------------------------------------------ r2944 | vasil | 2008-10-31 09:44:16 +0200 (Fri, 31 Oct 2008) | 12 lines branches/zip: Revert our temporary fix for "Bug#40360 Binlog related errors with binlog off" (r2935, r2936) and deploy MySQL's one, but put the function mysql_bin_log_is_engaged() inside mysql_addons.cc instead of in mysql's log.cc and use a different name for it so there is no collision when MySQL adds this function in log.cc. [note from the future: the windows part of this patch went into r2947] Approved by: Marko (https://svn.innodb.com/rb/r/41/) ------------------------------------------------------------------------ r2945 | sunny | 2008-10-31 09:44:45 +0200 (Fri, 31 Oct 2008) | 2 lines branches/zip: Update ChangeLog with r2943 info. ------------------------------------------------------------------------ r2946 | marko | 2008-10-31 10:18:47 +0200 (Fri, 31 Oct 2008) | 2 lines branches/zip: Revert the unintended change to univ.i that was made in r2943. ------------------------------------------------------------------------ r2947 | calvin | 2008-10-31 10:38:26 +0200 (Fri, 31 Oct 2008) | 6 lines branches/zip: Windows plugin part of r2944 r2944 has reference to mysql_bin_log.is_open(), which is new in InnoDB. Add two new entries and remove one duplicate in mysqld.def & mysqld_x64.def. ------------------------------------------------------------------------ r2948 | vasil | 2008-10-31 11:39:07 +0200 (Fri, 31 Oct 2008) | 9 lines branches/zip: Fix Mantis issue#106 plugin init error:InnoDB: stats_on_metadata in static InnoDB (flags=0x2401) differs from stats_on_metadata in dynamic InnoDB (fl Ignore the NOSYSVAR flag in addition to ignoring the READONLY flag. Approved by: Marko (https://svn.innodb.com/rb/r/42/) ------------------------------------------------------------------------ r2949 | vasil | 2008-10-31 11:47:56 +0200 (Fri, 31 Oct 2008) | 4 lines branches/zip: White-space cleanup in ChangeLog. ------------------------------------------------------------------------ r2951 | marko | 2008-10-31 14:21:43 +0200 (Fri, 31 Oct 2008) | 4 lines branches/zip: scripts/install_innodb_plugins_win.sql: New script, for installing the InnoDB plugins in Windows. Copied from scripts/install_innodb_plugins.sql. ------------------------------------------------------------------------ r2954 | calvin | 2008-11-04 09:15:26 +0200 (Tue, 04 Nov 2008) | 8 lines branches/zip: ignore the failure when builtin_innobase_plugin is not available. External variable builtin_innobase_plugin is not available when mysqld does not have a builtin InnoDB. The init of the Windows plugin should not fail in this case. Approved by: Marko (on IM) ------------------------------------------------------------------------ r2955 | calvin | 2008-11-04 12:43:14 +0200 (Tue, 04 Nov 2008) | 11 lines branches/zip: windows plugin - fix references to array variables. This problem surfaced when running new test innodb_bug40360.test. Both tx_isolation_names and binlog_format_names are name arrays, and should be defined as wdl_tx_isolation_names and wdl_binlog_format_names, not *wdl_tx_isolation_names and *wdl_binlog_format_names. Another array variable is all_charsets, which is already correctly defined. Approved by: Marko (on IM) ------------------------------------------------------------------------ r2986 | marko | 2008-11-11 09:28:37 +0200 (Tue, 11 Nov 2008) | 11 lines branches/zip: ha_innobase::create(): Remove the dependences on DICT_TF_ZSSIZE_MAX, so that the code can be compiled with a different uncompressed page size by redefining UNIV_PAGE_SIZE_SHIFT in univ.i. Currently, the allowed values are 12, 13, or 14 (4k, 8k, 16k). Make the default compressed page size half the uncompressed page size. The previous default was 8 kilobytes, which is the same when compiling with the default 16k uncompressed page size. rb://50 approved by Pekka Lampio and Sunny Bains. ------------------------------------------------------------------------
2008-11-11 11:21:16 +01:00
#include "handler0vars.h"
branches/zip: Implement the reporting of duplicate key values to MySQL. innobase_rec_to_mysql(): New function, for converting an InnoDB clustered index record to MySQL table->record[0]. TODO: convert integer fields. Currently, integer fields are in big-endian byte order instead of host byte order, and signed integer fields are offset by 0x80000000. innobase_rec_reset(): New function, for resetting table->record[0]. row_merge_build_indexes(): Add the parameter TABLE* table (the MySQL table handle) for reporting duplicate key values. dtuple_from_fields(): New function, to convert an array of dfield_t* to dtuple_t. dtuple_get_n_ext(): New function, to compute the number of externally stored fields. row_merge_dup_t: Structure for counting and reporting duplicate records. row_merge_dup_report(): Function for counting and reporting duplicate records. row_merge_tuple_cmp(), row_merge_tuple_sort(): Replace the ulint* n_dup parameter with row_merge_dup_t* dup. row_merge_buf_sort(): Add the parameter row_merge_dup_t* dup, which is NULL when sorting a non-unique index. row_merge_buf_write(), row_merge_heap_create(), row_merge_read_rec(), row_merge_cmp(), row_merge_read_clustered_index(), row_merge_blocks(), row_merge(), row_merge_sort(): Add const qualifiers. row_merge_read_clustered_index(): Use a common error handling branch err_exit. Invoke row_merge_buf_sort() differently on unique indexes. row_merge_blocks(): note TODO: We could invoke innobase_rec_to_mysql() to report duplicate key values when creating a clustered index.
2007-09-26 13:56:26 +02:00
/*****************************************************************
Copies an InnoDB column to a MySQL field. This function is
adapted from row_sel_field_store_in_mysql_format(). */
static
void
innobase_col_to_mysql(
/*==================*/
const dict_col_t* col, /* in: InnoDB column */
const uchar* data, /* in: InnoDB column data */
ulint len, /* in: length of data, in bytes */
Field* field) /* in/out: MySQL field */
{
uchar* ptr;
uchar* dest = field->ptr;
ulint flen = field->pack_length();
switch (col->mtype) {
case DATA_INT:
ut_ad(len == flen);
/* Convert integer data from Innobase to little-endian
format, sign bit restored to normal */
for (ptr = dest + len; ptr != dest; ) {
*--ptr = *data++;
}
if (!(field->flags & UNSIGNED_FLAG)) {
((byte*) dest)[len - 1] ^= 0x80;
}
break;
case DATA_VARCHAR:
case DATA_VARMYSQL:
case DATA_BINARY:
field->reset();
if (field->type() == MYSQL_TYPE_VARCHAR) {
/* This is a >= 5.0.3 type true VARCHAR. Store the
length of the data to the first byte or the first
two bytes of dest. */
dest = row_mysql_store_true_var_len(
dest, len, flen - field->key_length());
}
/* Copy the actual data */
memcpy(dest, data, len);
break;
case DATA_BLOB:
/* Store a pointer to the BLOB buffer to dest: the BLOB was
already copied to the buffer in row_sel_store_mysql_rec */
row_mysql_store_blob_ref(dest, flen, data, len);
break;
#ifdef UNIV_DEBUG
case DATA_MYSQL:
ut_ad(flen >= len);
ut_ad(col->mbmaxlen >= col->mbminlen);
ut_ad(col->mbmaxlen > col->mbminlen || flen == len);
memcpy(dest, data, len);
break;
default:
case DATA_SYS_CHILD:
case DATA_SYS:
/* These column types should never be shipped to MySQL. */
ut_ad(0);
case DATA_CHAR:
case DATA_FIXBINARY:
case DATA_FLOAT:
case DATA_DOUBLE:
case DATA_DECIMAL:
/* Above are the valid column types for MySQL data. */
ut_ad(flen == len);
#else /* UNIV_DEBUG */
default:
#endif /* UNIV_DEBUG */
memcpy(dest, data, len);
}
}
/*****************************************************************
Copies an InnoDB record to table->record[0]. */
extern "C" UNIV_INTERN
branches/zip: Implement the reporting of duplicate key values to MySQL. innobase_rec_to_mysql(): New function, for converting an InnoDB clustered index record to MySQL table->record[0]. TODO: convert integer fields. Currently, integer fields are in big-endian byte order instead of host byte order, and signed integer fields are offset by 0x80000000. innobase_rec_reset(): New function, for resetting table->record[0]. row_merge_build_indexes(): Add the parameter TABLE* table (the MySQL table handle) for reporting duplicate key values. dtuple_from_fields(): New function, to convert an array of dfield_t* to dtuple_t. dtuple_get_n_ext(): New function, to compute the number of externally stored fields. row_merge_dup_t: Structure for counting and reporting duplicate records. row_merge_dup_report(): Function for counting and reporting duplicate records. row_merge_tuple_cmp(), row_merge_tuple_sort(): Replace the ulint* n_dup parameter with row_merge_dup_t* dup. row_merge_buf_sort(): Add the parameter row_merge_dup_t* dup, which is NULL when sorting a non-unique index. row_merge_buf_write(), row_merge_heap_create(), row_merge_read_rec(), row_merge_cmp(), row_merge_read_clustered_index(), row_merge_blocks(), row_merge(), row_merge_sort(): Add const qualifiers. row_merge_read_clustered_index(): Use a common error handling branch err_exit. Invoke row_merge_buf_sort() differently on unique indexes. row_merge_blocks(): note TODO: We could invoke innobase_rec_to_mysql() to report duplicate key values when creating a clustered index.
2007-09-26 13:56:26 +02:00
void
innobase_rec_to_mysql(
/*==================*/
TABLE* table, /* in/out: MySQL table */
const rec_t* rec, /* in: record */
const dict_index_t* index, /* in: index */
branches/zip: Implement the reporting of duplicate key values to MySQL. innobase_rec_to_mysql(): New function, for converting an InnoDB clustered index record to MySQL table->record[0]. TODO: convert integer fields. Currently, integer fields are in big-endian byte order instead of host byte order, and signed integer fields are offset by 0x80000000. innobase_rec_reset(): New function, for resetting table->record[0]. row_merge_build_indexes(): Add the parameter TABLE* table (the MySQL table handle) for reporting duplicate key values. dtuple_from_fields(): New function, to convert an array of dfield_t* to dtuple_t. dtuple_get_n_ext(): New function, to compute the number of externally stored fields. row_merge_dup_t: Structure for counting and reporting duplicate records. row_merge_dup_report(): Function for counting and reporting duplicate records. row_merge_tuple_cmp(), row_merge_tuple_sort(): Replace the ulint* n_dup parameter with row_merge_dup_t* dup. row_merge_buf_sort(): Add the parameter row_merge_dup_t* dup, which is NULL when sorting a non-unique index. row_merge_buf_write(), row_merge_heap_create(), row_merge_read_rec(), row_merge_cmp(), row_merge_read_clustered_index(), row_merge_blocks(), row_merge(), row_merge_sort(): Add const qualifiers. row_merge_read_clustered_index(): Use a common error handling branch err_exit. Invoke row_merge_buf_sort() differently on unique indexes. row_merge_blocks(): note TODO: We could invoke innobase_rec_to_mysql() to report duplicate key values when creating a clustered index.
2007-09-26 13:56:26 +02:00
const ulint* offsets) /* in: rec_get_offsets(
rec, index, ...) */
{
uint n_fields = table->s->fields;
uint i;
ut_ad(n_fields == dict_table_get_n_user_cols(index->table));
for (i = 0; i < n_fields; i++) {
Field* field = table->field[i];
ulint ipos;
ulint ilen;
const uchar* ifield;
field->reset();
branches/zip: Implement the reporting of duplicate key values to MySQL. innobase_rec_to_mysql(): New function, for converting an InnoDB clustered index record to MySQL table->record[0]. TODO: convert integer fields. Currently, integer fields are in big-endian byte order instead of host byte order, and signed integer fields are offset by 0x80000000. innobase_rec_reset(): New function, for resetting table->record[0]. row_merge_build_indexes(): Add the parameter TABLE* table (the MySQL table handle) for reporting duplicate key values. dtuple_from_fields(): New function, to convert an array of dfield_t* to dtuple_t. dtuple_get_n_ext(): New function, to compute the number of externally stored fields. row_merge_dup_t: Structure for counting and reporting duplicate records. row_merge_dup_report(): Function for counting and reporting duplicate records. row_merge_tuple_cmp(), row_merge_tuple_sort(): Replace the ulint* n_dup parameter with row_merge_dup_t* dup. row_merge_buf_sort(): Add the parameter row_merge_dup_t* dup, which is NULL when sorting a non-unique index. row_merge_buf_write(), row_merge_heap_create(), row_merge_read_rec(), row_merge_cmp(), row_merge_read_clustered_index(), row_merge_blocks(), row_merge(), row_merge_sort(): Add const qualifiers. row_merge_read_clustered_index(): Use a common error handling branch err_exit. Invoke row_merge_buf_sort() differently on unique indexes. row_merge_blocks(): note TODO: We could invoke innobase_rec_to_mysql() to report duplicate key values when creating a clustered index.
2007-09-26 13:56:26 +02:00
ipos = dict_index_get_nth_col_pos(index, i);
if (UNIV_UNLIKELY(ipos == ULINT_UNDEFINED)) {
null_field:
branches/zip: Implement the reporting of duplicate key values to MySQL. innobase_rec_to_mysql(): New function, for converting an InnoDB clustered index record to MySQL table->record[0]. TODO: convert integer fields. Currently, integer fields are in big-endian byte order instead of host byte order, and signed integer fields are offset by 0x80000000. innobase_rec_reset(): New function, for resetting table->record[0]. row_merge_build_indexes(): Add the parameter TABLE* table (the MySQL table handle) for reporting duplicate key values. dtuple_from_fields(): New function, to convert an array of dfield_t* to dtuple_t. dtuple_get_n_ext(): New function, to compute the number of externally stored fields. row_merge_dup_t: Structure for counting and reporting duplicate records. row_merge_dup_report(): Function for counting and reporting duplicate records. row_merge_tuple_cmp(), row_merge_tuple_sort(): Replace the ulint* n_dup parameter with row_merge_dup_t* dup. row_merge_buf_sort(): Add the parameter row_merge_dup_t* dup, which is NULL when sorting a non-unique index. row_merge_buf_write(), row_merge_heap_create(), row_merge_read_rec(), row_merge_cmp(), row_merge_read_clustered_index(), row_merge_blocks(), row_merge(), row_merge_sort(): Add const qualifiers. row_merge_read_clustered_index(): Use a common error handling branch err_exit. Invoke row_merge_buf_sort() differently on unique indexes. row_merge_blocks(): note TODO: We could invoke innobase_rec_to_mysql() to report duplicate key values when creating a clustered index.
2007-09-26 13:56:26 +02:00
field->set_null();
continue;
}
ifield = rec_get_nth_field(rec, offsets, ipos, &ilen);
/* Assign the NULL flag */
if (ilen == UNIV_SQL_NULL) {
ut_ad(field->real_maybe_null());
goto null_field;
branches/zip: Implement the reporting of duplicate key values to MySQL. innobase_rec_to_mysql(): New function, for converting an InnoDB clustered index record to MySQL table->record[0]. TODO: convert integer fields. Currently, integer fields are in big-endian byte order instead of host byte order, and signed integer fields are offset by 0x80000000. innobase_rec_reset(): New function, for resetting table->record[0]. row_merge_build_indexes(): Add the parameter TABLE* table (the MySQL table handle) for reporting duplicate key values. dtuple_from_fields(): New function, to convert an array of dfield_t* to dtuple_t. dtuple_get_n_ext(): New function, to compute the number of externally stored fields. row_merge_dup_t: Structure for counting and reporting duplicate records. row_merge_dup_report(): Function for counting and reporting duplicate records. row_merge_tuple_cmp(), row_merge_tuple_sort(): Replace the ulint* n_dup parameter with row_merge_dup_t* dup. row_merge_buf_sort(): Add the parameter row_merge_dup_t* dup, which is NULL when sorting a non-unique index. row_merge_buf_write(), row_merge_heap_create(), row_merge_read_rec(), row_merge_cmp(), row_merge_read_clustered_index(), row_merge_blocks(), row_merge(), row_merge_sort(): Add const qualifiers. row_merge_read_clustered_index(): Use a common error handling branch err_exit. Invoke row_merge_buf_sort() differently on unique indexes. row_merge_blocks(): note TODO: We could invoke innobase_rec_to_mysql() to report duplicate key values when creating a clustered index.
2007-09-26 13:56:26 +02:00
}
field->set_notnull();
innobase_col_to_mysql(
dict_field_get_col(
dict_index_get_nth_field(index, ipos)),
ifield, ilen, field);
branches/zip: Implement the reporting of duplicate key values to MySQL. innobase_rec_to_mysql(): New function, for converting an InnoDB clustered index record to MySQL table->record[0]. TODO: convert integer fields. Currently, integer fields are in big-endian byte order instead of host byte order, and signed integer fields are offset by 0x80000000. innobase_rec_reset(): New function, for resetting table->record[0]. row_merge_build_indexes(): Add the parameter TABLE* table (the MySQL table handle) for reporting duplicate key values. dtuple_from_fields(): New function, to convert an array of dfield_t* to dtuple_t. dtuple_get_n_ext(): New function, to compute the number of externally stored fields. row_merge_dup_t: Structure for counting and reporting duplicate records. row_merge_dup_report(): Function for counting and reporting duplicate records. row_merge_tuple_cmp(), row_merge_tuple_sort(): Replace the ulint* n_dup parameter with row_merge_dup_t* dup. row_merge_buf_sort(): Add the parameter row_merge_dup_t* dup, which is NULL when sorting a non-unique index. row_merge_buf_write(), row_merge_heap_create(), row_merge_read_rec(), row_merge_cmp(), row_merge_read_clustered_index(), row_merge_blocks(), row_merge(), row_merge_sort(): Add const qualifiers. row_merge_read_clustered_index(): Use a common error handling branch err_exit. Invoke row_merge_buf_sort() differently on unique indexes. row_merge_blocks(): note TODO: We could invoke innobase_rec_to_mysql() to report duplicate key values when creating a clustered index.
2007-09-26 13:56:26 +02:00
}
}
/*****************************************************************
Resets table->record[0]. */
extern "C" UNIV_INTERN
branches/zip: Implement the reporting of duplicate key values to MySQL. innobase_rec_to_mysql(): New function, for converting an InnoDB clustered index record to MySQL table->record[0]. TODO: convert integer fields. Currently, integer fields are in big-endian byte order instead of host byte order, and signed integer fields are offset by 0x80000000. innobase_rec_reset(): New function, for resetting table->record[0]. row_merge_build_indexes(): Add the parameter TABLE* table (the MySQL table handle) for reporting duplicate key values. dtuple_from_fields(): New function, to convert an array of dfield_t* to dtuple_t. dtuple_get_n_ext(): New function, to compute the number of externally stored fields. row_merge_dup_t: Structure for counting and reporting duplicate records. row_merge_dup_report(): Function for counting and reporting duplicate records. row_merge_tuple_cmp(), row_merge_tuple_sort(): Replace the ulint* n_dup parameter with row_merge_dup_t* dup. row_merge_buf_sort(): Add the parameter row_merge_dup_t* dup, which is NULL when sorting a non-unique index. row_merge_buf_write(), row_merge_heap_create(), row_merge_read_rec(), row_merge_cmp(), row_merge_read_clustered_index(), row_merge_blocks(), row_merge(), row_merge_sort(): Add const qualifiers. row_merge_read_clustered_index(): Use a common error handling branch err_exit. Invoke row_merge_buf_sort() differently on unique indexes. row_merge_blocks(): note TODO: We could invoke innobase_rec_to_mysql() to report duplicate key values when creating a clustered index.
2007-09-26 13:56:26 +02:00
void
innobase_rec_reset(
/*===============*/
TABLE* table) /* in/out: MySQL table */
{
uint n_fields = table->s->fields;
uint i;
for (i = 0; i < n_fields; i++) {
table->field[i]->set_default();
}
}
/**********************************************************************
Removes the filename encoding of a database and table name. */
static
void
innobase_convert_tablename(
/*=======================*/
char* s) /* in: identifier; out: decoded identifier */
{
uint errors;
char* slash = strchr(s, '/');
if (slash) {
char* t;
/* Temporarily replace the '/' with NUL. */
*slash = 0;
/* Convert the database name. */
strconvert(&my_charset_filename, s, system_charset_info,
s, slash - s + 1, &errors);
t = s + strlen(s);
ut_ad(slash >= t);
/* Append a '.' after the database name. */
*t++ = '.';
slash++;
/* Convert the table name. */
strconvert(&my_charset_filename, slash, system_charset_info,
t, slash - t + strlen(slash), &errors);
} else {
strconvert(&my_charset_filename, s,
system_charset_info, s, strlen(s), &errors);
}
}
/***********************************************************************
This function checks that index keys are sensible. */
static
int
innobase_check_index_keys(
/*======================*/
/* out: 0 or error number */
const KEY* key_info, /* in: Indexes to be created */
ulint num_of_keys) /* in: Number of indexes to
be created */
{
ulint key_num;
ut_ad(key_info);
ut_ad(num_of_keys);
for (key_num = 0; key_num < num_of_keys; key_num++) {
const KEY& key = key_info[key_num];
/* Check that the same index name does not appear
twice in indexes to be created. */
for (ulint i = 0; i < key_num; i++) {
const KEY& key2 = key_info[i];
if (0 == strcmp(key.name, key2.name)) {
sql_print_error("InnoDB: key name `%s` appears"
" twice in CREATE INDEX\n",
key.name);
return(ER_WRONG_NAME_FOR_INDEX);
}
}
/* Check that MySQL does not try to create a column
prefix index field on an inappropriate data type and
that the same colum does not appear twice in the index. */
for (ulint i = 0; i < key.key_parts; i++) {
const KEY_PART_INFO& key_part1
= key.key_part[i];
const Field* field
= key_part1.field;
ibool is_unsigned;
switch (get_innobase_type_from_mysql_type(
&is_unsigned, field)) {
default:
break;
case DATA_INT:
case DATA_FLOAT:
case DATA_DOUBLE:
case DATA_DECIMAL:
if (field->type() == MYSQL_TYPE_VARCHAR) {
if (key_part1.length
>= field->pack_length()
- ((Field_varstring*) field)
->length_bytes) {
break;
}
} else {
if (key_part1.length
>= field->pack_length()) {
break;
}
}
sql_print_error("InnoDB: MySQL is trying to"
" create a column prefix"
" index field on an"
" inappropriate data type."
" column `%s`,"
" index `%s`.\n",
field->field_name,
key.name);
return(ER_WRONG_KEY_COLUMN);
}
for (ulint j = 0; j < i; j++) {
const KEY_PART_INFO& key_part2
= key.key_part[j];
if (strcmp(key_part1.field->field_name,
key_part2.field->field_name)) {
continue;
}
sql_print_error("InnoDB: column `%s`"
" is not allowed to occur"
" twice in index `%s`.\n",
key_part1.field->field_name,
key.name);
return(ER_WRONG_KEY_COLUMN);
}
}
}
return(0);
}
/***********************************************************************
Create index field definition for key part */
static
void
innobase_create_index_field_def(
/*============================*/
KEY_PART_INFO* key_part, /* in: MySQL key definition */
mem_heap_t* heap, /* in: memory heap */
merge_index_field_t* index_field) /* out: index field
definition for key_part */
{
Field* field;
ibool is_unsigned;
ulint col_type;
DBUG_ENTER("innobase_create_index_field_def");
ut_ad(key_part);
ut_ad(index_field);
field = key_part->field;
ut_a(field);
col_type = get_innobase_type_from_mysql_type(&is_unsigned, field);
if (DATA_BLOB == col_type
|| (key_part->length < field->pack_length()
&& field->type() != MYSQL_TYPE_VARCHAR)
|| (field->type() == MYSQL_TYPE_VARCHAR
&& key_part->length < field->pack_length()
- ((Field_varstring*)field)->length_bytes)) {
index_field->prefix_len = key_part->length;
} else {
index_field->prefix_len = 0;
}
index_field->field_name = mem_heap_strdup(heap, field->field_name);
DBUG_VOID_RETURN;
}
/***********************************************************************
Create index definition for key */
static
void
innobase_create_index_def(
/*======================*/
KEY* key, /* in: key definition */
bool new_primary, /* in: TRUE=generating
a new primary key
on the table */
bool key_primary, /* in: TRUE if this key
is a primary key */
merge_index_def_t* index, /* out: index definition */
mem_heap_t* heap) /* in: heap where memory
is allocated */
{
ulint i;
ulint len;
ulint n_fields = key->key_parts;
char* index_name;
DBUG_ENTER("innobase_create_index_def");
index->fields = (merge_index_field_t*) mem_heap_alloc(
heap, n_fields * sizeof *index->fields);
index->ind_type = 0;
index->n_fields = n_fields;
len = strlen(key->name) + 1;
index->name = index_name = (char*) mem_heap_alloc(heap,
len + !new_primary);
if (UNIV_LIKELY(!new_primary)) {
*index_name++ = TEMP_INDEX_PREFIX;
}
memcpy(index_name, key->name, len);
if (key->flags & HA_NOSAME) {
index->ind_type |= DICT_UNIQUE;
}
if (key_primary) {
index->ind_type |= DICT_CLUSTERED;
}
for (i = 0; i < n_fields; i++) {
innobase_create_index_field_def(&key->key_part[i], heap,
&index->fields[i]);
}
DBUG_VOID_RETURN;
}
/***********************************************************************
Copy index field definition */
static
void
innobase_copy_index_field_def(
/*==========================*/
const dict_field_t* field, /* in: definition to copy */
merge_index_field_t* index_field) /* out: copied definition */
{
DBUG_ENTER("innobase_copy_index_field_def");
DBUG_ASSERT(field != NULL);
DBUG_ASSERT(index_field != NULL);
index_field->field_name = field->name;
index_field->prefix_len = field->prefix_len;
DBUG_VOID_RETURN;
}
/***********************************************************************
Copy index definition for the index */
static
void
innobase_copy_index_def(
/*====================*/
const dict_index_t* index, /* in: index definition to copy */
merge_index_def_t* new_index,/* out: Index definition */
mem_heap_t* heap) /* in: heap where allocated */
{
ulint n_fields;
ulint i;
DBUG_ENTER("innobase_copy_index_def");
/* Note that we take only those fields that user defined to be
in the index. In the internal representation more colums were
added and those colums are not copied .*/
n_fields = index->n_user_defined_cols;
new_index->fields = (merge_index_field_t*) mem_heap_alloc(
heap, n_fields * sizeof *new_index->fields);
/* When adding a PRIMARY KEY, we may convert a previous
clustered index to a secondary index (UNIQUE NOT NULL). */
new_index->ind_type = index->type & ~DICT_CLUSTERED;
new_index->n_fields = n_fields;
new_index->name = index->name;
for (i = 0; i < n_fields; i++) {
innobase_copy_index_field_def(&index->fields[i],
&new_index->fields[i]);
}
DBUG_VOID_RETURN;
}
/***********************************************************************
Create an index table where indexes are ordered as follows:
IF a new primary key is defined for the table THEN
1) New primary key
2) Original secondary indexes
3) New secondary indexes
ELSE
1) All new indexes in the order they arrive from MySQL
ENDIF
*/
static
merge_index_def_t*
innobase_create_key_def(
/*====================*/
/* out: key definitions or NULL */
trx_t* trx, /* in: trx */
const dict_table_t*table, /* in: table definition */
mem_heap_t* heap, /* in: heap where space for key
definitions are allocated */
KEY* key_info, /* in: Indexes to be created */
ulint& n_keys) /* in/out: Number of indexes to
be created */
{
ulint i = 0;
merge_index_def_t* indexdef;
merge_index_def_t* indexdefs;
bool new_primary;
DBUG_ENTER("innobase_create_key_def");
indexdef = indexdefs = (merge_index_def_t*)
mem_heap_alloc(heap, sizeof *indexdef
* (n_keys + UT_LIST_GET_LEN(table->indexes)));
/* If there is a primary key, it is always the first index
defined for the table. */
new_primary = !my_strcasecmp(system_charset_info,
key_info->name, "PRIMARY");
/* If there is a UNIQUE INDEX consisting entirely of NOT NULL
columns, MySQL will treat it as a PRIMARY KEY unless the
table already has one. */
if (!new_primary && (key_info->flags & HA_NOSAME)
&& row_table_got_default_clust_index(table)) {
uint key_part = key_info->key_parts;
new_primary = TRUE;
while (key_part--) {
if (key_info->key_part[key_part].key_type
& FIELDFLAG_MAYBE_NULL) {
new_primary = FALSE;
break;
}
}
}
if (new_primary) {
const dict_index_t* index;
/* Create the PRIMARY key index definition */
innobase_create_index_def(&key_info[i++], TRUE, TRUE,
indexdef++, heap);
row_mysql_lock_data_dictionary(trx);
index = dict_table_get_first_index(table);
/* Copy the index definitions of the old table. Skip
the old clustered index if it is a generated clustered
index or a PRIMARY KEY. If the clustered index is a
UNIQUE INDEX, it must be converted to a secondary index. */
if (dict_index_get_nth_col(index, 0)->mtype == DATA_SYS
|| !my_strcasecmp(system_charset_info,
index->name, "PRIMARY")) {
index = dict_table_get_next_index(index);
}
while (index) {
innobase_copy_index_def(index, indexdef++, heap);
index = dict_table_get_next_index(index);
}
row_mysql_unlock_data_dictionary(trx);
}
/* Create definitions for added secondary indexes. */
while (i < n_keys) {
innobase_create_index_def(&key_info[i++], new_primary, FALSE,
indexdef++, heap);
}
n_keys = indexdef - indexdefs;
DBUG_RETURN(indexdefs);
}
/***********************************************************************
Create a temporary tablename using query id, thread id, and id */
static
char*
innobase_create_temporary_tablename(
/*================================*/
/* out: temporary tablename */
mem_heap_t* heap, /* in: memory heap */
char id, /* in: identifier [0-9a-zA-Z] */
const char* table_name) /* in: table name */
{
char* name;
ulint len;
static const char suffix[] = "@0023 "; /* "# " */
len = strlen(table_name);
name = (char*) mem_heap_alloc(heap, len + sizeof suffix);
memcpy(name, table_name, len);
memcpy(name + len, suffix, sizeof suffix);
name[len + (sizeof suffix - 2)] = id;
return(name);
}
/***********************************************************************
Create indexes. */
UNIV_INTERN
int
ha_innobase::add_index(
/*===================*/
/* out: 0 or error number */
TABLE* table, /* in: Table where indexes are created */
KEY* key_info, /* in: Indexes to be created */
uint num_of_keys) /* in: Number of indexes to be created */
{
dict_index_t** index; /* Index to be created */
dict_table_t* innodb_table; /* InnoDB table in dictionary */
dict_table_t* indexed_table; /* Table where indexes are created */
merge_index_def_t* index_defs; /* Index definitions */
mem_heap_t* heap; /* Heap for index definitions */
trx_t* trx; /* Transaction */
ulint num_of_idx;
ulint num_created = 0;
ibool dict_locked = FALSE;
ulint new_primary;
ulint error;
DBUG_ENTER("ha_innobase::add_index");
ut_a(table);
ut_a(key_info);
ut_a(num_of_keys);
if (srv_created_new_raw || srv_force_recovery) {
DBUG_RETURN(HA_ERR_WRONG_COMMAND);
}
update_thd();
heap = mem_heap_create(1024);
/* In case MySQL calls this in the middle of a SELECT query, release
possible adaptive hash latch to avoid deadlocks of threads. */
trx_search_latch_release_if_reserved(prebuilt->trx);
/* Create a background transaction for the operations on
the data dictionary tables. */
branches/innodb+: Merge revisions 3602:3931 from branches/zip: ------------------------------------------------------------------------ r3607 | marko | 2008-12-30 22:33:31 +0200 (Tue, 30 Dec 2008) | 20 lines branches/zip: Remove the dependency on the MySQL HASH table implementation. Use the InnoDB hash table for keeping track of INNOBASE_SHARE objects. struct st_innobase_share: Make table_name const uchar*. Add the member table_name_hash. innobase_open_tables: Change the type from HASH to hash_table_t*. innobase_get_key(): Remove. innobase_fold_name(): New function, for computing the fold value for the InnoDB hash table. get_share(), free_share(): Use the InnoDB hash functions. innobase_end(): Free innobase_open_tables before shutting down InnoDB. Shutting down InnoDB will invalidate all memory allocated via InnoDB. rb://65 approved by Heikki Tuuri. This addresses Issue #104. ------------------------------------------------------------------------ r3608 | marko | 2008-12-30 22:45:04 +0200 (Tue, 30 Dec 2008) | 22 lines branches/zip: When setting the PAGE_LEVEL of a compressed B-tree page from or to 0, compress the page at the same time. This is necessary, because the column information stored on the compressed page will differ between leaf and non-leaf pages. Leaf pages are identified by PAGE_LEVEL=0. This bug was reported as Issue #150. Document the similarity between btr_page_create() and btr_page_empty(). Make the function signature of btr_page_empty() identical with btr_page_create(). (This will add the parameter "level".) btr_root_raise_and_insert(): Replace some code with a call to btr_page_empty(). btr_attach_half_pages(): Assert that the page level has already been set on both block and new_block. Do not set it again. btr_discard_only_page_on_level(): Document that this function is probably never called. Make it work on any height tree. (Tested on 2-high tree by disabling btr_lift_page_up().) rb://68 ------------------------------------------------------------------------ r3612 | marko | 2009-01-02 11:02:44 +0200 (Fri, 02 Jan 2009) | 14 lines branches/zip: Merge c2998 from branches/6.0, so that the same InnoDB Plugin source tree will work both under 5.1 and 6.0. Do not add the test case innodb_ctype_ldml.test, because it would not work under MySQL 5.1. Refuse to create tables whose columns contain collation IDs above 255. This removes an assertion failure that was introduced in WL#4164 (Two-byte collation IDs). create_table_def(): Do not fail an assertion if a column contains a charset-collation ID greater than 256. Instead, issue an error and refuse to create the table. The original change (branches/6.0 r2998) was rb://51 approved by Calvin Sun. ------------------------------------------------------------------------ r3613 | inaam | 2009-01-02 15:10:50 +0200 (Fri, 02 Jan 2009) | 6 lines branches/zip: Implement the parameter innodb_use_sys_malloc (false by default), for disabling InnoDB's internal memory allocator and using system malloc/free instead. rb://62 approved by Marko ------------------------------------------------------------------------ r3614 | marko | 2009-01-02 15:55:12 +0200 (Fri, 02 Jan 2009) | 1 line branches/zip: ChangeLog: Document r3608 and r3613. ------------------------------------------------------------------------ r3615 | marko | 2009-01-02 15:57:51 +0200 (Fri, 02 Jan 2009) | 1 line branches/zip: ChangeLog: Clarify the impact of r3608. ------------------------------------------------------------------------ r3616 | marko | 2009-01-03 00:23:30 +0200 (Sat, 03 Jan 2009) | 1 line branches/zip: srv_suspend_mysql_thread(): Add some clarifying comments. ------------------------------------------------------------------------ r3618 | marko | 2009-01-05 12:54:53 +0200 (Mon, 05 Jan 2009) | 15 lines branches/zip: Merge revisions 3598:3601 from branches/5.1: ------------------------------------------------------------------------ r3601 | marko | 2008-12-22 16:05:19 +0200 (Mon, 22 Dec 2008) | 9 lines branches/5.1: Make SET SESSION TRANSACTION ISOLATION LEVEL READ COMMITTED a true replacement of SET GLOBAL INNODB_LOCKS_UNSAFE_FOR_BINLOG=1. This fixes an error that was introduced in r370, causing semi-consistent read not to not unlock rows in READ COMMITTED mode. (Bug #41671, Issue #146) rb://67 approved by Heikki Tuuri ------------------------------------------------------------------------ ------------------------------------------------------------------------ r3623 | vasil | 2009-01-06 09:56:32 +0200 (Tue, 06 Jan 2009) | 7 lines branches/zip: Add patch to fix the failing main.variables mysql-test. It started failing after the variable innodb_use_sys_malloc was added because it matches '%alloc%' and the test is badly written and expects that no new variables like that will ever be added. ------------------------------------------------------------------------ r3795 | marko | 2009-01-07 16:17:47 +0200 (Wed, 07 Jan 2009) | 7 lines branches/zip: row_merge_tuple_cmp(): Do not report a duplicate key value if any of the fields are NULL. While the tuples are equal in the sorting order, SQL NULL is defined to be logically inequal to anything else. (Bug #41904) rb://70 approved by Heikki Tuuri ------------------------------------------------------------------------ r3796 | marko | 2009-01-07 16:19:32 +0200 (Wed, 07 Jan 2009) | 1 line branches/zip: Add the tests that were forgotten from r3795. ------------------------------------------------------------------------ r3797 | marko | 2009-01-07 16:22:18 +0200 (Wed, 07 Jan 2009) | 22 lines branches/zip: Do not call trx_allocate_for_mysql() directly, but use helper functions that initialize some members of the transaction struct. (Bug #41680) innobase_trx_init(): New function: initialize some fields of a transaction struct from a MySQL THD object. innobase_trx_allocate(): New function: allocate and initialize a transaction struct. check_trx_exists(): Use the above two functions. ha_innobase::delete_table(), ha_innobase::rename_table(), ha_innobase::add_index(), ha_innobase::final_drop_index(): Use innobase_trx_allocate(). innobase_drop_database(): In the Windows plugin, initialize the trx_t specially, because the THD is not available. Otherwise, use innobase_trx_allocate(). rb://69 accepted by Heikki Tuuri ------------------------------------------------------------------------ r3798 | marko | 2009-01-07 16:42:42 +0200 (Wed, 07 Jan 2009) | 8 lines branches/zip: row_merge_drop_temp_indexes(): Do not lock the rows of SYS_INDEXES when looking for partially created indexes. Use the transaction isolation level READ UNCOMMITTED to avoid interfering with locks held by incomplete transactions that will be rolled back in a subsequent step in the recovery. (Issue #152) Approved by Heikki Tuuri ------------------------------------------------------------------------ r3852 | vasil | 2009-01-08 22:10:10 +0200 (Thu, 08 Jan 2009) | 4 lines branches/zip: Add ChangeLog entries for r3795 r3796 r3797 r3798. ------------------------------------------------------------------------ r3866 | marko | 2009-01-09 15:09:51 +0200 (Fri, 09 Jan 2009) | 2 lines branches/zip: buf_flush_try_page(): Move some common code from each switch case before the switch block. ------------------------------------------------------------------------ r3867 | marko | 2009-01-09 15:13:14 +0200 (Fri, 09 Jan 2009) | 2 lines branches/zip: buf_flush_try_page(): Introduce the variable is_compressed for caching the result of buf_page_get_state(bpage) == BUF_BLOCK_FILE_PAGE. ------------------------------------------------------------------------ r3868 | marko | 2009-01-09 15:40:11 +0200 (Fri, 09 Jan 2009) | 4 lines branches/zip: buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove unused code. Change the parameter to buf_block_t* block and assert that block->state == BUF_BLOCK_FILE_PAGE. This is part of Issue #155. ------------------------------------------------------------------------ r3873 | marko | 2009-01-09 22:27:40 +0200 (Fri, 09 Jan 2009) | 17 lines branches/zip: Some non-functional changes related to Issue #155. buf_page_struct: Note that space and offset are also protected by buf_pool_mutex. They are only assigned to by buf_block_set_file_page(). Thus, it suffices for buf_flush_batch() to hold just buf_pool_mutex when checking these fields. buf_flush_try_page(): Rename "locked" to "is_s_latched", per Heikki's request. buf_flush_batch(): Move the common statement mutex_exit(block_mutex) from all if-else if-else branches before the if block. Remove the redundant test (buf_pool->init_flush[flush_type] == FALSE) that was apparently copied from buf_flush_write_complete(). buf_flush_write_block_low(): Note why it is safe not to hold buf_pool_mutex or block_mutex. Enumerate the assumptions in debug assertions. ------------------------------------------------------------------------ r3874 | marko | 2009-01-09 23:09:06 +0200 (Fri, 09 Jan 2009) | 4 lines branches/zip: Add comments related to Issue #155. buf_flush_try_page(): Note why it is safe to access bpage without holding buf_pool_mutex or block_mutex. ------------------------------------------------------------------------ r3875 | marko | 2009-01-09 23:15:12 +0200 (Fri, 09 Jan 2009) | 11 lines branches/zip: Non-functional change: Tighten debug assertions and remove dead code. buf_flush_ready_for_flush(), buf_flush_try_page(): Assert that flush_type is one of BUF_FLUSH_LRU or BUF_FLUSH_LIST. The flush_type comes from buf_flush_batch(), which already asserts this. The assertion holds for all calls in the source code. buf_flush_try_page(): Remove the dead case BUF_FLUSH_SINGLE_PAGE of switch (flush_type). ------------------------------------------------------------------------ r3879 | marko | 2009-01-12 12:46:44 +0200 (Mon, 12 Jan 2009) | 14 lines branches/zip: Simplify the flushing of dirty pages from the buffer pool. buf_flush_try_page(): Rename to buf_flush_page(), and change the return type to void. Replace the parameters space, offset with bpage, and remove the second page hash lookup. Note and assert that both buf_pool_mutex and block_mutex must now be held upon entering the function. They will still be released by this function. buf_flush_try_neighbors(): Replace buf_flush_try_page() with buf_flush_page(). Make the logic easier to follow by not negating the precondition of buf_flush_page(). rb://73 approved by Sunny Bains. This is related to Issue #157. ------------------------------------------------------------------------ r3880 | marko | 2009-01-12 13:24:37 +0200 (Mon, 12 Jan 2009) | 2 lines branches/zip: buf_flush_page(): Fix a comment that should have been fixed in r3879. Spotted by Sunny. ------------------------------------------------------------------------ r3881 | marko | 2009-01-12 14:25:22 +0200 (Mon, 12 Jan 2009) | 2 lines branches/zip: buf_page_get_newest_modification(): Use the block mutex instead of the buffer pool mutex. This is related to Issue #157. ------------------------------------------------------------------------ r3882 | marko | 2009-01-12 14:40:08 +0200 (Mon, 12 Jan 2009) | 3 lines branches/zip: struct mtr_struct: Remove the unused field magic_n unless UNIV_DEBUG is defined. mtr->magic_n is only assigned to and checked in UNIV_DEBUG builds. ------------------------------------------------------------------------ r3883 | marko | 2009-01-12 14:48:59 +0200 (Mon, 12 Jan 2009) | 1 line branches/zip: Non-functional change: Use ut_d when assigning to mtr->state. ------------------------------------------------------------------------ r3884 | marko | 2009-01-12 18:56:11 +0200 (Mon, 12 Jan 2009) | 16 lines branches/zip: Non-functional change: Add some debug assertions and comments. buf_page_t: Note that the LRU fields are protected by buf_pool_mutex only, not block->mutex or buf_pool_zip_mutex. buf_page_get_freed_page_clock(): Note that this is sometimes invoked without mutex protection. buf_pool_get_oldest_modification(): Note that the result may be out of date. buf_page_get_LRU_position(), buf_page_is_old(): Assert that the buffer pool mutex is being held. buf_page_release(): Assert that dirty blocks are in the flush list. ------------------------------------------------------------------------ r3896 | marko | 2009-01-13 09:30:26 +0200 (Tue, 13 Jan 2009) | 2 lines branches/zip: buf_flush_try_neighbors(): Fix a bug that was introduced in r3879 (rb://73). ------------------------------------------------------------------------ r3900 | marko | 2009-01-13 10:32:24 +0200 (Tue, 13 Jan 2009) | 1 line branches/zip: Fix some comments to say buf_pool_mutex. ------------------------------------------------------------------------ r3907 | marko | 2009-01-13 11:54:01 +0200 (Tue, 13 Jan 2009) | 3 lines branches/zip: row_merge_create_temporary_table(): On error, row_create_table_for_mysql() already frees new_table. Do not attempt to free it again. ------------------------------------------------------------------------ r3908 | marko | 2009-01-13 12:34:32 +0200 (Tue, 13 Jan 2009) | 1 line branches/zip: Enable HASH_ASSERT_OWNED independently of UNIV_SYNC_DEBUG. ------------------------------------------------------------------------ r3914 | marko | 2009-01-13 21:46:22 +0200 (Tue, 13 Jan 2009) | 37 lines branches/zip: In hash table lookups, assert that the traversed items satisfy some conditions when UNIV_DEBUG is defined. HASH_SEARCH(): New parameter: ASSERTION. All users will pass an appropriate ut_ad() or nothing. dict_table_add_to_columns(): Assert that the table being added to the data dictionary cache is not already being pointed to by the name_hash and id_hash tables. HASH_SEARCH_ALL(): New macro, for use in dict_table_add_to_columns(). dict_mem_table_free(): Set ut_d(table->cached = FALSE), so that we can check ut_ad(table->cached) when traversing the hash tables, as in HASH_SEARCH(name_hash, dict_sys->table_hash, ...) and HASH_SEARCH(id_hash, dict_sys->table_id_hash, ...). dict_table_get_low(), dict_table_get_on_id_low(): Assert ut_ad(!table || table->cached). fil_space_get_by_id(): Check ut_ad(space->magic_n == FIL_SPACE_MAGIC_N) in HASH_SEARCH(hash, fil_system->spaces, ...). fil_space_get_by_name(): Check ut_ad(space->magic_n == FIL_SPACE_MAGIC_N) in HASH_SEARCH(name_hash, fil_system->name_hash, ...). buf_buddy_block_free(): Check that the blocks are in valid state in HASH_SEARCH(hash, buf_pool->zip_hash, ...). buf_page_hash_get(): Check that the blocks are in valid state in HASH_SEARCH(hash, buf_pool->page_hash, ...). get_share(), free_share(): Check ut_ad(share->use_count > 0) in HASH_SEARCH(table_name_hash, innobase_open_tables, ...). This was posted as rb://75 for tracking down errors similar to Issue #153. ------------------------------------------------------------------------ r3931 | marko | 2009-01-14 16:06:22 +0200 (Wed, 14 Jan 2009) | 26 lines branches/zip: Merge revisions 3601:3930 from branches/5.1: ------------------------------------------------------------------------ r3911 | sunny | 2009-01-13 14:15:24 +0200 (Tue, 13 Jan 2009) | 13 lines branches/5.1: Fix Bug#38187 Error 153 when creating savepoints InnoDB previously treated savepoints as a stack e.g., SAVEPOINT a; SAVEPOINT b; SAVEPOINT c; SAVEPOINT b; <- This would delete b and c. This fix changes the behavior to: SAVEPOINT a; SAVEPOINT b; SAVEPOINT c; SAVEPOINT b; <- Does not delete savepoint c ------------------------------------------------------------------------ r3930 | marko | 2009-01-14 15:51:30 +0200 (Wed, 14 Jan 2009) | 4 lines branches/5.1: dict_load_table(): If dict_load_indexes() fails, invoke dict_table_remove_from_cache() instead of dict_mem_table_free(), so that the data dictionary will not point to freed data. (Bug #42075, Issue #153, rb://76 approved by Heikki Tuuri) ------------------------------------------------------------------------ ------------------------------------------------------------------------
2009-01-14 15:25:45 +01:00
trx = innobase_trx_allocate(user_thd);
trx_start_if_not_started(trx);
innodb_table = indexed_table
= dict_table_get(prebuilt->table->name, FALSE);
/* Check that index keys are sensible */
error = innobase_check_index_keys(key_info, num_of_keys);
if (UNIV_UNLIKELY(error)) {
err_exit:
mem_heap_free(heap);
trx_general_rollback_for_mysql(trx, FALSE, NULL);
trx_free_for_mysql(trx);
trx_commit_for_mysql(prebuilt->trx);
DBUG_RETURN(error);
}
/* Create table containing all indexes to be built in this
alter table add index so that they are in the correct order
in the table. */
num_of_idx = num_of_keys;
index_defs = innobase_create_key_def(
trx, innodb_table, heap, key_info, num_of_idx);
new_primary = DICT_CLUSTERED & index_defs[0].ind_type;
/* Allocate memory for dictionary index definitions */
index = (dict_index_t**) mem_heap_alloc(
heap, num_of_idx * sizeof *index);
/* Flag this transaction as a dictionary operation, so that
the data dictionary will be locked in crash recovery. */
trx_set_dict_operation(trx, TRX_DICT_OP_INDEX);
branches/zip: Fast index creation: Remove the ROW_PREBUILT_OBSOLETE nonsense. Active transactions must not switch table or index definitions on the fly, for several reasons, including the following: * copied indexes do not carry any history or locking information; that is, rollbacks, read views, and record locking would be broken * huge potential for race conditions, inconsistent reads and writes, loss of data, and corruption Instead of trying to track down if the table was changed during a transaction, acquire appropriate locks that protect the creation and dropping of indexes. innodb-index.test: Test the locking of CREATE INDEX and DROP INDEX. Test that consistent reads work across dropped indexes. lock_rec_insert_check_and_lock(): Relax the lock_table_has() assertion. When inserting a record into an index, the table must be at least IX-locked. However, when an index is being created, an IS-lock on the table is sufficient. row_merge_lock_table(): Add the parameter enum lock_mode mode, which must be LOCK_X or LOCK_S. row_merge_drop_table(): Assert that n_mysql_handles_opened == 0. Unconditionally drop the table. ha_innobase::add_index(): Acquire an X or S lock on the table, as appropriate. After acquiring an X lock, assert that n_mysql_handles_opened == 1. Remove the comments about dropping tables in the background. ha_innobase::final_drop_index(): Acquire an X lock on the table. dict_table_t: Remove version_number, to_be_dropped, and prebuilts. ins_node_t: Remove table_version_number. enum lock_mode: Move the definition from lock0lock.h to lock0types.h. ROW_PREBUILT_OBSOLETE, row_update_prebuilt(), row_prebuilt_table_obsolete(): Remove. row_prebuilt_t: Remove the declaration from row0types.h. row_drop_table_for_mysql_no_commit(): Always print a warning if a table was added to the background drop queue.
2007-12-17 16:49:59 +01:00
/* Acquire a lock on the table before creating any indexes. */
error = row_merge_lock_table(prebuilt->trx, innodb_table,
branches/zip: Fast index creation: Remove the ROW_PREBUILT_OBSOLETE nonsense. Active transactions must not switch table or index definitions on the fly, for several reasons, including the following: * copied indexes do not carry any history or locking information; that is, rollbacks, read views, and record locking would be broken * huge potential for race conditions, inconsistent reads and writes, loss of data, and corruption Instead of trying to track down if the table was changed during a transaction, acquire appropriate locks that protect the creation and dropping of indexes. innodb-index.test: Test the locking of CREATE INDEX and DROP INDEX. Test that consistent reads work across dropped indexes. lock_rec_insert_check_and_lock(): Relax the lock_table_has() assertion. When inserting a record into an index, the table must be at least IX-locked. However, when an index is being created, an IS-lock on the table is sufficient. row_merge_lock_table(): Add the parameter enum lock_mode mode, which must be LOCK_X or LOCK_S. row_merge_drop_table(): Assert that n_mysql_handles_opened == 0. Unconditionally drop the table. ha_innobase::add_index(): Acquire an X or S lock on the table, as appropriate. After acquiring an X lock, assert that n_mysql_handles_opened == 1. Remove the comments about dropping tables in the background. ha_innobase::final_drop_index(): Acquire an X lock on the table. dict_table_t: Remove version_number, to_be_dropped, and prebuilts. ins_node_t: Remove table_version_number. enum lock_mode: Move the definition from lock0lock.h to lock0types.h. ROW_PREBUILT_OBSOLETE, row_update_prebuilt(), row_prebuilt_table_obsolete(): Remove. row_prebuilt_t: Remove the declaration from row0types.h. row_drop_table_for_mysql_no_commit(): Always print a warning if a table was added to the background drop queue.
2007-12-17 16:49:59 +01:00
new_primary ? LOCK_X : LOCK_S);
if (UNIV_UNLIKELY(error != DB_SUCCESS)) {
goto error_handling;
}
/* Latch the InnoDB data dictionary exclusively so that no deadlocks
or lock waits can happen in it during an index create operation. */
row_mysql_lock_data_dictionary(trx);
dict_locked = TRUE;
/* If a new primary key is defined for the table we need
to drop the original table and rebuild all indexes. */
if (UNIV_UNLIKELY(new_primary)) {
branches/zip: Fast index creation: Remove the ROW_PREBUILT_OBSOLETE nonsense. Active transactions must not switch table or index definitions on the fly, for several reasons, including the following: * copied indexes do not carry any history or locking information; that is, rollbacks, read views, and record locking would be broken * huge potential for race conditions, inconsistent reads and writes, loss of data, and corruption Instead of trying to track down if the table was changed during a transaction, acquire appropriate locks that protect the creation and dropping of indexes. innodb-index.test: Test the locking of CREATE INDEX and DROP INDEX. Test that consistent reads work across dropped indexes. lock_rec_insert_check_and_lock(): Relax the lock_table_has() assertion. When inserting a record into an index, the table must be at least IX-locked. However, when an index is being created, an IS-lock on the table is sufficient. row_merge_lock_table(): Add the parameter enum lock_mode mode, which must be LOCK_X or LOCK_S. row_merge_drop_table(): Assert that n_mysql_handles_opened == 0. Unconditionally drop the table. ha_innobase::add_index(): Acquire an X or S lock on the table, as appropriate. After acquiring an X lock, assert that n_mysql_handles_opened == 1. Remove the comments about dropping tables in the background. ha_innobase::final_drop_index(): Acquire an X lock on the table. dict_table_t: Remove version_number, to_be_dropped, and prebuilts. ins_node_t: Remove table_version_number. enum lock_mode: Move the definition from lock0lock.h to lock0types.h. ROW_PREBUILT_OBSOLETE, row_update_prebuilt(), row_prebuilt_table_obsolete(): Remove. row_prebuilt_t: Remove the declaration from row0types.h. row_drop_table_for_mysql_no_commit(): Always print a warning if a table was added to the background drop queue.
2007-12-17 16:49:59 +01:00
/* This transaction should be the only one
operating on the table. */
ut_a(innodb_table->n_mysql_handles_opened == 1);
char* new_table_name = innobase_create_temporary_tablename(
heap, '1', innodb_table->name);
/* Clone the table. */
trx_set_dict_operation(trx, TRX_DICT_OP_TABLE);
indexed_table = row_merge_create_temporary_table(
new_table_name, index_defs, innodb_table, trx);
if (!indexed_table) {
switch (trx->error_state) {
case DB_TABLESPACE_ALREADY_EXISTS:
case DB_DUPLICATE_KEY:
innobase_convert_tablename(new_table_name);
my_error(HA_ERR_TABLE_EXIST, MYF(0),
new_table_name);
error = HA_ERR_TABLE_EXIST;
break;
default:
error = convert_error_code_to_mysql(
trx->error_state, innodb_table->flags,
user_thd);
}
row_mysql_unlock_data_dictionary(trx);
goto err_exit;
}
trx->table_id = indexed_table->id;
}
/* Create the indexes in SYS_INDEXES and load into dictionary. */
for (ulint i = 0; i < num_of_idx; i++) {
index[i] = row_merge_create_index(trx, indexed_table,
&index_defs[i]);
if (!index[i]) {
error = trx->error_state;
goto error_handling;
}
num_created++;
}
ut_ad(error == DB_SUCCESS);
/* Commit the data dictionary transaction in order to release
the table locks on the system tables. Unfortunately, this
means that if MySQL crashes while creating a new primary key
inside row_merge_build_indexes(), indexed_table will not be
dropped on crash recovery. Thus, it will become orphaned. */
trx_commit_for_mysql(trx);
row_mysql_unlock_data_dictionary(trx);
dict_locked = FALSE;
ut_a(trx->n_active_thrs == 0);
ut_a(UT_LIST_GET_LEN(trx->signals) == 0);
if (UNIV_UNLIKELY(new_primary)) {
/* A primary key is to be built. Acquire an exclusive
table lock also on the table that is being created. */
ut_ad(indexed_table != innodb_table);
error = row_merge_lock_table(prebuilt->trx, indexed_table,
LOCK_X);
if (UNIV_UNLIKELY(error != DB_SUCCESS)) {
goto error_handling;
}
}
/* Read the clustered index of the table and build indexes
based on this information using temporary files and merge sort. */
error = row_merge_build_indexes(prebuilt->trx,
innodb_table, indexed_table,
branches/zip: Implement the reporting of duplicate key values to MySQL. innobase_rec_to_mysql(): New function, for converting an InnoDB clustered index record to MySQL table->record[0]. TODO: convert integer fields. Currently, integer fields are in big-endian byte order instead of host byte order, and signed integer fields are offset by 0x80000000. innobase_rec_reset(): New function, for resetting table->record[0]. row_merge_build_indexes(): Add the parameter TABLE* table (the MySQL table handle) for reporting duplicate key values. dtuple_from_fields(): New function, to convert an array of dfield_t* to dtuple_t. dtuple_get_n_ext(): New function, to compute the number of externally stored fields. row_merge_dup_t: Structure for counting and reporting duplicate records. row_merge_dup_report(): Function for counting and reporting duplicate records. row_merge_tuple_cmp(), row_merge_tuple_sort(): Replace the ulint* n_dup parameter with row_merge_dup_t* dup. row_merge_buf_sort(): Add the parameter row_merge_dup_t* dup, which is NULL when sorting a non-unique index. row_merge_buf_write(), row_merge_heap_create(), row_merge_read_rec(), row_merge_cmp(), row_merge_read_clustered_index(), row_merge_blocks(), row_merge(), row_merge_sort(): Add const qualifiers. row_merge_read_clustered_index(): Use a common error handling branch err_exit. Invoke row_merge_buf_sort() differently on unique indexes. row_merge_blocks(): note TODO: We could invoke innobase_rec_to_mysql() to report duplicate key values when creating a clustered index.
2007-09-26 13:56:26 +02:00
index, num_of_idx, table);
error_handling:
#ifdef UNIV_DEBUG
/* TODO: At the moment we can't handle the following statement
in our debugging code below:
alter table t drop index b, add index (b);
The fix will have to parse the SQL and note that the index
being added has the same name as the the one being dropped and
ignore that in the dup index check.*/
//dict_table_check_for_dup_indexes(prebuilt->table);
#endif
/* After an error, remove all those index definitions from the
dictionary which were defined. */
switch (error) {
const char* old_name;
char* tmp_name;
case DB_SUCCESS:
ut_a(!dict_locked);
row_mysql_lock_data_dictionary(trx);
dict_locked = TRUE;
if (!new_primary) {
error = row_merge_rename_indexes(trx, indexed_table);
if (error != DB_SUCCESS) {
row_merge_drop_indexes(trx, indexed_table,
index, num_created);
}
goto convert_error;
}
/* If a new primary key was defined for the table and
there was no error at this point, we can now rename
the old table as a temporary table, rename the new
temporary table as the old table and drop the old table. */
old_name = innodb_table->name;
tmp_name = innobase_create_temporary_tablename(heap, '2',
old_name);
error = row_merge_rename_tables(innodb_table, indexed_table,
tmp_name, trx);
if (error != DB_SUCCESS) {
row_merge_drop_table(trx, indexed_table);
switch (error) {
case DB_TABLESPACE_ALREADY_EXISTS:
case DB_DUPLICATE_KEY:
innobase_convert_tablename(tmp_name);
my_error(HA_ERR_TABLE_EXIST, MYF(0), tmp_name);
error = HA_ERR_TABLE_EXIST;
break;
default:
goto convert_error;
}
break;
}
trx_commit_for_mysql(prebuilt->trx);
row_prebuilt_free(prebuilt, TRUE);
prebuilt = row_create_prebuilt(indexed_table);
branches/zip: Fast index creation: Remove the ROW_PREBUILT_OBSOLETE nonsense. Active transactions must not switch table or index definitions on the fly, for several reasons, including the following: * copied indexes do not carry any history or locking information; that is, rollbacks, read views, and record locking would be broken * huge potential for race conditions, inconsistent reads and writes, loss of data, and corruption Instead of trying to track down if the table was changed during a transaction, acquire appropriate locks that protect the creation and dropping of indexes. innodb-index.test: Test the locking of CREATE INDEX and DROP INDEX. Test that consistent reads work across dropped indexes. lock_rec_insert_check_and_lock(): Relax the lock_table_has() assertion. When inserting a record into an index, the table must be at least IX-locked. However, when an index is being created, an IS-lock on the table is sufficient. row_merge_lock_table(): Add the parameter enum lock_mode mode, which must be LOCK_X or LOCK_S. row_merge_drop_table(): Assert that n_mysql_handles_opened == 0. Unconditionally drop the table. ha_innobase::add_index(): Acquire an X or S lock on the table, as appropriate. After acquiring an X lock, assert that n_mysql_handles_opened == 1. Remove the comments about dropping tables in the background. ha_innobase::final_drop_index(): Acquire an X lock on the table. dict_table_t: Remove version_number, to_be_dropped, and prebuilts. ins_node_t: Remove table_version_number. enum lock_mode: Move the definition from lock0lock.h to lock0types.h. ROW_PREBUILT_OBSOLETE, row_update_prebuilt(), row_prebuilt_table_obsolete(): Remove. row_prebuilt_t: Remove the declaration from row0types.h. row_drop_table_for_mysql_no_commit(): Always print a warning if a table was added to the background drop queue.
2007-12-17 16:49:59 +01:00
indexed_table->n_mysql_handles_opened++;
error = row_merge_drop_table(trx, innodb_table);
goto convert_error;
case DB_TOO_BIG_RECORD:
my_error(HA_ERR_TO_BIG_ROW, MYF(0));
goto error;
case DB_PRIMARY_KEY_IS_NULL:
my_error(ER_PRIMARY_CANT_HAVE_NULL, MYF(0));
/* fall through */
case DB_DUPLICATE_KEY:
error:
prebuilt->trx->error_info = NULL;
/* fall through */
default:
if (new_primary) {
row_merge_drop_table(trx, indexed_table);
} else {
if (!dict_locked) {
row_mysql_lock_data_dictionary(trx);
dict_locked = TRUE;
}
row_merge_drop_indexes(trx, indexed_table,
index, num_created);
}
convert_error:
error = convert_error_code_to_mysql(error,
innodb_table->flags,
user_thd);
}
mem_heap_free(heap);
trx_commit_for_mysql(trx);
if (prebuilt->trx) {
trx_commit_for_mysql(prebuilt->trx);
}
if (dict_locked) {
row_mysql_unlock_data_dictionary(trx);
}
trx_free_for_mysql(trx);
/* There might be work for utility threads.*/
srv_active_wake_master_thread();
DBUG_RETURN(error);
}
/***********************************************************************
Prepare to drop some indexes of a table. */
UNIV_INTERN
int
ha_innobase::prepare_drop_index(
/*============================*/
/* out: 0 or error number */
TABLE* table, /* in: Table where indexes are dropped */
uint* key_num, /* in: Key nums to be dropped */
uint num_of_keys) /* in: Number of keys to be dropped */
{
trx_t* trx;
int err = 0;
uint n_key;
DBUG_ENTER("ha_innobase::prepare_drop_index");
ut_ad(table);
ut_ad(key_num);
ut_ad(num_of_keys);
if (srv_created_new_raw || srv_force_recovery) {
DBUG_RETURN(HA_ERR_WRONG_COMMAND);
}
update_thd();
trx_search_latch_release_if_reserved(prebuilt->trx);
trx = prebuilt->trx;
/* Test and mark all the indexes to be dropped */
row_mysql_lock_data_dictionary(trx);
/* Check that none of the indexes have previously been flagged
for deletion. */
{
const dict_index_t* index
= dict_table_get_first_index(prebuilt->table);
do {
ut_a(!index->to_be_dropped);
index = dict_table_get_next_index(index);
} while (index);
}
for (n_key = 0; n_key < num_of_keys; n_key++) {
const KEY* key;
dict_index_t* index;
key = table->key_info + key_num[n_key];
index = dict_table_get_index_on_name_and_min_id(
prebuilt->table, key->name);
if (!index) {
sql_print_error("InnoDB could not find key n:o %u "
"with name %s for table %s",
key_num[n_key],
key ? key->name : "NULL",
prebuilt->table->name);
err = HA_ERR_KEY_NOT_FOUND;
goto func_exit;
}
/* Refuse to drop the clustered index. It would be
better to automatically generate a clustered index,
but mysql_alter_table() will call this method only
after ha_innobase::add_index(). */
if (dict_index_is_clust(index)) {
my_error(ER_REQUIRES_PRIMARY_KEY, MYF(0));
err = -1;
goto func_exit;
}
index->to_be_dropped = TRUE;
}
/* If FOREIGN_KEY_CHECK = 1 you may not drop an index defined
for a foreign key constraint because InnoDB requires that both
tables contain indexes for the constraint. Note that CREATE
INDEX id ON table does a CREATE INDEX and DROP INDEX, and we
can ignore here foreign keys because a new index for the
foreign key has already been created.
We check for the foreign key constraints after marking the
candidate indexes for deletion, because when we check for an
equivalent foreign index we don't want to select an index that
is later deleted. */
if (trx->check_foreigns
&& thd_sql_command(user_thd) != SQLCOM_CREATE_INDEX) {
dict_index_t* index;
for (index = dict_table_get_first_index(prebuilt->table);
index;
index = dict_table_get_next_index(index)) {
dict_foreign_t* foreign;
if (!index->to_be_dropped) {
continue;
}
/* Check if the index is referenced. */
foreign = dict_table_get_referenced_constraint(
prebuilt->table, index);
if (foreign) {
index_needed:
trx_set_detailed_error(
trx,
"Index needed in foreign key "
"constraint");
trx->error_info = index;
err = HA_ERR_DROP_INDEX_FK;
break;
} else {
/* Check if this index references some
other table */
foreign = dict_table_get_foreign_constraint(
prebuilt->table, index);
if (foreign) {
ut_a(foreign->foreign_index == index);
/* Search for an equivalent index that
the foreign key constraint could use
if this index were to be deleted. */
if (!dict_foreign_find_equiv_index(
foreign)) {
goto index_needed;
}
}
}
}
} else if (thd_sql_command(user_thd) == SQLCOM_CREATE_INDEX) {
/* This is a drop of a foreign key constraint index that
was created by MySQL when the constraint was added. MySQL
does this when the user creates an index explicitly which
can be used in place of the automatically generated index. */
dict_index_t* index;
for (index = dict_table_get_first_index(prebuilt->table);
index;
index = dict_table_get_next_index(index)) {
dict_foreign_t* foreign;
if (!index->to_be_dropped) {
continue;
}
/* Check if this index references some other table */
foreign = dict_table_get_foreign_constraint(
prebuilt->table, index);
if (foreign == NULL) {
continue;
}
ut_a(foreign->foreign_index == index);
/* Search for an equivalent index that the
foreign key constraint could use if this index
were to be deleted. */
if (!dict_foreign_find_equiv_index(foreign)) {
trx_set_detailed_error(
trx,
"Index needed in foreign key "
"constraint");
trx->error_info = foreign->foreign_index;
err = HA_ERR_DROP_INDEX_FK;
break;
}
}
}
func_exit:
if (err) {
/* Undo our changes since there was some sort of error. */
dict_index_t* index
= dict_table_get_first_index(prebuilt->table);
do {
index->to_be_dropped = FALSE;
index = dict_table_get_next_index(index);
} while (index);
}
row_mysql_unlock_data_dictionary(trx);
DBUG_RETURN(err);
}
/***********************************************************************
Drop the indexes that were passed to a successful prepare_drop_index(). */
UNIV_INTERN
int
ha_innobase::final_drop_index(
/*==========================*/
/* out: 0 or error number */
TABLE* table) /* in: Table where indexes are dropped */
{
dict_index_t* index; /* Index to be dropped */
trx_t* trx; /* Transaction */
branches/zip: Fast index creation: Remove the ROW_PREBUILT_OBSOLETE nonsense. Active transactions must not switch table or index definitions on the fly, for several reasons, including the following: * copied indexes do not carry any history or locking information; that is, rollbacks, read views, and record locking would be broken * huge potential for race conditions, inconsistent reads and writes, loss of data, and corruption Instead of trying to track down if the table was changed during a transaction, acquire appropriate locks that protect the creation and dropping of indexes. innodb-index.test: Test the locking of CREATE INDEX and DROP INDEX. Test that consistent reads work across dropped indexes. lock_rec_insert_check_and_lock(): Relax the lock_table_has() assertion. When inserting a record into an index, the table must be at least IX-locked. However, when an index is being created, an IS-lock on the table is sufficient. row_merge_lock_table(): Add the parameter enum lock_mode mode, which must be LOCK_X or LOCK_S. row_merge_drop_table(): Assert that n_mysql_handles_opened == 0. Unconditionally drop the table. ha_innobase::add_index(): Acquire an X or S lock on the table, as appropriate. After acquiring an X lock, assert that n_mysql_handles_opened == 1. Remove the comments about dropping tables in the background. ha_innobase::final_drop_index(): Acquire an X lock on the table. dict_table_t: Remove version_number, to_be_dropped, and prebuilts. ins_node_t: Remove table_version_number. enum lock_mode: Move the definition from lock0lock.h to lock0types.h. ROW_PREBUILT_OBSOLETE, row_update_prebuilt(), row_prebuilt_table_obsolete(): Remove. row_prebuilt_t: Remove the declaration from row0types.h. row_drop_table_for_mysql_no_commit(): Always print a warning if a table was added to the background drop queue.
2007-12-17 16:49:59 +01:00
int err;
DBUG_ENTER("ha_innobase::final_drop_index");
ut_ad(table);
if (srv_created_new_raw || srv_force_recovery) {
DBUG_RETURN(HA_ERR_WRONG_COMMAND);
}
update_thd();
trx_search_latch_release_if_reserved(prebuilt->trx);
/* Create a background transaction for the operations on
the data dictionary tables. */
branches/innodb+: Merge revisions 3602:3931 from branches/zip: ------------------------------------------------------------------------ r3607 | marko | 2008-12-30 22:33:31 +0200 (Tue, 30 Dec 2008) | 20 lines branches/zip: Remove the dependency on the MySQL HASH table implementation. Use the InnoDB hash table for keeping track of INNOBASE_SHARE objects. struct st_innobase_share: Make table_name const uchar*. Add the member table_name_hash. innobase_open_tables: Change the type from HASH to hash_table_t*. innobase_get_key(): Remove. innobase_fold_name(): New function, for computing the fold value for the InnoDB hash table. get_share(), free_share(): Use the InnoDB hash functions. innobase_end(): Free innobase_open_tables before shutting down InnoDB. Shutting down InnoDB will invalidate all memory allocated via InnoDB. rb://65 approved by Heikki Tuuri. This addresses Issue #104. ------------------------------------------------------------------------ r3608 | marko | 2008-12-30 22:45:04 +0200 (Tue, 30 Dec 2008) | 22 lines branches/zip: When setting the PAGE_LEVEL of a compressed B-tree page from or to 0, compress the page at the same time. This is necessary, because the column information stored on the compressed page will differ between leaf and non-leaf pages. Leaf pages are identified by PAGE_LEVEL=0. This bug was reported as Issue #150. Document the similarity between btr_page_create() and btr_page_empty(). Make the function signature of btr_page_empty() identical with btr_page_create(). (This will add the parameter "level".) btr_root_raise_and_insert(): Replace some code with a call to btr_page_empty(). btr_attach_half_pages(): Assert that the page level has already been set on both block and new_block. Do not set it again. btr_discard_only_page_on_level(): Document that this function is probably never called. Make it work on any height tree. (Tested on 2-high tree by disabling btr_lift_page_up().) rb://68 ------------------------------------------------------------------------ r3612 | marko | 2009-01-02 11:02:44 +0200 (Fri, 02 Jan 2009) | 14 lines branches/zip: Merge c2998 from branches/6.0, so that the same InnoDB Plugin source tree will work both under 5.1 and 6.0. Do not add the test case innodb_ctype_ldml.test, because it would not work under MySQL 5.1. Refuse to create tables whose columns contain collation IDs above 255. This removes an assertion failure that was introduced in WL#4164 (Two-byte collation IDs). create_table_def(): Do not fail an assertion if a column contains a charset-collation ID greater than 256. Instead, issue an error and refuse to create the table. The original change (branches/6.0 r2998) was rb://51 approved by Calvin Sun. ------------------------------------------------------------------------ r3613 | inaam | 2009-01-02 15:10:50 +0200 (Fri, 02 Jan 2009) | 6 lines branches/zip: Implement the parameter innodb_use_sys_malloc (false by default), for disabling InnoDB's internal memory allocator and using system malloc/free instead. rb://62 approved by Marko ------------------------------------------------------------------------ r3614 | marko | 2009-01-02 15:55:12 +0200 (Fri, 02 Jan 2009) | 1 line branches/zip: ChangeLog: Document r3608 and r3613. ------------------------------------------------------------------------ r3615 | marko | 2009-01-02 15:57:51 +0200 (Fri, 02 Jan 2009) | 1 line branches/zip: ChangeLog: Clarify the impact of r3608. ------------------------------------------------------------------------ r3616 | marko | 2009-01-03 00:23:30 +0200 (Sat, 03 Jan 2009) | 1 line branches/zip: srv_suspend_mysql_thread(): Add some clarifying comments. ------------------------------------------------------------------------ r3618 | marko | 2009-01-05 12:54:53 +0200 (Mon, 05 Jan 2009) | 15 lines branches/zip: Merge revisions 3598:3601 from branches/5.1: ------------------------------------------------------------------------ r3601 | marko | 2008-12-22 16:05:19 +0200 (Mon, 22 Dec 2008) | 9 lines branches/5.1: Make SET SESSION TRANSACTION ISOLATION LEVEL READ COMMITTED a true replacement of SET GLOBAL INNODB_LOCKS_UNSAFE_FOR_BINLOG=1. This fixes an error that was introduced in r370, causing semi-consistent read not to not unlock rows in READ COMMITTED mode. (Bug #41671, Issue #146) rb://67 approved by Heikki Tuuri ------------------------------------------------------------------------ ------------------------------------------------------------------------ r3623 | vasil | 2009-01-06 09:56:32 +0200 (Tue, 06 Jan 2009) | 7 lines branches/zip: Add patch to fix the failing main.variables mysql-test. It started failing after the variable innodb_use_sys_malloc was added because it matches '%alloc%' and the test is badly written and expects that no new variables like that will ever be added. ------------------------------------------------------------------------ r3795 | marko | 2009-01-07 16:17:47 +0200 (Wed, 07 Jan 2009) | 7 lines branches/zip: row_merge_tuple_cmp(): Do not report a duplicate key value if any of the fields are NULL. While the tuples are equal in the sorting order, SQL NULL is defined to be logically inequal to anything else. (Bug #41904) rb://70 approved by Heikki Tuuri ------------------------------------------------------------------------ r3796 | marko | 2009-01-07 16:19:32 +0200 (Wed, 07 Jan 2009) | 1 line branches/zip: Add the tests that were forgotten from r3795. ------------------------------------------------------------------------ r3797 | marko | 2009-01-07 16:22:18 +0200 (Wed, 07 Jan 2009) | 22 lines branches/zip: Do not call trx_allocate_for_mysql() directly, but use helper functions that initialize some members of the transaction struct. (Bug #41680) innobase_trx_init(): New function: initialize some fields of a transaction struct from a MySQL THD object. innobase_trx_allocate(): New function: allocate and initialize a transaction struct. check_trx_exists(): Use the above two functions. ha_innobase::delete_table(), ha_innobase::rename_table(), ha_innobase::add_index(), ha_innobase::final_drop_index(): Use innobase_trx_allocate(). innobase_drop_database(): In the Windows plugin, initialize the trx_t specially, because the THD is not available. Otherwise, use innobase_trx_allocate(). rb://69 accepted by Heikki Tuuri ------------------------------------------------------------------------ r3798 | marko | 2009-01-07 16:42:42 +0200 (Wed, 07 Jan 2009) | 8 lines branches/zip: row_merge_drop_temp_indexes(): Do not lock the rows of SYS_INDEXES when looking for partially created indexes. Use the transaction isolation level READ UNCOMMITTED to avoid interfering with locks held by incomplete transactions that will be rolled back in a subsequent step in the recovery. (Issue #152) Approved by Heikki Tuuri ------------------------------------------------------------------------ r3852 | vasil | 2009-01-08 22:10:10 +0200 (Thu, 08 Jan 2009) | 4 lines branches/zip: Add ChangeLog entries for r3795 r3796 r3797 r3798. ------------------------------------------------------------------------ r3866 | marko | 2009-01-09 15:09:51 +0200 (Fri, 09 Jan 2009) | 2 lines branches/zip: buf_flush_try_page(): Move some common code from each switch case before the switch block. ------------------------------------------------------------------------ r3867 | marko | 2009-01-09 15:13:14 +0200 (Fri, 09 Jan 2009) | 2 lines branches/zip: buf_flush_try_page(): Introduce the variable is_compressed for caching the result of buf_page_get_state(bpage) == BUF_BLOCK_FILE_PAGE. ------------------------------------------------------------------------ r3868 | marko | 2009-01-09 15:40:11 +0200 (Fri, 09 Jan 2009) | 4 lines branches/zip: buf_flush_insert_into_flush_list(), buf_flush_insert_sorted_into_flush_list(): Remove unused code. Change the parameter to buf_block_t* block and assert that block->state == BUF_BLOCK_FILE_PAGE. This is part of Issue #155. ------------------------------------------------------------------------ r3873 | marko | 2009-01-09 22:27:40 +0200 (Fri, 09 Jan 2009) | 17 lines branches/zip: Some non-functional changes related to Issue #155. buf_page_struct: Note that space and offset are also protected by buf_pool_mutex. They are only assigned to by buf_block_set_file_page(). Thus, it suffices for buf_flush_batch() to hold just buf_pool_mutex when checking these fields. buf_flush_try_page(): Rename "locked" to "is_s_latched", per Heikki's request. buf_flush_batch(): Move the common statement mutex_exit(block_mutex) from all if-else if-else branches before the if block. Remove the redundant test (buf_pool->init_flush[flush_type] == FALSE) that was apparently copied from buf_flush_write_complete(). buf_flush_write_block_low(): Note why it is safe not to hold buf_pool_mutex or block_mutex. Enumerate the assumptions in debug assertions. ------------------------------------------------------------------------ r3874 | marko | 2009-01-09 23:09:06 +0200 (Fri, 09 Jan 2009) | 4 lines branches/zip: Add comments related to Issue #155. buf_flush_try_page(): Note why it is safe to access bpage without holding buf_pool_mutex or block_mutex. ------------------------------------------------------------------------ r3875 | marko | 2009-01-09 23:15:12 +0200 (Fri, 09 Jan 2009) | 11 lines branches/zip: Non-functional change: Tighten debug assertions and remove dead code. buf_flush_ready_for_flush(), buf_flush_try_page(): Assert that flush_type is one of BUF_FLUSH_LRU or BUF_FLUSH_LIST. The flush_type comes from buf_flush_batch(), which already asserts this. The assertion holds for all calls in the source code. buf_flush_try_page(): Remove the dead case BUF_FLUSH_SINGLE_PAGE of switch (flush_type). ------------------------------------------------------------------------ r3879 | marko | 2009-01-12 12:46:44 +0200 (Mon, 12 Jan 2009) | 14 lines branches/zip: Simplify the flushing of dirty pages from the buffer pool. buf_flush_try_page(): Rename to buf_flush_page(), and change the return type to void. Replace the parameters space, offset with bpage, and remove the second page hash lookup. Note and assert that both buf_pool_mutex and block_mutex must now be held upon entering the function. They will still be released by this function. buf_flush_try_neighbors(): Replace buf_flush_try_page() with buf_flush_page(). Make the logic easier to follow by not negating the precondition of buf_flush_page(). rb://73 approved by Sunny Bains. This is related to Issue #157. ------------------------------------------------------------------------ r3880 | marko | 2009-01-12 13:24:37 +0200 (Mon, 12 Jan 2009) | 2 lines branches/zip: buf_flush_page(): Fix a comment that should have been fixed in r3879. Spotted by Sunny. ------------------------------------------------------------------------ r3881 | marko | 2009-01-12 14:25:22 +0200 (Mon, 12 Jan 2009) | 2 lines branches/zip: buf_page_get_newest_modification(): Use the block mutex instead of the buffer pool mutex. This is related to Issue #157. ------------------------------------------------------------------------ r3882 | marko | 2009-01-12 14:40:08 +0200 (Mon, 12 Jan 2009) | 3 lines branches/zip: struct mtr_struct: Remove the unused field magic_n unless UNIV_DEBUG is defined. mtr->magic_n is only assigned to and checked in UNIV_DEBUG builds. ------------------------------------------------------------------------ r3883 | marko | 2009-01-12 14:48:59 +0200 (Mon, 12 Jan 2009) | 1 line branches/zip: Non-functional change: Use ut_d when assigning to mtr->state. ------------------------------------------------------------------------ r3884 | marko | 2009-01-12 18:56:11 +0200 (Mon, 12 Jan 2009) | 16 lines branches/zip: Non-functional change: Add some debug assertions and comments. buf_page_t: Note that the LRU fields are protected by buf_pool_mutex only, not block->mutex or buf_pool_zip_mutex. buf_page_get_freed_page_clock(): Note that this is sometimes invoked without mutex protection. buf_pool_get_oldest_modification(): Note that the result may be out of date. buf_page_get_LRU_position(), buf_page_is_old(): Assert that the buffer pool mutex is being held. buf_page_release(): Assert that dirty blocks are in the flush list. ------------------------------------------------------------------------ r3896 | marko | 2009-01-13 09:30:26 +0200 (Tue, 13 Jan 2009) | 2 lines branches/zip: buf_flush_try_neighbors(): Fix a bug that was introduced in r3879 (rb://73). ------------------------------------------------------------------------ r3900 | marko | 2009-01-13 10:32:24 +0200 (Tue, 13 Jan 2009) | 1 line branches/zip: Fix some comments to say buf_pool_mutex. ------------------------------------------------------------------------ r3907 | marko | 2009-01-13 11:54:01 +0200 (Tue, 13 Jan 2009) | 3 lines branches/zip: row_merge_create_temporary_table(): On error, row_create_table_for_mysql() already frees new_table. Do not attempt to free it again. ------------------------------------------------------------------------ r3908 | marko | 2009-01-13 12:34:32 +0200 (Tue, 13 Jan 2009) | 1 line branches/zip: Enable HASH_ASSERT_OWNED independently of UNIV_SYNC_DEBUG. ------------------------------------------------------------------------ r3914 | marko | 2009-01-13 21:46:22 +0200 (Tue, 13 Jan 2009) | 37 lines branches/zip: In hash table lookups, assert that the traversed items satisfy some conditions when UNIV_DEBUG is defined. HASH_SEARCH(): New parameter: ASSERTION. All users will pass an appropriate ut_ad() or nothing. dict_table_add_to_columns(): Assert that the table being added to the data dictionary cache is not already being pointed to by the name_hash and id_hash tables. HASH_SEARCH_ALL(): New macro, for use in dict_table_add_to_columns(). dict_mem_table_free(): Set ut_d(table->cached = FALSE), so that we can check ut_ad(table->cached) when traversing the hash tables, as in HASH_SEARCH(name_hash, dict_sys->table_hash, ...) and HASH_SEARCH(id_hash, dict_sys->table_id_hash, ...). dict_table_get_low(), dict_table_get_on_id_low(): Assert ut_ad(!table || table->cached). fil_space_get_by_id(): Check ut_ad(space->magic_n == FIL_SPACE_MAGIC_N) in HASH_SEARCH(hash, fil_system->spaces, ...). fil_space_get_by_name(): Check ut_ad(space->magic_n == FIL_SPACE_MAGIC_N) in HASH_SEARCH(name_hash, fil_system->name_hash, ...). buf_buddy_block_free(): Check that the blocks are in valid state in HASH_SEARCH(hash, buf_pool->zip_hash, ...). buf_page_hash_get(): Check that the blocks are in valid state in HASH_SEARCH(hash, buf_pool->page_hash, ...). get_share(), free_share(): Check ut_ad(share->use_count > 0) in HASH_SEARCH(table_name_hash, innobase_open_tables, ...). This was posted as rb://75 for tracking down errors similar to Issue #153. ------------------------------------------------------------------------ r3931 | marko | 2009-01-14 16:06:22 +0200 (Wed, 14 Jan 2009) | 26 lines branches/zip: Merge revisions 3601:3930 from branches/5.1: ------------------------------------------------------------------------ r3911 | sunny | 2009-01-13 14:15:24 +0200 (Tue, 13 Jan 2009) | 13 lines branches/5.1: Fix Bug#38187 Error 153 when creating savepoints InnoDB previously treated savepoints as a stack e.g., SAVEPOINT a; SAVEPOINT b; SAVEPOINT c; SAVEPOINT b; <- This would delete b and c. This fix changes the behavior to: SAVEPOINT a; SAVEPOINT b; SAVEPOINT c; SAVEPOINT b; <- Does not delete savepoint c ------------------------------------------------------------------------ r3930 | marko | 2009-01-14 15:51:30 +0200 (Wed, 14 Jan 2009) | 4 lines branches/5.1: dict_load_table(): If dict_load_indexes() fails, invoke dict_table_remove_from_cache() instead of dict_mem_table_free(), so that the data dictionary will not point to freed data. (Bug #42075, Issue #153, rb://76 approved by Heikki Tuuri) ------------------------------------------------------------------------ ------------------------------------------------------------------------
2009-01-14 15:25:45 +01:00
trx = innobase_trx_allocate(user_thd);
trx_start_if_not_started(trx);
/* Flag this transaction as a dictionary operation, so that
the data dictionary will be locked in crash recovery. */
trx_set_dict_operation(trx, TRX_DICT_OP_INDEX);
branches/zip: Fast index creation: Remove the ROW_PREBUILT_OBSOLETE nonsense. Active transactions must not switch table or index definitions on the fly, for several reasons, including the following: * copied indexes do not carry any history or locking information; that is, rollbacks, read views, and record locking would be broken * huge potential for race conditions, inconsistent reads and writes, loss of data, and corruption Instead of trying to track down if the table was changed during a transaction, acquire appropriate locks that protect the creation and dropping of indexes. innodb-index.test: Test the locking of CREATE INDEX and DROP INDEX. Test that consistent reads work across dropped indexes. lock_rec_insert_check_and_lock(): Relax the lock_table_has() assertion. When inserting a record into an index, the table must be at least IX-locked. However, when an index is being created, an IS-lock on the table is sufficient. row_merge_lock_table(): Add the parameter enum lock_mode mode, which must be LOCK_X or LOCK_S. row_merge_drop_table(): Assert that n_mysql_handles_opened == 0. Unconditionally drop the table. ha_innobase::add_index(): Acquire an X or S lock on the table, as appropriate. After acquiring an X lock, assert that n_mysql_handles_opened == 1. Remove the comments about dropping tables in the background. ha_innobase::final_drop_index(): Acquire an X lock on the table. dict_table_t: Remove version_number, to_be_dropped, and prebuilts. ins_node_t: Remove table_version_number. enum lock_mode: Move the definition from lock0lock.h to lock0types.h. ROW_PREBUILT_OBSOLETE, row_update_prebuilt(), row_prebuilt_table_obsolete(): Remove. row_prebuilt_t: Remove the declaration from row0types.h. row_drop_table_for_mysql_no_commit(): Always print a warning if a table was added to the background drop queue.
2007-12-17 16:49:59 +01:00
/* Lock the table exclusively, to ensure that no active
transaction depends on an index that is being dropped. */
err = convert_error_code_to_mysql(
row_merge_lock_table(prebuilt->trx, prebuilt->table, LOCK_X),
prebuilt->table->flags, user_thd);
branches/zip: Fast index creation: Remove the ROW_PREBUILT_OBSOLETE nonsense. Active transactions must not switch table or index definitions on the fly, for several reasons, including the following: * copied indexes do not carry any history or locking information; that is, rollbacks, read views, and record locking would be broken * huge potential for race conditions, inconsistent reads and writes, loss of data, and corruption Instead of trying to track down if the table was changed during a transaction, acquire appropriate locks that protect the creation and dropping of indexes. innodb-index.test: Test the locking of CREATE INDEX and DROP INDEX. Test that consistent reads work across dropped indexes. lock_rec_insert_check_and_lock(): Relax the lock_table_has() assertion. When inserting a record into an index, the table must be at least IX-locked. However, when an index is being created, an IS-lock on the table is sufficient. row_merge_lock_table(): Add the parameter enum lock_mode mode, which must be LOCK_X or LOCK_S. row_merge_drop_table(): Assert that n_mysql_handles_opened == 0. Unconditionally drop the table. ha_innobase::add_index(): Acquire an X or S lock on the table, as appropriate. After acquiring an X lock, assert that n_mysql_handles_opened == 1. Remove the comments about dropping tables in the background. ha_innobase::final_drop_index(): Acquire an X lock on the table. dict_table_t: Remove version_number, to_be_dropped, and prebuilts. ins_node_t: Remove table_version_number. enum lock_mode: Move the definition from lock0lock.h to lock0types.h. ROW_PREBUILT_OBSOLETE, row_update_prebuilt(), row_prebuilt_table_obsolete(): Remove. row_prebuilt_t: Remove the declaration from row0types.h. row_drop_table_for_mysql_no_commit(): Always print a warning if a table was added to the background drop queue.
2007-12-17 16:49:59 +01:00
row_mysql_lock_data_dictionary(trx);
branches/zip: Fast index creation: Remove the ROW_PREBUILT_OBSOLETE nonsense. Active transactions must not switch table or index definitions on the fly, for several reasons, including the following: * copied indexes do not carry any history or locking information; that is, rollbacks, read views, and record locking would be broken * huge potential for race conditions, inconsistent reads and writes, loss of data, and corruption Instead of trying to track down if the table was changed during a transaction, acquire appropriate locks that protect the creation and dropping of indexes. innodb-index.test: Test the locking of CREATE INDEX and DROP INDEX. Test that consistent reads work across dropped indexes. lock_rec_insert_check_and_lock(): Relax the lock_table_has() assertion. When inserting a record into an index, the table must be at least IX-locked. However, when an index is being created, an IS-lock on the table is sufficient. row_merge_lock_table(): Add the parameter enum lock_mode mode, which must be LOCK_X or LOCK_S. row_merge_drop_table(): Assert that n_mysql_handles_opened == 0. Unconditionally drop the table. ha_innobase::add_index(): Acquire an X or S lock on the table, as appropriate. After acquiring an X lock, assert that n_mysql_handles_opened == 1. Remove the comments about dropping tables in the background. ha_innobase::final_drop_index(): Acquire an X lock on the table. dict_table_t: Remove version_number, to_be_dropped, and prebuilts. ins_node_t: Remove table_version_number. enum lock_mode: Move the definition from lock0lock.h to lock0types.h. ROW_PREBUILT_OBSOLETE, row_update_prebuilt(), row_prebuilt_table_obsolete(): Remove. row_prebuilt_t: Remove the declaration from row0types.h. row_drop_table_for_mysql_no_commit(): Always print a warning if a table was added to the background drop queue.
2007-12-17 16:49:59 +01:00
if (UNIV_UNLIKELY(err)) {
/* Unmark the indexes to be dropped. */
for (index = dict_table_get_first_index(prebuilt->table);
index; index = dict_table_get_next_index(index)) {
index->to_be_dropped = FALSE;
}
branches/zip: Fast index creation: Remove the ROW_PREBUILT_OBSOLETE nonsense. Active transactions must not switch table or index definitions on the fly, for several reasons, including the following: * copied indexes do not carry any history or locking information; that is, rollbacks, read views, and record locking would be broken * huge potential for race conditions, inconsistent reads and writes, loss of data, and corruption Instead of trying to track down if the table was changed during a transaction, acquire appropriate locks that protect the creation and dropping of indexes. innodb-index.test: Test the locking of CREATE INDEX and DROP INDEX. Test that consistent reads work across dropped indexes. lock_rec_insert_check_and_lock(): Relax the lock_table_has() assertion. When inserting a record into an index, the table must be at least IX-locked. However, when an index is being created, an IS-lock on the table is sufficient. row_merge_lock_table(): Add the parameter enum lock_mode mode, which must be LOCK_X or LOCK_S. row_merge_drop_table(): Assert that n_mysql_handles_opened == 0. Unconditionally drop the table. ha_innobase::add_index(): Acquire an X or S lock on the table, as appropriate. After acquiring an X lock, assert that n_mysql_handles_opened == 1. Remove the comments about dropping tables in the background. ha_innobase::final_drop_index(): Acquire an X lock on the table. dict_table_t: Remove version_number, to_be_dropped, and prebuilts. ins_node_t: Remove table_version_number. enum lock_mode: Move the definition from lock0lock.h to lock0types.h. ROW_PREBUILT_OBSOLETE, row_update_prebuilt(), row_prebuilt_table_obsolete(): Remove. row_prebuilt_t: Remove the declaration from row0types.h. row_drop_table_for_mysql_no_commit(): Always print a warning if a table was added to the background drop queue.
2007-12-17 16:49:59 +01:00
goto func_exit;
}
/* Drop indexes marked to be dropped */
index = dict_table_get_first_index(prebuilt->table);
while (index) {
dict_index_t* next_index;
next_index = dict_table_get_next_index(index);
if (index->to_be_dropped) {
row_merge_drop_index(index, prebuilt->table, trx);
}
index = next_index;
}
/* Check that all flagged indexes were dropped. */
for (index = dict_table_get_first_index(prebuilt->table);
index; index = dict_table_get_next_index(index)) {
ut_a(!index->to_be_dropped);
}
#ifdef UNIV_DEBUG
dict_table_check_for_dup_indexes(prebuilt->table);
#endif
branches/zip: Fast index creation: Remove the ROW_PREBUILT_OBSOLETE nonsense. Active transactions must not switch table or index definitions on the fly, for several reasons, including the following: * copied indexes do not carry any history or locking information; that is, rollbacks, read views, and record locking would be broken * huge potential for race conditions, inconsistent reads and writes, loss of data, and corruption Instead of trying to track down if the table was changed during a transaction, acquire appropriate locks that protect the creation and dropping of indexes. innodb-index.test: Test the locking of CREATE INDEX and DROP INDEX. Test that consistent reads work across dropped indexes. lock_rec_insert_check_and_lock(): Relax the lock_table_has() assertion. When inserting a record into an index, the table must be at least IX-locked. However, when an index is being created, an IS-lock on the table is sufficient. row_merge_lock_table(): Add the parameter enum lock_mode mode, which must be LOCK_X or LOCK_S. row_merge_drop_table(): Assert that n_mysql_handles_opened == 0. Unconditionally drop the table. ha_innobase::add_index(): Acquire an X or S lock on the table, as appropriate. After acquiring an X lock, assert that n_mysql_handles_opened == 1. Remove the comments about dropping tables in the background. ha_innobase::final_drop_index(): Acquire an X lock on the table. dict_table_t: Remove version_number, to_be_dropped, and prebuilts. ins_node_t: Remove table_version_number. enum lock_mode: Move the definition from lock0lock.h to lock0types.h. ROW_PREBUILT_OBSOLETE, row_update_prebuilt(), row_prebuilt_table_obsolete(): Remove. row_prebuilt_t: Remove the declaration from row0types.h. row_drop_table_for_mysql_no_commit(): Always print a warning if a table was added to the background drop queue.
2007-12-17 16:49:59 +01:00
func_exit:
trx_commit_for_mysql(trx);
trx_commit_for_mysql(prebuilt->trx);
row_mysql_unlock_data_dictionary(trx);
/* Flush the log to reduce probability that the .frm files and
the InnoDB data dictionary get out-of-sync if the user runs
with innodb_flush_log_at_trx_commit = 0 */
log_buffer_flush_to_disk();
trx_free_for_mysql(trx);
/* Tell the InnoDB server that there might be work for
utility threads: */
srv_active_wake_master_thread();
branches/zip: Fast index creation: Remove the ROW_PREBUILT_OBSOLETE nonsense. Active transactions must not switch table or index definitions on the fly, for several reasons, including the following: * copied indexes do not carry any history or locking information; that is, rollbacks, read views, and record locking would be broken * huge potential for race conditions, inconsistent reads and writes, loss of data, and corruption Instead of trying to track down if the table was changed during a transaction, acquire appropriate locks that protect the creation and dropping of indexes. innodb-index.test: Test the locking of CREATE INDEX and DROP INDEX. Test that consistent reads work across dropped indexes. lock_rec_insert_check_and_lock(): Relax the lock_table_has() assertion. When inserting a record into an index, the table must be at least IX-locked. However, when an index is being created, an IS-lock on the table is sufficient. row_merge_lock_table(): Add the parameter enum lock_mode mode, which must be LOCK_X or LOCK_S. row_merge_drop_table(): Assert that n_mysql_handles_opened == 0. Unconditionally drop the table. ha_innobase::add_index(): Acquire an X or S lock on the table, as appropriate. After acquiring an X lock, assert that n_mysql_handles_opened == 1. Remove the comments about dropping tables in the background. ha_innobase::final_drop_index(): Acquire an X lock on the table. dict_table_t: Remove version_number, to_be_dropped, and prebuilts. ins_node_t: Remove table_version_number. enum lock_mode: Move the definition from lock0lock.h to lock0types.h. ROW_PREBUILT_OBSOLETE, row_update_prebuilt(), row_prebuilt_table_obsolete(): Remove. row_prebuilt_t: Remove the declaration from row0types.h. row_drop_table_for_mysql_no_commit(): Always print a warning if a table was added to the background drop queue.
2007-12-17 16:49:59 +01:00
DBUG_RETURN(err);
}