October 2020 - developers - lists.mariadb.org

Re: [Maria-developers] 2830b9c290e: MDEV-23926: Fix warnings generated during compilation of plugin/auth_pam/mapper/pam_user_map.c on MacOS
by Sergei Golubchik 22 Oct '20

22 Oct '20

Hi, Dmitry! On Oct 22, Dmitry Shulga wrote: > revision-id: 2830b9c290e (mariadb-10.2.31-489-g2830b9c290e) > parent(s): 8984d779103 > author: Dmitry Shulga <dmitry.shulga(a)mariadb.com> > committer: Dmitry Shulga <dmitry.shulga(a)mariadb.com> > timestamp: 2020-10-22 00:42:10 +0700 > message: > > MDEV-23926: Fix warnings generated during compilation of plugin/auth_pam/mapper/pam_user_map.c on MacOS > > Compiler warnings like one listed below are generated during server build on MacOS: > > > --- > plugin/auth_pam/CMakeLists.txt | 26 ++++++++++++++++++++++++++ > plugin/auth_pam/config.h.cmake | 1 + > plugin/auth_pam/mapper/pam_user_map.c | 30 +++++++++++++++++++----------- > 3 files changed, 46 insertions(+), 11 deletions(-) Nice! Could you also - in a second follow-up commit, not to confuse with the pure bug fix - do a small cleanup and move all other checks from plugin/auth_pam/CMakeLists.txt into your new config.h.cmake ? Currently they're added to the command line, like IF(HAVE_PAM_EXT_H) ADD_DEFINITIONS(-DHAVE_PAM_EXT_H) ENDIF() but as you've introduced config.h.cmake, there's no need to put them all on the command line anymore. You'll need to include config_auth_pam.h into more source files. And, I suppose. add config_auth_pam.h to .gitignore too. ok to push these two commits when done. Thanks! Regards, Sergei VP of MariaDB Server Engineering and security(a)mariadb.org

1 0

Re: [Maria-developers] 65af1d07771: MDEV-23178: Qualified asterisk not supported in INSERT .. RETURNING
by Sergei Golubchik 22 Oct '20

22 Oct '20

Hi, Rucha! On Oct 22, Rucha Deodhar wrote: > revision-id: 65af1d07771 (mariadb-10.5.4-108-g65af1d07771) > parent(s): bbd70fcc43c > author: Rucha Deodhar <rucha.deodhar(a)mariadb.com> > committer: Rucha Deodhar <rucha.deodhar(a)mariadb.com> > timestamp: 2020-08-06 17:23:49 +0530 > message: > > MDEV-23178: Qualified asterisk not supported in INSERT .. RETURNING > > Anaylsis: When we have INSERT/REPLACE...RETURNING and have qualified asterisk, > table_name is not NULL and context->table_list is either NULL or has > incorrect reference because context->table_list has tables from the FROM clause. > context->table_list has incorrect reference (has table from the FROM clause > instead of table we are inserting into) for INSERT/REPLACE...SELECT...RETURNING > because we have a FROM clause from the SELECT statement. > For INSERT/REPLACE...RETURNING it is NULL because there is no FROM clause. > Fix: If table_name is not null, check if we have INSERT/REPLACE...RETURNING. > If so, the reference should be the table we are inserting into and not > the table in the FROM clause. > > diff --git a/mysql-test/main/insert_returning.result b/mysql-test/main/insert_returning.result > index e664e02bedc..e1bab6f0329 100644 > --- a/mysql-test/main/insert_returning.result > +++ b/mysql-test/main/insert_returning.result > @@ -89,6 +89,9 @@ total val1 id1 && id1 id1 UPPER(val1) f(id1) good tests! > diff --git a/sql/sql_base.cc b/sql/sql_base.cc > index 45ce4be3eb5..4ce7b034029 100644 > --- a/sql/sql_base.cc > +++ b/sql/sql_base.cc > @@ -7959,7 +7959,7 @@ insert_fields(THD *thd, Name_resolution_context *context, const char *db_name, > bool any_privileges, uint *hidden_bit_fields) > { > Field_iterator_table_ref field_iterator; > - bool found; > + bool found, is_insert_or_replace_returning= false; > char name_buff[SAFE_NAME_LEN+1]; > DBUG_ENTER("insert_fields"); > DBUG_PRINT("arena", ("stmt arena: %p",thd->stmt_arena)); > @@ -7978,13 +7978,31 @@ insert_fields(THD *thd, Name_resolution_context *context, const char *db_name, > > found= FALSE; > > + if ((thd->lex->sql_command == SQLCOM_INSERT_SELECT || > + thd->lex->sql_command == SQLCOM_REPLACE_SELECT || > + thd->lex->sql_command == SQLCOM_INSERT || > + thd->lex->sql_command == SQLCOM_REPLACE) && > + thd->lex->has_returning()) > + is_insert_or_replace_returning= true; What about INSERT t1 SELECT t2.* FROM t2 RETURNING t1.*; can your is_insert_or_replace_returning distinguish between the first and the second asterisk? Regards, Sergei VP of MariaDB Server Engineering and security(a)mariadb.org

1 0

Re: [Maria-developers] 01612f594b3: MDEV-23769: MTR can abort before it prints the test result summary
by Sergei Golubchik 22 Oct '20

22 Oct '20

Hi, Anel! On Oct 22, Anel Husakovic wrote: > revision-id: 01612f594b3 (mariadb-10.1.43-300-g01612f594b3) > parent(s): 82301aea4f2 > author: Anel Husakovic <anel(a)mariadb.org> > committer: Anel Husakovic <anel(a)mariadb.org> > timestamp: 2020-10-03 02:34:03 +0200 > message: > > MDEV-23769: MTR can abort before it prints the test result summary > > - Patch is solving generating report on warning > - Patch is doing cosmetic fix of `current_test` log file which holds the old log value of test `CURRENT TEST:..` in `mark_log()` in case of `unknown option` and as such > the logic which is using it's content doesn't output valid log content and doesn't generate valid `$test->{'comment'}` message. more is not always better. please remove everything in the commit comment below this line - 400 lines that don't explain anything beyond what your *one line* of the comment already did. And now you have to press PageDown many times to get to my main comment about the fix :) > Here are the test cases observed. Note that only in 6. scenario there is `print $child "HELLO\n";` problem in printing (to delete?). > > 1. Valid test case, but warning in server log > ``` > $ ./mtr MDEV-23769 --force > Logging: ./mtr MDEV-23769 --force > vardir: /home/anel/mariadb/10.1-mtr-in-source/mysql-test/var > Checking leftover processes... > Removing old var directory... > Creating var directory '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var'... > Checking supported features... > MariaDB Version 10.1.47-MariaDB-debug > - SSL connections supported > - binaries are debug compiled > Sphinx 'indexer' binary not found, sphinx suite will be skipped > Collecting tests... > Installing system database... > > ============================================================================== > > TEST RESULT TIME (ms) or COMMENT > -------------------------------------------------------------------------- > > worker[1] Using MTR_BUILD_THREAD 300, with reserved ports 16000..16019 > show databases; > Database > information_schema > mtr > mysql > performance_schema > test > main.MDEV-23769 [ fail ] Found warnings/errors in server log file! > Test ended at 2020-10-03 02:13:00 > line > 2020-10-03 2:13:00 140309996242496 [Warning] option 'lock_wait_timeout': unsigned value 18446744073709551615 adjusted to 31536000 > ^ Found warnings in /home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/mysqld.1.err > ok > > - saving '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.MDEV-23769/' to '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.MDEV-23769/' > -------------------------------------------------------------------------- > The servers were restarted 0 times > Spent 0.000 of 4 seconds executing testcases > mysql-test-run: WARNING: Got errors/warnings while running tests, please examine '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/warnings' for details. > > Completed: Failed 1/1 tests, 0.00% were successful. > > Failing test(s): main.MDEV-23769 > > The log files in var/log may give you some hint of what went wrong. > > If you want to report this error, please read first the documentation > at http://dev.mysql.com/doc/mysql/en/mysql-test-suite.html > > mysql-test-run: *** ERROR: there were failing test cases > ``` > > 2. Valid test cases (same), but warning in server log (Note reported single test case) > ``` > $ ./mtr MDEV-23769 MDEV-23769 --force > Logging: ./mtr MDEV-23769 MDEV-23769 --force > vardir: /home/anel/mariadb/10.1-mtr-in-source/mysql-test/var > Checking leftover processes... > Removing old var directory... > Creating var directory '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var'... > Checking supported features... > MariaDB Version 10.1.47-MariaDB-debug > - SSL connections supported > - binaries are debug compiled > Sphinx 'indexer' binary not found, sphinx suite will be skipped > Collecting tests... > Installing system database... > > ============================================================================== > > TEST RESULT TIME (ms) or COMMENT > -------------------------------------------------------------------------- > > worker[1] Using MTR_BUILD_THREAD 300, with reserved ports 16000..16019 > show databases; > Database > information_schema > mtr > mysql > performance_schema > test > main.MDEV-23769 [ fail ] Found warnings/errors in server log file! > Test ended at 2020-10-03 02:15:05 > line > 2020-10-03 2:15:04 139868136609344 [Warning] option 'lock_wait_timeout': unsigned value 18446744073709551615 adjusted to 31536000 > ^ Found warnings in /home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/mysqld.1.err > ok > > - saving '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.MDEV-23769/' to '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.MDEV-23769/' > show databases; > Database > information_schema > mtr > mysql > performance_schema > test > main.MDEV-23769 [ fail ] Found warnings/errors in server log file! > Test ended at 2020-10-03 02:15:05 > line > 2020-10-03 2:15:05 140513782085184 [Warning] option 'lock_wait_timeout': unsigned value 18446744073709551615 adjusted to 31536000 > ^ Found warnings in /home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/mysqld.1.err > ok > > - saving '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.MDEV-23769/' to '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.MDEV-23769/' > -------------------------------------------------------------------------- > The servers were restarted 1 times > Spent 0.000 of 5 seconds executing testcases > mysql-test-run: WARNING: Got errors/warnings while running tests, please examine '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/warnings' for details. > > Completed: Failed 2/2 tests, 0.00% were successful. > > Failing test(s): main.MDEV-23769 > > The log files in var/log may give you some hint of what went wrong. > > If you want to report this error, please read first the documentation > at http://dev.mysql.com/doc/mysql/en/mysql-test-suite.html > > mysql-test-run: *** ERROR: there were failing test cases > ``` > > 3. Valid test cases (different test cases), but warning in server log (Note one passed one failed) > > ``` > $ ./mtr MDEV-23769 1st --force > Logging: ./mtr MDEV-23769 1st --force > vardir: /home/anel/mariadb/10.1-mtr-in-source/mysql-test/var > Checking leftover processes... > Removing old var directory... > Creating var directory '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var'... > Checking supported features... > MariaDB Version 10.1.47-MariaDB-debug > - SSL connections supported > - binaries are debug compiled > Sphinx 'indexer' binary not found, sphinx suite will be skipped > Collecting tests... > Installing system database... > > ============================================================================== > > TEST RESULT TIME (ms) or COMMENT > -------------------------------------------------------------------------- > > worker[1] Using MTR_BUILD_THREAD 300, with reserved ports 16000..16019 > show databases; > Database > information_schema > mtr > mysql > performance_schema > test > main.MDEV-23769 [ fail ] Found warnings/errors in server log file! > Test ended at 2020-10-03 02:17:09 > line > 2020-10-03 2:17:08 140658805515840 [Warning] option 'lock_wait_timeout': unsigned value 18446744073709551615 adjusted to 31536000 > ^ Found warnings in /home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/mysqld.1.err > ok > > - saving '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.MDEV-23769/' to '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.MDEV-23769/' > main.1st [ pass ] 1 > -------------------------------------------------------------------------- > The servers were restarted 1 times > Spent 0.001 of 4 seconds executing testcases > mysql-test-run: WARNING: Got errors/warnings while running tests, please examine '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/warnings' for details. > > Completed: Failed 1/2 tests, 50.00% were successful. > > Failing test(s): main.MDEV-23769 > > The log files in var/log may give you some hint of what went wrong. > > If you want to report this error, please read first the documentation > at http://dev.mysql.com/doc/mysql/en/mysql-test-suite.html > > mysql-test-run: *** ERROR: there were failing test cases > > ``` > > 4. Now for example 3. create artificial failure with warnings in server log (note, both test failed) > > ``` > > $ ./mtr MDEV-23769 1st --force > Logging: ./mtr MDEV-23769 1st --force > vardir: /home/anel/mariadb/10.1-mtr-in-source/mysql-test/var > Checking leftover processes... > Removing old var directory... > Creating var directory '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var'... > Checking supported features... > MariaDB Version 10.1.47-MariaDB-debug > - SSL connections supported > - binaries are debug compiled > Sphinx 'indexer' binary not found, sphinx suite will be skipped > Collecting tests... > Installing system database... > > ============================================================================== > > TEST RESULT TIME (ms) or COMMENT > -------------------------------------------------------------------------- > > worker[1] Using MTR_BUILD_THREAD 300, with reserved ports 16000..16019 > show databases; > Database > information_schema > mtr > mysql > performance_schema > test > main.MDEV-23769 [ fail ] Found warnings/errors in server log file! > Test ended at 2020-10-03 02:17:52 > line > 2020-10-03 2:17:51 140260493149760 [Warning] option 'lock_wait_timeout': unsigned value 18446744073709551615 adjusted to 31536000 > ^ Found warnings in /home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/mysqld.1.err > ok > > - saving '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.MDEV-23769/' to '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.MDEV-23769/' > main.1st [ fail ] > Test ended at 2020-10-03 02:17:52 > > CURRENT_TEST: main.1st > mysqltest: At line 5: End of line junk detected: "show tables in mysql > " > > The result from queries just before the failure was: > show databases; > Database > information_schema > mtr > mysql > performance_schema > test > > - saving '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.1st/' to '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.1st/' > -------------------------------------------------------------------------- > The servers were restarted 1 times > Spent 0.000 of 4 seconds executing testcases > mysql-test-run: WARNING: Got errors/warnings while running tests, please examine '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/warnings' for details. > > Completed: Failed 2/2 tests, 0.00% were successful. > > Failing test(s): main.MDEV-23769 main.1st > > The log files in var/log may give you some hint of what went wrong. > > If you want to report this error, please read first the documentation > at http://dev.mysql.com/doc/mysql/en/mysql-test-suite.html > > mysql-test-run: *** ERROR: there were failing test cases > > ``` > > 5. Generate unknown option, for the same example as 3. (note: both failed) > ``` > > $ ./mtr MDEV-23769 1st --mysqld=--xx --force > Logging: ./mtr MDEV-23769 1st --mysqld=--xx --force > vardir: /home/anel/mariadb/10.1-mtr-in-source/mysql-test/var > Checking leftover processes... > Removing old var directory... > Creating var directory '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var'... > Checking supported features... > MariaDB Version 10.1.47-MariaDB-debug > - SSL connections supported > - binaries are debug compiled > Sphinx 'indexer' binary not found, sphinx suite will be skipped > Collecting tests... > Installing system database... > > ============================================================================== > > TEST RESULT TIME (ms) or COMMENT > -------------------------------------------------------------------------- > > worker[1] Using MTR_BUILD_THREAD 300, with reserved ports 16000..16019 > worker[1] mysql-test-run: WARNING: Process [mysqld.1 - pid: 52531, winpid: 52531, exit: 512] died after mysql-test-run waited 0.3 seconds for /home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/run/mysqld.1.pid to be created. > main.MDEV-23769 [ fail ] > Test ended at 2020-10-03 02:19:01 > > CURRENT_TEST: main.MDEV-23769 > > Failed to start mysqld.1 > mysqltest failed but provided no output > > - saving '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.MDEV-23769/' to '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.MDEV-23769/' > ***Warnings generated in error logs during shutdown after running tests: main.MDEV-23769 > > 2020-10-03 2:19:01 140238937720384 [Warning] option 'lock_wait_timeout': unsigned value 18446744073709551615 adjusted to 31536000 > 2020-10-03 2:19:01 140238937720384 [ERROR] /home/anel/mariadb/10.1-mtr-in-source/sql/mysqld: unknown option '--xx' > 2020-10-03 2:19:01 140238937720384 [ERROR] Aborting > > worker[1] mysql-test-run: WARNING: Process [mysqld.1 - pid: 52537, winpid: 52537, exit: 512] died after mysql-test-run waited 0.3 seconds for /home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/run/mysqld.1.pid to be created. > main.1st [ fail ] > Test ended at 2020-10-03 02:19:02 > > CURRENT_TEST: main.1st > > Failed to start mysqld.1 > mysqltest failed but provided no output > > - saving '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.1st/' to '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.1st/' > ***Warnings generated in error logs during shutdown after running tests: main.1st > > 2020-10-03 2:19:02 140048201700928 [ERROR] /home/anel/mariadb/10.1-mtr-in-source/sql/mysqld: unknown option '--xx' > 2020-10-03 2:19:02 140048201700928 [ERROR] Aborting > > Can't use an undefined value as a symbol reference at ./mtr line 657. > -------------------------------------------------------------------------- > The servers were restarted 1 times > Spent 0.000 of 5 seconds executing testcases > > Completed: Failed 2/2 tests, 0.00% were successful. > > Failing test(s): main.MDEV-23769 main.1st > > The log files in var/log may give you some hint of what went wrong. > > If you want to report this error, please read first the documentation > at http://dev.mysql.com/doc/mysql/en/mysql-test-suite.html > > Errors/warnings were found in logfiles during server shutdown after running the > following sequence(s) of tests: > main.MDEV-23769 > main.1st > mysql-test-run: *** ERROR: there were failing test cases > > ``` > > 6. Again the same option, but with 2 same tests -> not perl error: `Can't use an undefined value as a symbol reference at ./mtr line 657. > ` ` print $child "HELLO\n";` (solution delete the line). > > ``` > $ ./mtr MDEV-23769 MDEV-23769 --mysqld=--xx --force > Logging: ./mtr MDEV-23769 MDEV-23769 --mysqld=--xx --force > vardir: /home/anel/mariadb/10.1-mtr-in-source/mysql-test/var > Checking leftover processes... > Removing old var directory... > Creating var directory '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var'... > Checking supported features... > MariaDB Version 10.1.47-MariaDB-debug > - SSL connections supported > - binaries are debug compiled > Sphinx 'indexer' binary not found, sphinx suite will be skipped > Collecting tests... > Installing system database... > > ============================================================================== > > TEST RESULT TIME (ms) or COMMENT > -------------------------------------------------------------------------- > > worker[1] Using MTR_BUILD_THREAD 300, with reserved ports 16000..16019 > worker[1] mysql-test-run: WARNING: Process [mysqld.1 - pid: 52595, winpid: 52595, exit: 512] died after mysql-test-run waited 0.3 seconds for /home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/run/mysqld.1.pid to be created. > main.MDEV-23769 [ fail ] > Test ended at 2020-10-03 02:20:12 > > CURRENT_TEST: main.MDEV-23769 > > Failed to start mysqld.1 > mysqltest failed but provided no output > > - saving '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.MDEV-23769/' to '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.MDEV-23769/' > ***Warnings generated in error logs during shutdown after running tests: main.MDEV-23769 > > 2020-10-03 2:20:12 139815724283456 [Warning] option 'lock_wait_timeout': unsigned value 18446744073709551615 adjusted to 31536000 > 2020-10-03 2:20:12 139815724283456 [ERROR] /home/anel/mariadb/10.1-mtr-in-source/sql/mysqld: unknown option '--xx' > 2020-10-03 2:20:12 139815724283456 [ERROR] Aborting > > worker[1] mysql-test-run: WARNING: Process [mysqld.1 - pid: 52601, winpid: 52601, exit: 512] died after mysql-test-run waited 0.3 seconds for /home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/run/mysqld.1.pid to be created. > main.MDEV-23769 [ fail ] > Test ended at 2020-10-03 02:20:12 > > CURRENT_TEST: main.MDEV-23769 > > Failed to start mysqld.1 > mysqltest failed but provided no output > > - saving '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.MDEV-23769/' to '/home/anel/mariadb/10.1-mtr-in-source/mysql-test/var/log/main.MDEV-23769/' > ***Warnings generated in error logs during shutdown after running tests: main.MDEV-23769 > > 2020-10-03 2:20:12 140650797745728 [Warning] option 'lock_wait_timeout': unsigned value 18446744073709551615 adjusted to 31536000 > 2020-10-03 2:20:12 140650797745728 [ERROR] /home/anel/mariadb/10.1-mtr-in-source/sql/mysqld: unknown option '--xx' > 2020-10-03 2:20:12 140650797745728 [ERROR] Aborting > > Can't use an undefined value as a symbol reference at ./mtr line 657. > -------------------------------------------------------------------------- > The servers were restarted 1 times > Spent 0.000 of 5 seconds executing testcases > > Completed: Failed 2/2 tests, 0.00% were successful. > > Failing test(s): main.MDEV-23769 > > The log files in var/log may give you some hint of what went wrong. > > If you want to report this error, please read first the documentation > at http://dev.mysql.com/doc/mysql/en/mysql-test-suite.html > > Errors/warnings were found in logfiles during server shutdown after running the > following sequence(s) of tests: > main.MDEV-23769 > main.MDEV-23769 > mysql-test-run: *** ERROR: there were failing test cases > > ``` > > diff --git a/mysql-test/disabled.def b/mysql-test/disabled.def > index ee30c4f4d3c..fdb62a7e5f5 100644 > --- a/mysql-test/disabled.def > +++ b/mysql-test/disabled.def > @@ -20,3 +20,4 @@ ssl_crl_clrpath : broken upstream > file_contents : MDEV-6526 these files are not installed anymore > max_statement_time : cannot possibly work, depends on timing > partition_open_files_limit : open_files_limit check broken by MDEV-18360 > +MDEV-23769 : Intentionally disabled, for validating purposes. I don't know if it makes any sense. A dead test that nobody will ever run. May be it's better not to add it at all? In the past I've added tests like that too, the "mtr" and "mtr2" suites. I regret it now :) Tests only help if they're run. > diff --git a/mysql-test/mysql-test-run.pl b/mysql-test/mysql-test-run.pl > index 52920530132..8d761fd70cf 100755 > --- a/mysql-test/mysql-test-run.pl > +++ b/mysql-test/mysql-test-run.pl > @@ -1005,6 +1005,13 @@ sub run_worker ($) { > stop_servers(reverse all_servers()); > if(check_warnings_post_shutdown($server)) { > # Warnings appeared in log file(s) during final server shutdown. > + # Make sure first to run statistics > + my $tests= collect_test_cases($opt_reorder, $opt_suites, \@opt_cases, \@opt_skip_test_list); > + my ($prefix, $fail, $completed, $extra_warnings)= > + run_test_server($server, $tests, $opt_parallel); > + mtr_print_line(); > + print_total_times($opt_parallel) if $opt_report_times; > + mtr_report_stats($prefix, $fail, $completed, $extra_warnings); I don't understand the fix. You're saying that if a warning happens on shutdown then mtr won't print the result summary. This doesn't seem to be the case, I've tried few times to repeat it - modified mtr to always pretend there's a warning on shutdown, and modified mysqld to always issue a warning on shutdown - and in all cases I saw the result summary all right. > exit(1); > } > } > @@ -3971,6 +3978,7 @@ sub run_testcase ($$) { > if (start_servers($tinfo)) > { > report_failure_and_restart($tinfo); > + unlink $path_current_testlog; > return 1; > } > } > diff --git a/mysql-test/t/MDEV-23769.opt b/mysql-test/t/MDEV-23769.opt > new file mode 100644 > index 00000000000..fd8dbcab53f > --- /dev/null > +++ b/mysql-test/t/MDEV-23769.opt > @@ -0,0 +1 @@ > +--lock-wait-timeout=-1 > diff --git a/mysql-test/t/MDEV-23769.test b/mysql-test/t/MDEV-23769.test > new file mode 100644 > index 00000000000..9938c7cf644 > --- /dev/null > +++ b/mysql-test/t/MDEV-23769.test > @@ -0,0 +1,4 @@ > +--echo # > +--echo # MDEV-23769: MTR can abort before it prints the test result summary > +--echo # > +show databases; > Regards, Sergei VP of MariaDB Server Engineering and security(a)mariadb.org

1 0

Re: [Maria-developers] 8c745467560: MDEV-23836: Assertion `! is_set() || m_can_overwrite_status' in
by Sergei Golubchik 21 Oct '20

21 Oct '20

Hi, Rucha! what happens in this bug? Why does the server need to open_table() after my_ok()? On Oct 21, Rucha Deodhar wrote: > revision-id: 8c745467560 (mariadb-10.5.4-226-g8c745467560) > parent(s): 308f8350c7b > author: Rucha Deodhar <rucha.deodhar(a)mariadb.com> > committer: Rucha Deodhar <rucha.deodhar(a)mariadb.com> > timestamp: 2020-10-20 11:53:22 +0530 > message: > > MDEV-23836: Assertion `! is_set() || m_can_overwrite_status' in > Diagnostics_area::set_error_status (interrupted ALTER TABLE under LOCK) > > Analysis: Assertion failure happens because we don't ignore kill query > even if 'OK' is already sent. > Fix: Check if we already sent 'OK'. If sent, ignore kill query. > > diff --git a/mysql-test/main/alter_table.result b/mysql-test/main/alter_table.result > index 197e52b35ca..204bfa5249c 100644 > --- a/mysql-test/main/alter_table.result > +++ b/mysql-test/main/alter_table.result > @@ -3322,5 +3322,23 @@ Note 1176 Key 'x' doesn't exist in table 't1' > unlock tables; > drop table t1; > # > +# MDEV-23836: Assertion `! is_set() || m_can_overwrite_status' in > +# Diagnostics_area::set_error_status (interrupted ALTER TABLE under LOCK) > +# > +set @max_session_mem_used_save= @@max_session_mem_used; > +CREATE TABLE t1 (a INT); > +SELECT * FROM t1; > +a > +ALTER TABLE x MODIFY xx INT; > +ERROR 42S02: Table 'test.x' doesn't exist > +SET SESSION max_session_mem_used= 8192; > +LOCK TABLE t1 WRITE; > +ALTER TABLE t1 CHANGE COLUMN IF EXISTS b c INT; > +Warnings: > +Note 1054 Unknown column 'b' in 't1' > +set SESSION max_session_mem_used = @max_session_mem_used_save; > +UNLOCK TABLES; > +DROP TABLE t1; > +# > # End of 10.5 tests > # > diff --git a/mysql-test/main/alter_table.test b/mysql-test/main/alter_table.test > index 6fe61a3222a..d09c4378339 100644 > --- a/mysql-test/main/alter_table.test > +++ b/mysql-test/main/alter_table.test > @@ -2529,6 +2529,27 @@ alter table t1 rename key if exists x to xx; > unlock tables; > drop table t1; > > +--echo # > +--echo # MDEV-23836: Assertion `! is_set() || m_can_overwrite_status' in > +--echo # Diagnostics_area::set_error_status (interrupted ALTER TABLE under LOCK) > +--echo # > + > +set @max_session_mem_used_save= @@max_session_mem_used; > + > +CREATE TABLE t1 (a INT); > +SELECT * FROM t1; > + > +--error ER_NO_SUCH_TABLE > +ALTER TABLE x MODIFY xx INT; > + > +SET SESSION max_session_mem_used= 8192; > +LOCK TABLE t1 WRITE; > +ALTER TABLE t1 CHANGE COLUMN IF EXISTS b c INT; > + > +set SESSION max_session_mem_used = @max_session_mem_used_save; > +UNLOCK TABLES; > +DROP TABLE t1; > + > --echo # > --echo # End of 10.5 tests > --echo # > diff --git a/sql/sql_base.cc b/sql/sql_base.cc > index e6d0ab13e75..e8d39917ba6 100644 > --- a/sql/sql_base.cc > +++ b/sql/sql_base.cc > @@ -1682,7 +1682,9 @@ bool open_table(THD *thd, TABLE_LIST *table_list, Open_table_context *ot_ctx) > > if (!(flags & MYSQL_OPEN_IGNORE_KILLED) && thd->killed) > { > - thd->send_kill_message(); > + /* If we already sent 'ok', we can ignore any kill query statements */ > + if (! thd->get_stmt_da()->is_set()) > + thd->send_kill_message(); > DBUG_RETURN(TRUE); > } > Regards, Sergei VP of MariaDB Server Engineering and security(a)mariadb.org

1 0

Re: [Maria-developers] 1060cc4235a: MDEV-23500: Assertion `!is_set() || (m_status == DA_OK_BULK && is_bulk_op())'
by Sergei Golubchik 21 Oct '20

21 Oct '20

Hi, Rucha! On Oct 21, Rucha Deodhar wrote: > revision-id: 1060cc4235a (mariadb-10.2.31-438-g1060cc4235a) > parent(s): 80075ba011f > author: Rucha Deodhar <rucha.deodhar(a)mariadb.com> > committer: Rucha Deodhar <rucha.deodhar(a)mariadb.com> > timestamp: 2020-09-22 17:12:57 +0530 > message: > > MDEV-23500: Assertion `!is_set() || (m_status == DA_OK_BULK && is_bulk_op())' > failed in Diagnostics_area::set_ok_status upon multi-table update > > Analysis: When NULL field is copied to NOT-NULL so there is error but > this error is not reported so we get the failure. > Fix: Return the error state. > > diff --git a/mysql-test/r/update.result b/mysql-test/r/update.result > index 9e19abc4e9c..c1d37fc36e7 100644 > --- a/mysql-test/r/update.result > +++ b/mysql-test/r/update.result > @@ -731,4 +731,14 @@ INSERT INTO t2 VALUES > UPDATE t1,t2 SET t1.i1 = -39 WHERE t2.d1 <> t1.i1 AND t2.d1 = t1.d2; > ERROR 22007: Incorrect datetime value: '19' for column `test`.`t1`.`i1` at row 1 > DROP TABLE t1,t2; > +# > +# MDEV-23500: Assertion `!is_set() || (m_status == DA_OK_BULK && > +# is_bulk_op())' failed in Diagnostics_area::set_ok_status upon > +# multi-table update > +# > +CREATE TABLE t (f INT NOT NULL); > +INSERT INTO t VALUES (0),(0); > +UPDATE t AS t1 LEFT JOIN t AS t2 ON t1.f <=> t2.f SET t2.f = NULL; > +ERROR 01000: Data truncated for column 'f' at row 4 with a single-table update I get MariaDB [test]> update t set f=NULL; ERROR 1048 (23000): Column 'f' cannot be null should multi-table update produce the same error? > +DROP TABLE t; Regards, Sergei VP of MariaDB Server Engineering and security(a)mariadb.org

1 0

Re: [Maria-developers] 9c2570e6c42: MDEV-22422: Assertion `! is_set()' failed in Diagnostics_area::set_eof_status
by Sergei Golubchik 21 Oct '20

21 Oct '20

Hi, Rucha! On Oct 21, Rucha Deodhar wrote: > revision-id: 9c2570e6c42 (mariadb-10.1.43-279-g9c2570e6c42) > parent(s): 5768f57d243 > author: Rucha Deodhar <rucha.deodhar(a)mariadb.com> > committer: Rucha Deodhar <rucha.deodhar(a)mariadb.com> > timestamp: 2020-09-22 17:18:37 +0530 > message: > > MDEV-22422: Assertion `! is_set()' failed in Diagnostics_area::set_eof_status > > Analysis: The error is not returned when the statement can't be used. > And so we still go on to search for keywords. > Fix: Return the error state. > > diff --git a/mysql-test/t/errors.test b/mysql-test/t/errors.test > index 55461002fd4..e2b10559299 100644 > --- a/mysql-test/t/errors.test > +++ b/mysql-test/t/errors.test > @@ -213,3 +213,22 @@ select * from seq_1_to_1000; > --enable_result_log > # We may not be able to execute any more queries with this connection > # because of too little memory# > + > +--echo # > +--echo # 10.1 Test > +--echo # > +--echo # MDEV-22422: Assertion `! is_set()' failed in > +--echo # Diagnostics_area::set_eof_status > +--echo # > + > +SET @saved_max_session_mem_used = @@max_session_mem_used; > + > +SET max_session_mem_used = 50000; > +--error ER_OPTION_PREVENTS_STATEMENT > +help 'it is going to crash'; > + > +SET @@max_session_mem_used=@saved_max_session_mem_used; > + > +--echo # > +--echo # End of 10.1 Test > +--echo # > diff --git a/sql/sql_help.cc b/sql/sql_help.cc > index cd9da566e51..654a3915b29 100644 > --- a/sql/sql_help.cc > +++ b/sql/sql_help.cc > @@ -736,6 +736,9 @@ static bool mysqld_help_internal(THD *thd, const char *mask) > &name, &description, &example); > delete select; > > + if (thd->is_error()) > + goto error; does the error happen during the first count_topics= search_topics() ? There are lots of places below that look kind of fragile too. May be they're safe. Or may be by increasing max_session_mem_used somewhat you can get mysqld_help_internal to crash below again and again? Could you look into that please? > if (count_topics == 0) > { > int UNINIT_VAR(key_id); > Regards, Sergei VP of MariaDB Server Engineering and security(a)mariadb.org

1 0

Re: [Maria-developers] 6729bf5c031: MDEV-22313: Neither SHOW CREATE USER nor SHOW GRANTS prints a user's default role
by Sergei Golubchik 21 Oct '20

21 Oct '20

Hi, Anel! On Oct 21, Anel Husakovic wrote: > revision-id: 6729bf5c031 (mariadb-10.1.43-112-g6729bf5c031) > parent(s): ad4b70562bb > author: Anel Husakovic <anel(a)mariadb.org> > committer: Anel Husakovic <anel(a)mariadb.org> > timestamp: 2020-04-22 20:48:55 +0200 > message: > > MDEV-22313: Neither SHOW CREATE USER nor SHOW GRANTS prints a user's default role > > diff --git a/mysql-test/t/grant5.test b/mysql-test/t/grant5.test > index 649bba7d1ca..5eefd9b9185 100644 > --- a/mysql-test/t/grant5.test > +++ b/mysql-test/t/grant5.test > @@ -33,3 +33,16 @@ REVOKE EXECUTE ON PROCEDURE sp FROM u; > --error ER_TABLE_NOT_LOCKED > REVOKE PROCESS ON *.* FROM u; > DROP TABLE t1; > + > +# > +# MDEV-22313: Neither SHOW CREATE USER nor SHOW GRANTS prints a user's default role > +# > +CREATE ROLE test_role; > +CREATE USER test_user; > +GRANT test_role TO test_user; > +SET DEFAULT ROLE test_role FOR test_user; > +SHOW GRANTS FOR test_user; > +SET DEFAULT ROLE NONE for test_user; > +SHOW GRANTS FOR test_user; add a test for SHOW GRANTS without FOR username > +DROP USER test_user; > +DROP ROLE test_role; > diff --git a/sql/sql_acl.cc b/sql/sql_acl.cc > index b6a6f806e50..90dd552ec43 100644 > --- a/sql/sql_acl.cc > @@ -7891,6 +7893,10 @@ bool mysql_show_grants(THD *thd, LEX_USER *lex_user) > if (show_role_grants(thd, username, hostname, acl_user, buff, sizeof(buff))) > goto end; > > + /* Show default role to acl_user */ > + if (show_default_role(thd, username, hostname, acl_user, buff, sizeof(buff))) > + goto end; Please, print SET DEFAULT ROLE after all grants for this user. > + > /* Add first global access grants */ > if (show_global_privileges(thd, acl_user, FALSE, buff, sizeof(buff))) > goto end; > @@ -7963,6 +7969,42 @@ static ROLE_GRANT_PAIR *find_role_grant_pair(const LEX_STRING *u, > my_hash_search(&acl_roles_mappings, (uchar*)pair_key.ptr(), key_length); > } > > +static bool show_default_role(THD *thd, const char *username, > + const char *hostname, ACL_USER_BASE *acl_entry, why did you cast acl_entry (which you know is ACL_USER) to the base class ACL_USER_BASE? why did you pass the username if you don't use it? > + char *buff, size_t buffsize) > +{ > + Protocol *protocol= thd->protocol; > + LEX_STRING host= {const_cast<char*>(hostname), strlen(hostname)}; This doesn't make much sense. You only create this LEX_STRING to use it for String::append(LEX_STRING *). But you can directly use String::append(const char*) and you won't need any LEX_STRING here. Furthermore, you can print acl_entry->host.hostname just as you print acl_entry->user, so you don't need the above hostname argument at all. > + String def_role(buff,sizeof(buff),system_charset_info); > + def_role.length(0); you don't use def_role anywhere. > + LEX_STRING def_rolename= ((ACL_USER *)acl_entry)->default_rolename; > + if (def_rolename.length) > + { > + String def_str(buff,sizeof(buff),system_charset_info); > + def_str.length(0); > + def_str.append("SET DEFAULT ROLE "); prefer to use STRING_WITH_LEN() for string literals. E.g. here it should be def_str.append(STRING_WITH_LEN("SET DEFAULT ROLE ")); > + def_str.append(def_rolename.str, def_rolename.length, > + system_charset_info); here you can do def_str.append(&def_rolename); > + def_str.append(" FOR '"); > + def_str.append(acl_entry->user.str, acl_entry->user.length, > + system_charset_info); > + if (!(acl_entry->flags & IS_ROLE)) better put an assert here not an if > + { > + def_str.append(STRING_WITH_LEN("'@'")); > + def_str.append(&host); > + } > + def_str.append('\''); > + protocol->prepare_for_resend(); > + protocol->store(def_str.ptr(),def_str.length(),def_str.charset()); > + if (protocol->write()) > + { > + return TRUE; > + } > + } > + return FALSE; > +} > + > static bool show_role_grants(THD *thd, const char *username, > const char *hostname, ACL_USER_BASE *acl_entry, > char *buff, size_t buffsize) > Regards, Sergei VP of MariaDB Server Engineering and security(a)mariadb.org

1 0

Re: [Maria-developers] 8c34a242aa2: MDEV-20787: Script dgcov.pl does not work
by Sergei Golubchik 21 Oct '20

21 Oct '20

Hi, Anel! On Oct 21, Anel Husakovic wrote: > revision-id: 8c34a242aa2 (mariadb-10.2.31-478-g8c34a242aa2) > parent(s): 01ffccd6a4e > author: Anel Husakovic <anel(a)mariadb.org> > committer: Anel Husakovic <anel(a)mariadb.org> > timestamp: 2020-10-06 00:00:58 +0200 > message: > > MDEV-20787: Script dgcov.pl does not work > > - This patch is done after doing test on 00f9df29224ada9cdbb796c99f4 > where difference was found between versions of gcov. don't refer to commits that are not pushed and are not going to be pushed. Because if you push this one you'll have a commit forever in the history that refers to an non-existing commit. It isn't helpful. > - This patch doesn't take gcov versions in account since is extracting > filename of a `.gcov` file, which is generated with `gcov -i > <source-file>.gcda` during `dgcov.pl --generate`. > - There is a check if file exists. > Reason for that is because some generated `gcov` files are outliers from above > rule (that a file generated in intermediate format(gcov) has the same source name). > Example 1: `aestables.cpp.gcda` is not generating `aestables.cpp.gcov`, > but only the files used in #include `modes.hpp.gcov runtime.hpp.gcov` > Example 2: `my_sha256.cc.gcda` is generating `my_sha.ic.gcov` > - Make sure to test `MTR_BINDIR` which gets set during out-of-source > build and generate error when running `./mysql-test/mtr --gcov alias` What is a problem with out of source builds? If they don't work, perhaps it would be better to fix dgcov to work with out of source builds? > - During the testing: > 10.1 results are stored in > - `<basedir>//mysql-test-gcov.msg` and errors in > - `<basedir>//mysql-test-gcov.err` > 10.2 results should be stored in > `./mysql-test/var/last_changes.dgcov`, but only empty results are > obtained (tests `1st` and `alias`). I don't understand this ^^^ paragraph. And also because your patch is 10.2+ so who cares what it does on 10.1? > diff --git a/mysql-test/dgcov.pl b/mysql-test/dgcov.pl > index fbc5540e697..f2bbdad4d20 100755 > --- a/mysql-test/dgcov.pl > +++ b/mysql-test/dgcov.pl > @@ -161,8 +161,14 @@ sub gcov_one_file { > system($cmd)==0 or die "system($cmd): $? $!"; > } > > - # now, read the generated file > - open FH, '<', "$_.gcov" or die "open(<$_.gcov): $!"; > + (my $filename = $_)=~ s/\.[^.]+$//; # remove extension In your previous patch you either removed the extension or not based on gcov version. Why did you abandon this logic? > + my $gcov_file_path= $File::Find::dir."/$filename.gcov"; File::Find documentation seems to be saying that $File::Find::dir ends with a slash (/) so you don't need to prefix $filename.gcov with it. > + if (! -f $gcov_file_path) > + { > + return; > + } may be return unless -f $gcov_file_path it's perl after all :) > + open FH, '<', "$gcov_file_path" or die "open(<$gcov_file_path): $!"; you don't need double quotes around "$gcov_file_path" - they don't serve any purpose. > + > my $fname; > while (<FH>) { > chomp; Regards, Sergei VP of MariaDB Server Engineering and security(a)mariadb.org

1 0

Re: [Maria-developers] f83c1615d97: MDEV-20207: Assertion `! is_set()' failed in Diagnostics_area::set_eof_status
by Sergei Golubchik 21 Oct '20

21 Oct '20

Hi, Rucha! On Oct 21, Rucha Deodhar wrote: > revision-id: f83c1615d97 (mariadb-10.1.43-240-gf83c1615d97) > parent(s): 7e9ffc69ecd > author: Rucha Deodhar <rucha.deodhar(a)mariadb.com> > committer: Rucha Deodhar <rucha.deodhar(a)mariadb.com> > timestamp: 2020-08-05 00:33:03 +0530 > message: > > MDEV-20207: Assertion `! is_set()' failed in Diagnostics_area::set_eof_status > upon HANDLER READ > > Analysis: The error state is not stored while checking condition and key name. > Fix: Return true while checking condition and key name if error is reported > because geometry object can't be created from the data in the index value for > HANDLER READ. > > diff --git a/mysql-test/suite/handler/aria.test b/mysql-test/suite/handler/aria.test > index 1913d2b791c..50ef43ce5b2 100644 > --- a/mysql-test/suite/handler/aria.test > +++ b/mysql-test/suite/handler/aria.test > @@ -80,3 +80,23 @@ HANDLER t1 CLOSE; > DROP TABLE t1; > > --echo End of 5.1 tests > + > +--echo # > +--echo # 10.1 Test > +--echo # > +--echo # MDEV-20207: Assertion `! is_set()' failed in > +--echo # Diagnostics_area::set_eof_status upon HANDLER READ > +--echo # > + > +DROP TABLE IF EXISTS t1; there's no need to start the test from DROP TABLE IF EXISTS t1; historically people customarily did that. before mtr had after-test checks. but now all tests are supposed to clean up after itself, you do not need to worry about tests leaving tables around. > + > +CREATE TABLE t1 (a POINT, KEY(a)); > +HANDLER t1 OPEN h; > + > +--error ER_CANT_CREATE_GEOMETRY_OBJECT > +HANDLER h READ a = (0); I don't think it should be an error. SELECT * FROM t1 WHERE a=(0) -- this does not fail, it just doesn't find anything. I think HANDLER should do the same. > + > +HANDLER h CLOSE; > +DROP TABLE t1; note that all handler test files source handler.inc. that is, if you want a test to run for all engines, you do not need to add it to all individual test files, add it to handler.inc Regards, Sergei VP of MariaDB Server Engineering and security(a)mariadb.org

1 0

Re: [Maria-developers] e3f45b2f9ea: MDEV-10267 Add ngram fulltext parser plugin
by Sergei Golubchik 21 Oct '20

21 Oct '20

Hi, Rinat! On Oct 21, Rinat Ibragimov wrote: > > > + > > > + // Avoid creating partial n-grams by stopping at word boundaries. > > > + // Underscore is treated as a word character too, since identifiers in > > > + // programming languages often contain them. > > > + if (!is_alnum(char_type) && !is_underscore(cs, next, end)) { > > > + start= next + char_len; > > > + next= start; > > > + n_chars= 0; > > > + continue; > > > > I don't think it's a good idea. On the opposite, I would rather create > > ngrams over word boundaries, so that, say, I'd split "n-gram plugin" to > > > > "n-g", "-gr", "gra", "ram", "am ", "m p", " pl", "plu", "lug", "ugi", "gin". > > I see how this can increase index size: more different strings are > included. Implementation will become a tiny bit simpler. But can it > make searching better somehow? Yes, the idea is that it'll rank "One can use n-gram plugin in MariaDB fulltext search" higher than "Plugins are useful for extending functionality of a program" because some ngrams encode that "n-gram" and "plugin" words must be together as a phrase. > I can't decide. From my point of view, the current approach is fine. > Please pick a variant, and I'll try to implement that. No, I cannot guess which approach will produce more relevant searches. Implement something and then we test what works better > > > + } > > > + > > > + next += char_len; > > > + n_chars++; > > > + > > > + if (n_chars == ngram_token_size) { > > > + bool_info->position= static_cast<uint>(start - doc); > > > > nope, we don't have MYSQL_FTPARSER_FULL_BOOLEAN_INFO::position and this > > plugin is definitely not the reason to break the API and add it. > > There is a patch in the patchset that adds the "position" field. Without > that field, > it's possible to trigger an assertion in InnoDB's FTS code by executing: > > INSTALL PLUGIN simple_parser SONAME 'mypluglib'; > CREATE TABLE articles( > a TEXT DEFAULT NULL, > b TEXT DEFAULT NULL, > FULLTEXT (a, b) WITH PARSER simple_parser > ) ENGINE=InnoDB; > INSERT INTO articles VALUES ('111', '1234 1234 1234'); > DROP TABLE articles; > UNINSTALL PLUGIN simple_parser; > > Same assertion may be triggered by any fts plugin. Are you sure you > want that patch to be removed? I don't think InnoDB assert is a reason to extend the fulltext plugin API. What about this fix instead: diff --git a/storage/innobase/fts/fts0fts.cc b/storage/innobase/fts/fts0fts.cc --- a/storage/innobase/fts/fts0fts.cc +++ b/storage/innobase/fts/fts0fts.cc @@ -4669,7 +4669,7 @@ fts_tokenize_add_word_for_parser( ut_ad(boolean_info->position >= 0); position = boolean_info->position + fts_param->add_pos; */ - position = fts_param->add_pos; + position = fts_param->add_pos++; fts_add_token(result_doc, str, position); this fixed the crash for me. > > > + n_chars= ngram_token_size - 1; > > > + is_first= false; > > > + } > > > + } > > > + > > > + // If nothing was emitted yet, emit the remainder. > > > > what reminder, what is it for? > > It's "remainder", with "a" before "i". Remainder of the processed > string. That way strings that are too short for having even a single > n-gram, will still generate some index entries and will be > discoverable by exact queries. Perhaps the comment should say it? That if the string too short, it'll be emitted here? > > > + case MYSQL_FTPARSER_FULL_BOOLEAN_INFO: > > > + // Reusing InnoDB's boolean mode parser to handle all special characters. > > > + while (fts_get_word(cs, &start, end, &word, &bool_info)) { > > > > nope. InnoDB might not even be compiled it. Use param->mysql_parse() > > Copied fts_get_word() implementation into the plugin code. > > As far as I understand, param->mysql_parse() cannot be used here as it > splits a string into words like default fulltext search does. This > plugin however must generate n-grams. Of course, it can. Note that fts_get_word() doesn't generate n-grams either, it gets the whole word and the n-gram plugin later splits it into n-grams. Similarly param->mysql_parse() will extract words for you and you'll split them into n-grams. Regards, Sergei VP of MariaDB Server Engineering and security(a)mariadb.org

1 0