- developers - lists.mariadb.org

[Maria-developers] MariaDB test question (main.cast)
by sje＠cup.hp.com 28 Oct '11

28 Oct '11

I have downloaded, built and tested MariaDB 5.3.2 Beta on HP-UX 11.31 and I get two test failures, I have a question about one of tests that I was hoping someone could help me with. The main.cast test fails and the problem is with this part of the test: select cast(cast(20010203101112.1 as double) as datetime(1)); It is converting the string "20010203101112.1" to a double value and then converting the double value to a date. The test expects "2001-02-03 10:11:12.1" to be output but I am getting "2001-02-03 10:11:12.0". I think the problem is with the conversion into floating point. When I use gdb to break sql at double_to_datetime, on HP-UX the value I have is 20010203101112.098, on Linux I see 20010203101112.102. Neither value is completely correct because 20010203101112.1 cannot be represented exactly as an IEEE floating point value. So I have two questions. Where did the string get converted into a floating point value? I couldn't find where this is happening, I assume it is using strtod or something but breaking on that function didn't do anything. Secondly, given that 20010203101112.1 cannot be represented exactly as a double precision floating point value, is this a valid test? FYI: The other failure I get is main.cast, this test is trying to check for a stack overflow before it happens but IA64 HP-UX has two stacks, a register stack and a user data stack. In this test case the recursion of p_ere and p_ere_exp wind up using only the register stack because all the local variables and arguments are stored in registers. The stack check in check_stack_overrun is checking the user stack which doesn't overflow (or even change) and so we eventually die with a register stack overflow. There is a way to check the register stack on IA64 HP-UX, but it would require using IA64 assembly language so it may not be worth doing. Steve Ellcey sje(a)cup.hp.com

3 4

Re: [Maria-developers] [Commits] Rev 3100: Fix gcc 4.6 warnings about assigned but not used variables. in file:///home/bell/maria/bzr/work-maria-5.1-applebuild/
by Michael Widenius 27 Oct '11

27 Oct '11

Hi! >>>>> "sanja" == sanja <sanja(a)askmonty.org> writes: sanja> At file:///home/bell/maria/bzr/work-maria-5.1-applebuild/ sanja> ------------------------------------------------------------ sanja> revno: 3100 sanja> revision-id: sanja(a)askmonty.org-20111025142424-dntpnipgi5qxg2io sanja> parent: wlad(a)montyprogram.com-20111019185101-fw3oxkpr2o060qk4 sanja> committer: sanja(a)askmonty.org sanja> branch nick: work-maria-5.1-applebuild sanja> timestamp: Tue 2011-10-25 17:24:24 +0300 sanja> message: sanja> Fix gcc 4.6 warnings about assigned but not used variables. sanja> Fixed my_gethwaddr.c to allow compilation on Mac OS X. <cut> === modified file 'sql/sql_show.cc' --- a/sql/sql_show.cc 2011-05-12 11:56:08 +0000 +++ b/sql/sql_show.cc 2011-10-25 14:24:24 +0000 @@ -1215,7 +1215,7 @@ int store_create_info(THD *thd, TABLE_LI handler *file= table->file; TABLE_SHARE *share= table->s; HA_CREATE_INFO create_info; - bool show_table_options= FALSE; + bool show_table_options __attribute__ ((unused))= FALSE; bool foreign_db_mode= (thd->variables.sql_mode & (MODE_POSTGRESQL | MODE_ORACLE | MODE_MSSQL | Add instead: #ifdef WITH_PARTITION_STORAGE_ENGINE around the this the definition of the variable. @@ -1432,7 +1432,9 @@ int store_create_info(THD *thd, TABLE_LI packet->append(STRING_WITH_LEN("\n)")); if (!(thd->variables.sql_mode & MODE_NO_TABLE_OPTIONS) && !foreign_db_mode) { +#ifdef WITH_PARTITION_STORAGE_ENGINE show_table_options= TRUE; +#endif The above is correct; Better to do it this way. === modified file 'storage/maria/ma_blockrec.c' --- a/storage/maria/ma_blockrec.c 2011-07-24 08:25:28 +0000 +++ b/storage/maria/ma_blockrec.c 2011-10-25 14:24:24 +0000 @@ -4997,7 +4997,7 @@ static my_bool read_row_extent_info(MARI MARIA_EXTENT_CURSOR extent; MARIA_RECORD_POS *tail_pos; uchar *data, *end_of_data; - uint flag, row_extents, row_extents_size, field_lengths; + uint flag, row_extents, row_extents_size; uchar *extents, *end; DBUG_ENTER("read_row_extent_info"); @@ -5032,9 +5032,6 @@ static my_bool read_row_extent_info(MARI } info->cur_row.extents_count= row_extents; - if (share->base.max_field_lengths) - get_key_length(field_lengths, data); - The above is wrong. get_key_length will increment data, which is required as data it's used later. Better to leave field_lengths there for the future. (Please fix this in 5.3 too !) You can avoid the warning by doing: get_key_length= get_key_length; === modified file 'storage/maria/ma_dynrec.c' --- a/storage/maria/ma_dynrec.c 2011-02-25 12:55:40 +0000 +++ b/storage/maria/ma_dynrec.c 2011-10-25 14:24:24 +0000 @@ -1741,7 +1741,8 @@ int _ma_read_rnd_dynamic_record(MARIA_HA MARIA_RECORD_POS filepos, my_bool skip_deleted_blocks) { - int block_of_record, info_read; + int block_of_record; + int info_read; Why the above? === modified file 'storage/maria/ma_ft_update.c' --- a/storage/maria/ma_ft_update.c 2009-01-30 21:55:42 +0000 +++ b/storage/maria/ma_ft_update.c 2011-10-25 14:24:24 +0000 @@ -319,7 +319,6 @@ my_bool _ma_ft_convert_to_ft2(MARIA_HA * uchar *key_ptr= (uchar*) dynamic_array_ptr(da, 0), *end; uint length, key_length; MARIA_PINNED_PAGE tmp_page_link, *page_link= &tmp_page_link; - MARIA_KEY tmp_key; MARIA_PAGE page; DBUG_ENTER("_ma_ft_convert_to_ft2"); @@ -357,13 +356,8 @@ my_bool _ma_ft_convert_to_ft2(MARIA_HA * /* inserting the rest of key values */ end= (uchar*) dynamic_array_ptr(da, da->elements); - tmp_key.keyinfo= keyinfo; - tmp_key.data_length= keyinfo->keylength; - tmp_key.ref_length= 0; - tmp_key.flag= 0; for (key_ptr+=length; key_ptr < end; key_ptr+=keyinfo->keylength) { - tmp_key.data= key_ptr; if (_ma_ck_real_write_btree(info, key, &root, SEARCH_SAME)) DBUG_RETURN(1); } Wrong. The correct fix is: if (_ma_ck_real_write_btree(info, key, &root, SEARCH_SAME)) DBUG_RETURN(1); -> if (_ma_ck_real_write_btree(info, &tmp_key, &root, SEARCH_SAME)) DBUG_RETURN(1); Apparently free textsearch with Aria is not tested properly :( <cut> === modified file 'storage/maria/ma_open.c' --- a/storage/maria/ma_open.c 2011-07-24 08:25:28 +0000 +++ b/storage/maria/ma_open.c 2011-10-25 14:24:24 +0000 @@ -1802,7 +1802,6 @@ void _ma_set_index_pagecache_callbacks(P int _ma_open_datafile(MARIA_HA *info, MARIA_SHARE *share, const char *org_name, File file_to_dup __attribute__((unused))) { - char *data_name= share->data_file_name.str; char real_data_name[FN_REFLEN]; if (org_name) @@ -1816,7 +1815,6 @@ int _ma_open_datafile(MARIA_HA *info, MA my_errno= HA_WRONG_CREATE_OPTION; return 1; } - data_name= real_data_name; } } Wrong: The correct fix is: info->dfile.file= share->bitmap.file.file= my_open(share->data_file_name.str, share->mode | O_SHARE, MYF(MY_WME)); -> info->dfile.file= share->bitmap.file.file= my_open(data_name, share->mode | O_SHARE, MYF(MY_WME)); The above means that symlinking didn't work on windows with Aria files. === modified file 'storage/maria/ma_rt_split.c' --- a/storage/maria/ma_rt_split.c 2010-09-05 23:25:44 +0000 +++ b/storage/maria/ma_rt_split.c 2011-10-25 14:24:24 +0000 @@ -380,7 +380,6 @@ int maria_rtree_split_page(const MARIA_K SplitStruct *stop; double *coord_buf; double *next_coord; - double *old_coord; int n_dim; uchar *source_cur, *cur1, *cur2; uchar *new_page_buff, *log_internal_copy, *log_internal_copy_ptr, @@ -426,7 +425,6 @@ int maria_rtree_split_page(const MARIA_K maria_rtree_d_mbr(keyinfo->seg, key->data, key_data_length, cur->coords); cur->key= key->data; - old_coord= next_coord; myisam/rt_split.c has the exact same code. You should do the same change there. === modified file 'storage/myisam/mi_check.c' --- a/storage/myisam/mi_check.c 2011-05-02 17:58:45 +0000 +++ b/storage/myisam/mi_check.c 2011-10-25 14:24:24 +0000 @@ -946,7 +946,6 @@ int chk_data_link(HA_CHECK *param, MI_IN char llbuff[22],llbuff2[22],llbuff3[22]; ha_checksum intern_record_checksum; ha_checksum key_checksum[HA_MAX_POSSIBLE_KEY]; - my_bool static_row_size; MI_KEYDEF *keyinfo; MI_BLOCK_INFO block_info; DBUG_ENTER("chk_data_link"); @@ -971,7 +970,6 @@ int chk_data_link(HA_CHECK *param, MI_IN empty=info->s->pack.header_length; /* Check how to calculate checksum of rows */ - static_row_size=1; if (info->s->data_file_type == COMPRESSED_RECORD) { for (field=0 ; field < info->s->base.fields ; field++) @@ -979,7 +977,6 @@ int chk_data_link(HA_CHECK *param, MI_IN if (info->s->rec[field].base_type == FIELD_BLOB || info->s->rec[field].base_type == FIELD_VARCHAR) { - static_row_size=0; break; } } Please remove also the if and the loop around the above code. === modified file 'storage/pbxt/src/ha_pbxt.cc' --- a/storage/pbxt/src/ha_pbxt.cc 2010-11-26 22:37:34 +0000 +++ b/storage/pbxt/src/ha_pbxt.cc 2011-10-25 14:24:24 +0000 @@ -1559,7 +1559,7 @@ static int pbxt_rollback(handlerton *hto if (!all) self->st_stat_trans = FALSE; } - return 0; + return err; } #ifdef DRIZZLED @@ -2863,7 +2863,7 @@ int ha_pbxt::update_row(const byte * old * insert into t1 (val) values (1); */ if (table->found_next_number_field && new_data == table->record[0]) { - MX_LONGLONG_T nr; + MX_LONGLONG_T nr __attribute__ ((unused)); my_bitmap_map *old_map; old_map = mx_tmp_use_all_columns(table, table->read_set); Better to remove the nr value totally (and the call that sets it). This is safe as in the code: nr = table->found_next_number_field->val_int(); ha_set_auto_increment(pb_open_tab, table->found_next_number_field); The first thing ha_set_auto_increment() is doing is calling val_int() on the second argument. <cut> === modified file 'storage/pbxt/src/thread_xt.cc' --- a/storage/pbxt/src/thread_xt.cc 2010-09-28 13:05:45 +0000 +++ b/storage/pbxt/src/thread_xt.cc 2011-10-25 14:24:24 +0000 @@ -485,7 +485,8 @@ static void thr_free_resources(XTThreadP xtPublic void xt_bug(XTThreadPtr XT_UNUSED(self)) { - static int *bug_ptr = NULL; + static int *bug_ptr __attribute__ ((unused)); + bug_ptr= NULL; bug_ptr = NULL; } No reason to set this to null twice. Better to remove the variable altogether. === modified file 'storage/pbxt/src/xactlog_xt.cc' --- a/storage/pbxt/src/xactlog_xt.cc 2010-05-05 10:59:57 +0000 +++ b/storage/pbxt/src/xactlog_xt.cc 2011-10-25 14:24:24 +0000 @@ -2117,7 +2117,7 @@ xtBool XTDatabaseLog::xlog_seq_next(XTXa size_t rec_offset; size_t max_rec_len; size_t size; - u_int check_size = 1; + u_int check_size __attribute__ ((unused))= 1; Don't understand the above change. We both set check_size and test it in the above function. Are you sure that you get a warning about the above? /* Go to the next record (xseq_record_len must be initialized * to 0 for this to work. @@ -2629,7 +2629,7 @@ static void *xlog_wr_run_thread(XTThread { XTDatabaseHPtr db = (XTDatabaseHPtr) self->t_data; int count; - void *mysql_thread; + void *mysql_thread __attribute__ ((unused)); mysql_thread = myxt_create_thread(); === modified file 'storage/xtradb/btr/btr0cur.c' --- a/storage/xtradb/btr/btr0cur.c 2011-04-29 14:16:42 +0000 +++ b/storage/xtradb/btr/btr0cur.c 2011-10-25 14:24:24 +0000 @@ -3301,7 +3301,7 @@ static void btr_record_not_null_field_in_rec( /*=============================*/ - rec_t* rec, /*!< in: physical record */ + rec_t* rec __attribute__ ((unused)),/*!< in: physical record */ ulint n_unique, /*!< in: dict_index_get_n_unique(index), number of columns uniquely determine an index entry */ @@ -3321,9 +3321,8 @@ btr_record_not_null_field_in_rec( for (i = 0; i < n_unique; i++) { ulint rec_len; - byte* field; - field = rec_get_nth_field(rec, offsets, i, &rec_len); + rec_get_nth_field_offs(offsets, i, &rec_len); if (rec_len != UNIV_SQL_NULL) { n_not_null[i]++; === modified file 'storage/xtradb/buf/buf0buf.c' --- a/storage/xtradb/buf/buf0buf.c 2011-04-29 14:16:42 +0000 +++ b/storage/xtradb/buf/buf0buf.c 2011-10-25 14:24:24 +0000 @@ -3895,7 +3895,7 @@ buf_page_io_complete( enum buf_io_fix io_type; const ibool uncompressed = (buf_page_get_state(bpage) == BUF_BLOCK_FILE_PAGE); - enum buf_flush flush_type; + //enum buf_flush flush_type; mutex_t* block_mutex; ut_a(buf_page_in_file(bpage)); @@ -4051,7 +4051,7 @@ corrupt: //buf_pool_mutex_enter(); if (io_type == BUF_IO_WRITE) { - flush_type = buf_page_get_flush_type(bpage); + //flush_type = buf_page_get_flush_type(bpage); /* to keep consistency at buf_LRU_insert_zip_clean() */ //if (flush_type == BUF_FLUSH_LRU) { /* optimistic! */ mutex_enter(&LRU_list_mutex); Remove the variable from the start of the function and enum buf_flush flush_type; to the if. That way, all the comments are in one place. Ok to push after the above changes. Regards, Monty

3 5

Re: [Maria-developers] FYI optimizer problems fixed in maria-5.5
by Sergei Petrunia 27 Oct '11

27 Oct '11

Hello! JFYI I've just pushed the fix for table_elim.test into 5.5-main. With this, I consider my involvement with 5.5 merge to be done, i.e. I don't have anything on my todo that's related to 5.5. BR Sergei -- Sergei Petrunia, Software Developer Monty Program AB, http://askmonty.org Blog: http://s.petrunia.net/blog

2 1

Re: [Maria-developers] [Commits] Rev 3249: Fix of LP BUG#872775. in file:///home/bell/maria/bzr/work-maria-5.3-lpb872775/
by Sergei Petrunia 26 Oct '11

26 Oct '11

Hi Sanja, Please add a comment around the defitions of nest_level / nest_level_base that nest_levels are local to the query or VIEW, and that view merge procedure does not re-calculate them. Ok to push after the above is addressed. On Tue, Oct 25, 2011 at 11:59:37AM +0300, sanja(a)askmonty.org wrote: > At file:///home/bell/maria/bzr/work-maria-5.3-lpb872775/ > > ------------------------------------------------------------ > revno: 3249 > revision-id: sanja(a)askmonty.org-20111025085936-h0jyyxtbv3qhp4f4 > parent: igor(a)askmonty.org-20111023124603-sjv665ajk25b4dgi > committer: sanja(a)askmonty.org > branch nick: work-maria-5.3-lpb872775 > timestamp: Tue 2011-10-25 11:59:36 +0300 > message: > Fix of LP BUG#872775. > > The problem was that merged views has its own nest_level numbering => > when we compare nest levels we should take into considiration basis (i.e. 0 level), > if it is different then nest levels are not comparable. > === modified file 'mysql-test/r/subselect_cache.result' > --- a/mysql-test/r/subselect_cache.result 2011-07-19 20:19:10 +0000 > +++ b/mysql-test/r/subselect_cache.result 2011-10-25 08:59:36 +0000 > @@ -1,3 +1,5 @@ > +drop table if exists t1,t2,t3,t4,t5; > +drop view if exists v1; > set optimizer_switch='subquery_cache=on'; > create table t1 (a int, b int); > insert into t1 values (1,2),(3,4),(1,2),(3,4),(3,4),(4,5),(4,5),(5,6),(5,6),(4,5); > @@ -3354,5 +3356,31 @@ f1 f2 f3 f3 > 7 0 0 0 > 7 0 0 0 > drop tables t1, t2, t3; > +# > +# Test of LP BUG#872775 view with "outer references" bug > +# > +set @@optimizer_switch= default; > +set optimizer_switch='subquery_cache=on'; > +CREATE TABLE t1 (a int) ; > +CREATE TABLE t2 (b int, c varchar(1) NOT NULL ) ; > +INSERT INTO t2 VALUES (1,'x'),(2,'y'); > +CREATE TABLE t3 (a int) ; > +CREATE TABLE t4 ( pk int(11) NOT NULL , b int(11) NOT NULL ) ; > +INSERT INTO t4 VALUES (26,9),(27,5),(28,0),(29,3); > +CREATE OR REPLACE VIEW v1 AS > +SELECT t2.b > +FROM t1 > +JOIN t2 > +WHERE t2 .c > ( > +SELECT t2.c FROM t3 > +); > +SELECT * FROM t4 WHERE b NOT IN ( SELECT * FROM v1 ); > +pk b > +26 9 > +27 5 > +28 0 > +29 3 > +drop view v1; > +drop table t1,t2,t3,t4; > # restore default > set @@optimizer_switch= default; > > === modified file 'mysql-test/t/subselect_cache.test' > --- a/mysql-test/t/subselect_cache.test 2011-07-19 20:19:10 +0000 > +++ b/mysql-test/t/subselect_cache.test 2011-10-25 08:59:36 +0000 > @@ -1,4 +1,10 @@ > > +--disable_warnings > +drop table if exists t1,t2,t3,t4,t5; > +drop view if exists v1; > +--enable_warnings > + > + > set optimizer_switch='subquery_cache=on'; > > create table t1 (a int, b int); > @@ -1664,5 +1670,33 @@ FROM t1 > WHERE t2.f1 OR t2.f3 ); > drop tables t1, t2, t3; > > +--echo # > +--echo # Test of LP BUG#872775 view with "outer references" bug > +--echo # > +set @@optimizer_switch= default; > +set optimizer_switch='subquery_cache=on'; > +CREATE TABLE t1 (a int) ; > + > +CREATE TABLE t2 (b int, c varchar(1) NOT NULL ) ; > +INSERT INTO t2 VALUES (1,'x'),(2,'y'); > + > +CREATE TABLE t3 (a int) ; > + > +CREATE TABLE t4 ( pk int(11) NOT NULL , b int(11) NOT NULL ) ; > +INSERT INTO t4 VALUES (26,9),(27,5),(28,0),(29,3); > + > +CREATE OR REPLACE VIEW v1 AS > +SELECT t2.b > +FROM t1 > +JOIN t2 > +WHERE t2 .c > ( > + SELECT t2.c FROM t3 > + ); > + > +SELECT * FROM t4 WHERE b NOT IN ( SELECT * FROM v1 ); > + > +drop view v1; > +drop table t1,t2,t3,t4; > + > --echo # restore default > set @@optimizer_switch= default; > > === modified file 'sql/item.cc' > --- a/sql/item.cc 2011-10-22 07:14:27 +0000 > +++ b/sql/item.cc 2011-10-25 08:59:36 +0000 > @@ -755,7 +755,9 @@ bool Item_ident::remove_dependence_proce > bool Item_ident::collect_outer_ref_processor(uchar *param) > { > Collect_deps_prm *prm= (Collect_deps_prm *)param; > - if (depended_from && depended_from->nest_level < prm->nest_level) > + if (depended_from && > + depended_from->nest_level_base == prm->nest_level_base && > + depended_from->nest_level < prm->nest_level) > prm->parameters->add_unique(this, &cmp_items); > return FALSE; > } > > === modified file 'sql/item.h' > --- a/sql/item.h 2011-10-22 07:14:27 +0000 > +++ b/sql/item.h 2011-10-25 08:59:36 +0000 > @@ -505,6 +505,7 @@ typedef void (*Cond_traverser) (const It > class Item_equal; > class COND_EQUAL; > > +class st_select_lex_unit; > > class Item { > Item(const Item &); /* Prevent use of these */ > @@ -1159,8 +1160,10 @@ public: > } > struct Collect_deps_prm > { > - int nest_level; > List<Item> *parameters; > + /* unit from which we count nest_level */ > + st_select_lex_unit *nest_level_base; > + int nest_level; > }; > /** > Collect outer references > > === modified file 'sql/item_subselect.cc' > --- a/sql/item_subselect.cc 2011-10-22 07:14:27 +0000 > +++ b/sql/item_subselect.cc 2011-10-25 08:59:36 +0000 > @@ -599,7 +599,9 @@ bool Item_subselect::exec() > > void Item_subselect::get_cache_parameters(List<Item> &parameters) > { > - Collect_deps_prm prm= { unit->first_select()->nest_level, &parameters }; > + Collect_deps_prm prm= {&parameters, > + unit->first_select()->nest_level_base, > + unit->first_select()->nest_level}; > walk(&Item::collect_outer_ref_processor, TRUE, (uchar*)&prm); > } > > > === modified file 'sql/item_sum.cc' > --- a/sql/item_sum.cc 2011-07-19 20:19:10 +0000 > +++ b/sql/item_sum.cc 2011-10-25 08:59:36 +0000 > @@ -359,7 +359,9 @@ bool Item_sum::collect_outer_ref_process > { > Collect_deps_prm *prm= (Collect_deps_prm *)param; > SELECT_LEX *ds; > - if ((ds= depended_from()) && ds->nest_level < prm->nest_level) > + if ((ds= depended_from()) && > + ds->nest_level_base == prm->nest_level_base && > + ds->nest_level < prm->nest_level) > prm->parameters->add_unique(this, &cmp_items); > return FALSE; > } > > === modified file 'sql/sql_lex.cc' > --- a/sql/sql_lex.cc 2011-10-22 07:14:27 +0000 > +++ b/sql/sql_lex.cc 2011-10-25 08:59:36 +0000 > @@ -339,6 +339,7 @@ void lex_start(THD *thd) > lex->event_parse_data= NULL; > lex->profile_options= PROFILE_NONE; > lex->nest_level=0 ; > + lex->select_lex.nest_level_base= &lex->unit; > lex->allow_sum_func= 0; > lex->in_sum_func= NULL; > lex->protect_against_global_read_lock= FALSE; > > === modified file 'sql/sql_lex.h' > --- a/sql/sql_lex.h 2011-09-22 22:13:38 +0000 > +++ b/sql/sql_lex.h 2011-10-25 08:59:36 +0000 > @@ -684,6 +684,7 @@ public: > ulong table_join_options; > uint in_sum_expr; > uint select_number; /* number of select (used for EXPLAIN) */ > + SELECT_LEX_UNIT *nest_level_base; /* unit from which we count nest_level */ > int nest_level; /* nesting level of select */ > Item_sum *inner_sum_func_list; /* list of sum func in nested selects */ > uint with_wild; /* item list contain '*' */ > > === modified file 'sql/sql_parse.cc' > --- a/sql/sql_parse.cc 2011-10-11 10:55:42 +0000 > +++ b/sql/sql_parse.cc 2011-10-25 08:59:36 +0000 > @@ -5917,6 +5917,7 @@ mysql_new_select(LEX *lex, bool move_dow > DBUG_RETURN(1); > } > select_lex->nest_level= lex->nest_level; > + select_lex->nest_level_base= &thd->lex->unit; > if (move_down) > { > SELECT_LEX_UNIT *unit; > > _______________________________________________ > commits mailing list > commits(a)mariadb.org > https://lists.askmonty.org/cgi-bin/mailman/listinfo/commits -- BR Sergey -- Sergei Petrunia, Software Developer Monty Program AB, http://askmonty.org Blog: http://s.petrunia.net/blog

1 0

Re: [Maria-developers] [Commits] Rev 3249: Fix of LP BUG#872775. in file:///home/bell/maria/bzr/work-maria-5.3-lpb872775/
by Sergei Petrunia 25 Oct '11

25 Oct '11

On Tue, Oct 25, 2011 at 11:59:37AM +0300, sanja(a)askmonty.org wrote: > At file:///home/bell/maria/bzr/work-maria-5.3-lpb872775/ > > ------------------------------------------------------------ > revno: 3249 > revision-id: sanja(a)askmonty.org-20111025085936-h0jyyxtbv3qhp4f4 > parent: igor(a)askmonty.org-20111023124603-sjv665ajk25b4dgi > committer: sanja(a)askmonty.org > branch nick: work-maria-5.3-lpb872775 > timestamp: Tue 2011-10-25 11:59:36 +0300 > message: > Fix of LP BUG#872775. > > The problem was that merged views has its own nest_level numbering => > when we compare nest levels we should take into considiration basis (i.e. 0 level), > if it is different then nest levels are not comparable. :grep nest_level_base *.h *.c* item.h|1165| st_select_lex_unit *nest_level_base; sql_lex.h|687| SELECT_LEX_UNIT *nest_level_base; /* unit from which we count nest_level */ item.cc|759| depended_from->nest_level_base == prm->nest_level_base && item_subselect.cc|603| unit->first_select()->nest_level_base, item_sum.cc|363| ds->nest_level_base == prm->nest_level_base && sql_lex.cc|342| lex->select_lex.nest_level_base= &lex->unit; sql_parse.cc|5920| select_lex->nest_level_base= &thd->lex->unit; The only places where nest_level_base is assigned are: sql_lex.cc|342| lex->select_lex.nest_level_base= &lex->unit; sql_parse.cc|5920| select_lex->nest_level_base= &thd->lex->unit; both of these initialize it to the top-level unit. So what is the point of having nest_level_base in this case? Please clarify :) > === modified file 'mysql-test/r/subselect_cache.result' > --- a/mysql-test/r/subselect_cache.result 2011-07-19 20:19:10 +0000 > +++ b/mysql-test/r/subselect_cache.result 2011-10-25 08:59:36 +0000 > @@ -1,3 +1,5 @@ > +drop table if exists t1,t2,t3,t4,t5; > +drop view if exists v1; > set optimizer_switch='subquery_cache=on'; > create table t1 (a int, b int); > insert into t1 values (1,2),(3,4),(1,2),(3,4),(3,4),(4,5),(4,5),(5,6),(5,6),(4,5); > @@ -3354,5 +3356,31 @@ f1 f2 f3 f3 > 7 0 0 0 > 7 0 0 0 > drop tables t1, t2, t3; > +# > +# Test of LP BUG#872775 view with "outer references" bug > +# > +set @@optimizer_switch= default; > +set optimizer_switch='subquery_cache=on'; > +CREATE TABLE t1 (a int) ; > +CREATE TABLE t2 (b int, c varchar(1) NOT NULL ) ; > +INSERT INTO t2 VALUES (1,'x'),(2,'y'); > +CREATE TABLE t3 (a int) ; > +CREATE TABLE t4 ( pk int(11) NOT NULL , b int(11) NOT NULL ) ; > +INSERT INTO t4 VALUES (26,9),(27,5),(28,0),(29,3); > +CREATE OR REPLACE VIEW v1 AS > +SELECT t2.b > +FROM t1 > +JOIN t2 > +WHERE t2 .c > ( > +SELECT t2.c FROM t3 > +); > +SELECT * FROM t4 WHERE b NOT IN ( SELECT * FROM v1 ); > +pk b > +26 9 > +27 5 > +28 0 > +29 3 > +drop view v1; > +drop table t1,t2,t3,t4; > # restore default > set @@optimizer_switch= default; > > === modified file 'mysql-test/t/subselect_cache.test' > --- a/mysql-test/t/subselect_cache.test 2011-07-19 20:19:10 +0000 > +++ b/mysql-test/t/subselect_cache.test 2011-10-25 08:59:36 +0000 > @@ -1,4 +1,10 @@ > > +--disable_warnings > +drop table if exists t1,t2,t3,t4,t5; > +drop view if exists v1; > +--enable_warnings > + > + > set optimizer_switch='subquery_cache=on'; > > create table t1 (a int, b int); > @@ -1664,5 +1670,33 @@ FROM t1 > WHERE t2.f1 OR t2.f3 ); > drop tables t1, t2, t3; > > +--echo # > +--echo # Test of LP BUG#872775 view with "outer references" bug > +--echo # > +set @@optimizer_switch= default; > +set optimizer_switch='subquery_cache=on'; > +CREATE TABLE t1 (a int) ; > + > +CREATE TABLE t2 (b int, c varchar(1) NOT NULL ) ; > +INSERT INTO t2 VALUES (1,'x'),(2,'y'); > + > +CREATE TABLE t3 (a int) ; > + > +CREATE TABLE t4 ( pk int(11) NOT NULL , b int(11) NOT NULL ) ; > +INSERT INTO t4 VALUES (26,9),(27,5),(28,0),(29,3); > + > +CREATE OR REPLACE VIEW v1 AS > +SELECT t2.b > +FROM t1 > +JOIN t2 > +WHERE t2 .c > ( > + SELECT t2.c FROM t3 > + ); > + > +SELECT * FROM t4 WHERE b NOT IN ( SELECT * FROM v1 ); > + > +drop view v1; > +drop table t1,t2,t3,t4; > + > --echo # restore default > set @@optimizer_switch= default; > > === modified file 'sql/item.cc' > --- a/sql/item.cc 2011-10-22 07:14:27 +0000 > +++ b/sql/item.cc 2011-10-25 08:59:36 +0000 > @@ -755,7 +755,9 @@ bool Item_ident::remove_dependence_proce > bool Item_ident::collect_outer_ref_processor(uchar *param) > { > Collect_deps_prm *prm= (Collect_deps_prm *)param; > - if (depended_from && depended_from->nest_level < prm->nest_level) > + if (depended_from && > + depended_from->nest_level_base == prm->nest_level_base && > + depended_from->nest_level < prm->nest_level) > prm->parameters->add_unique(this, &cmp_items); > return FALSE; > } > > === modified file 'sql/item.h' > --- a/sql/item.h 2011-10-22 07:14:27 +0000 > +++ b/sql/item.h 2011-10-25 08:59:36 +0000 > @@ -505,6 +505,7 @@ typedef void (*Cond_traverser) (const It > class Item_equal; > class COND_EQUAL; > > +class st_select_lex_unit; > > class Item { > Item(const Item &); /* Prevent use of these */ > @@ -1159,8 +1160,10 @@ public: > } > struct Collect_deps_prm > { > - int nest_level; > List<Item> *parameters; > + /* unit from which we count nest_level */ > + st_select_lex_unit *nest_level_base; > + int nest_level; > }; > /** > Collect outer references > > === modified file 'sql/item_subselect.cc' > --- a/sql/item_subselect.cc 2011-10-22 07:14:27 +0000 > +++ b/sql/item_subselect.cc 2011-10-25 08:59:36 +0000 > @@ -599,7 +599,9 @@ bool Item_subselect::exec() > > void Item_subselect::get_cache_parameters(List<Item> &parameters) > { > - Collect_deps_prm prm= { unit->first_select()->nest_level, &parameters }; > + Collect_deps_prm prm= {&parameters, > + unit->first_select()->nest_level_base, > + unit->first_select()->nest_level}; > walk(&Item::collect_outer_ref_processor, TRUE, (uchar*)&prm); > } > > > === modified file 'sql/item_sum.cc' > --- a/sql/item_sum.cc 2011-07-19 20:19:10 +0000 > +++ b/sql/item_sum.cc 2011-10-25 08:59:36 +0000 > @@ -359,7 +359,9 @@ bool Item_sum::collect_outer_ref_process > { > Collect_deps_prm *prm= (Collect_deps_prm *)param; > SELECT_LEX *ds; > - if ((ds= depended_from()) && ds->nest_level < prm->nest_level) > + if ((ds= depended_from()) && > + ds->nest_level_base == prm->nest_level_base && > + ds->nest_level < prm->nest_level) > prm->parameters->add_unique(this, &cmp_items); > return FALSE; > } > > === modified file 'sql/sql_lex.cc' > --- a/sql/sql_lex.cc 2011-10-22 07:14:27 +0000 > +++ b/sql/sql_lex.cc 2011-10-25 08:59:36 +0000 > @@ -339,6 +339,7 @@ void lex_start(THD *thd) > lex->event_parse_data= NULL; > lex->profile_options= PROFILE_NONE; > lex->nest_level=0 ; > + lex->select_lex.nest_level_base= &lex->unit; > lex->allow_sum_func= 0; > lex->in_sum_func= NULL; > lex->protect_against_global_read_lock= FALSE; > > === modified file 'sql/sql_lex.h' > --- a/sql/sql_lex.h 2011-09-22 22:13:38 +0000 > +++ b/sql/sql_lex.h 2011-10-25 08:59:36 +0000 > @@ -684,6 +684,7 @@ public: > ulong table_join_options; > uint in_sum_expr; > uint select_number; /* number of select (used for EXPLAIN) */ > + SELECT_LEX_UNIT *nest_level_base; /* unit from which we count nest_level */ > int nest_level; /* nesting level of select */ > Item_sum *inner_sum_func_list; /* list of sum func in nested selects */ > uint with_wild; /* item list contain '*' */ > > === modified file 'sql/sql_parse.cc' > --- a/sql/sql_parse.cc 2011-10-11 10:55:42 +0000 > +++ b/sql/sql_parse.cc 2011-10-25 08:59:36 +0000 > @@ -5917,6 +5917,7 @@ mysql_new_select(LEX *lex, bool move_dow > DBUG_RETURN(1); > } > select_lex->nest_level= lex->nest_level; > + select_lex->nest_level_base= &thd->lex->unit; > if (move_down) > { > SELECT_LEX_UNIT *unit; > BR Sergey -- Sergei Petrunia, Software Developer Monty Program AB, http://askmonty.org Blog: http://s.petrunia.net/blog

2 1

[Maria-developers] WL#243 New (by Psergey): Support >63 tables in a select
by worklog-noreply＠askmonty.org 24 Oct '11

24 Oct '11

----------------------------------------------------------------------- WORKLOG TASK -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- TASK...........: Support >63 tables in a select CREATION DATE..: Mon, 24 Oct 2011, 18:38 SUPERVISOR.....: IMPLEMENTOR....: COPIES TO......: CATEGORY.......: Server-RawIdeaBin TASK ID........: 243 (http://askmonty.org/worklog/?tid=243) VERSION........: WorkLog-4.0 STATUS.........: Un-Assigned PRIORITY.......: 60 WORKED HOURS...: 0 ESTIMATE.......: 0 (hours remain) ORIG. ESTIMATE.: 0 PROGRESS NOTES: DESCRIPTION: Quote: > Number of joins in a view >63 > MariaDB has a limit of 63 tables joined in a select statement. This > needs to be extended in order to handle the many number of attributes > that may be connected to an anchor. Preferably a much larger number than > 63 in order to be 'future proof'. ESTIMATED WORK TIME ESTIMATED COMPLETION DATE ----------------------------------------------------------------------- WorkLog (v4.0.0)

1 0

[Maria-developers] WL#242 New (by Psergey): Parameterized VIEWs
by worklog-noreply＠askmonty.org 24 Oct '11

24 Oct '11

----------------------------------------------------------------------- WORKLOG TASK -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- TASK...........: Parameterized VIEWs CREATION DATE..: Mon, 24 Oct 2011, 16:57 SUPERVISOR.....: IMPLEMENTOR....: COPIES TO......: CATEGORY.......: Server-RawIdeaBin TASK ID........: 242 (http://askmonty.org/worklog/?tid=242) VERSION........: Benchmarks-3.0 STATUS.........: Un-Assigned PRIORITY.......: 60 WORKED HOURS...: 0 ESTIMATE.......: 0 (hours remain) ORIG. ESTIMATE.: 0 PROGRESS NOTES: DESCRIPTION: Quote: > We need views which can take a specified number of parameters of a > specified type. For example, the syntax could be something like: > > CREATE VIEW exampleView (@timepoint DATE) AS > SELECT FROM WHERE colDate <= @timepoint; > > Implementation could be done as simple as doing a search and replace > with the variable and the parameter given when the view is used, such > as: > > SELECT * FROM exampleView('2001-01-01'); ESTIMATED WORK TIME ESTIMATED COMPLETION DATE ----------------------------------------------------------------------- WorkLog (v4.0.0)

1 0

Re: [Maria-developers] [Commits] Rev 3246: Fix bug lp:869036
by Sergei Golubchik 21 Oct '11

21 Oct '11

Hi, Timour! On Oct 21, timour(a)askmonty.org wrote: I don't understand, why do you need map_ptr_array (working memory provided by the caller). Looks like unnecessary complication to me. I'd do something like for (j= 0; j < bitmap_count; j++) DBUG_ASSERT(end_bit < bitmap_array[j]->n_bits); start_idx= start_bit/8/sizeof(my_bitmap_map); end_idx= end_bit/8/sizeof(my_bitmap_map); for (i= start_idx; i < end_idx; i++) { cur_res= ~0; for (j= 0; cur_res && j < bitmap_count; j++) cur_res &= bitmap_array[j]->bitmap[i]; if (cur_res) return TRUE; } cur_res= last_word_mask(end_bit); for (j= 0; cur_res && j < bitmap_count; j++) cur_res &= bitmap_array[j]->bitmap[end_idx+1]; return cur_res != 0; } and created last_word_mask() function by extracting the relevant part of create_last_word_mask() to a separate inline function. Besides, it may be that your implementation has a bug - you use a last word mask from the shortest bitmap, while you should've built it from end_bit. Now you risk taking more bits into account that end_bit dictates. > === modified file 'mysys/my_bitmap.c' > --- a/mysys/my_bitmap.c 2011-05-02 17:58:45 +0000 > +++ b/mysys/my_bitmap.c 2011-10-21 09:44:36 +0000 > @@ -410,6 +410,88 @@ void bitmap_intersect(MY_BITMAP *map, co > } > } > > + > +/* > + Check if there is some bit index between start_bit and end_bit, such that > + this is bit is set for all bitmaps in bitmap_list. > + > + SYNOPSIS > + bitmap_exists_intersection() > + bitmpap_array [in] a set of MY_BITMAPs > + map_ptr_array [in] wokring memory provided by the caller, same size as > + bitmpap_array > + bitmap_count [in] number of elements in bitmpap_array > + start_bit [in] beginning (inclusive) of the range of bits to search > + end_bit [in] end (inclusive) of the range of bits to search, must be > + no bigger than the bits of the shortest bitmap. > + > + NOTES > + This function assumes that for at least one of the bitmaps in bitmap_array all > + bits outside the range [start_bit, end_bit] are 0. As a result is not > + necessary to take care of the bits outside the range [start_bit, end_bit]. > + > + RETURN > + TRUE if an intersecion exists > + FALSE no intersection > +*/ > + > +my_bool bitmap_exists_intersection(const MY_BITMAP **bitmap_array, > + const my_bitmap_map **map_ptr_array, > + uint bitmap_count, > + uint start_bit, uint end_bit) > +{ > + const MY_BITMAP *cur_bitmap; > + my_bitmap_map cur_res, last_word_mask_inverted= 0; > + uint i, j, min_n_bits= UINT_MAX; > + uint end_word_idx= end_bit / 32; > + > + DBUG_ASSERT(bitmap_count && end_bit >= start_bit); > + for (j= 0; j < bitmap_count; j++) > + { > + cur_bitmap= bitmap_array[j]; > + DBUG_ASSERT(end_bit <= cur_bitmap->n_bits); > + /* > + Get the word that contains 'start_bit' so that intersection begins > + from that word instead of the first word of the bitmap. > + */ > + map_ptr_array[j]= cur_bitmap->bitmap + start_bit / 32; > + /* > + Find and save the last_word_mask of the shortest bitmap that > + contains end_bit in its last word. > + */ > + if (no_words_in_map(bitmap_array[j]) - 1 == end_word_idx && > + bitmap_array[j]->n_bits < min_n_bits) > + { > + min_n_bits= bitmap_array[j]->n_bits; > + last_word_mask_inverted= ~bitmap_array[j]->last_word_mask; > + } > + } > + > + /* Find the first word where the intersection of all bits is non-empty. */ > + for (i= 0; i <= end_word_idx; i++) > + { > + cur_res= 0xFFFFFFFF; > + for (j= 0; j < bitmap_count; j++) > + { > + cur_res &= *(map_ptr_array[j]); > + ++map_ptr_array[j]; > + } > + /* Clear not relevant bits using the longest last_word_mask. */ > + if (last_word_mask_inverted && i == end_word_idx) > + cur_res&= last_word_mask_inverted; > + /* > + If there is at least one non-zero bit, there was at least one sub-row > + with all bits set, that is one NULL-subrow. > + */ > + if (cur_res) > + return TRUE; > + } > + > + /* No intersection was found. */ > + return FALSE; > +} Regards, Sergei

1 0

[Maria-developers] optimizer_switch flags for next 5.3 release
by Sergei Petrunia 19 Oct '11

19 Oct '11

Hi, Written down notes from discussion about which flags should be enabled by default in the next 5.3 release: - Derived tables and keys in them should be enabled - BKA should not be enabled (doesn't make sense for 'default' query loads) - semi-joins should be enabled - Materialization (base) should be enabled - extra strategies for NULL-aware materialization - ??? - ICP should be enabled - Hash join could be enabled but in the next minor release. - Cost-based MRR should not be enabled as cost model is not tuned. when some feature is considered stable enough, its owner should apply for testing with {$current_optimizer_switch_default+new_feature}, and once that is passed, change the default in the main branch. BR Sergey -- Sergei Petrunia, Software Developer Monty Program AB, http://askmonty.org Blog: http://s.petrunia.net/blog

2 1

Re: [Maria-developers] Compiler warning about assigned but not used variables fixed
by Sergei Golubchik 18 Oct '11

18 Oct '11

Hi, Sanja! On Oct 17, sanja(a)askmonty.org wrote: > message: > Compiler warning about assigned but not used variables fixed. Pretty much ok. But please change ifdef's to __attribute__((unused)), as discussed. And see below about the commented out code, looks like you forgot to fix or uncomment it. > === modified file 'client/mysqltest.cc' > --- a/client/mysqltest.cc 2011-05-28 02:11:32 +0000 > +++ b/client/mysqltest.cc 2011-10-17 19:38:32 +0000 > @@ -5238,7 +5238,10 @@ void do_connect(struct st_command *comma > int con_port= opt_port; > char *con_options; > my_bool con_ssl= 0, con_compress= 0; > - my_bool con_pipe= 0, con_shm= 0; > + my_bool con_pipe= 0; > +#ifdef HAVE_SMEM > + my_bool con_shm= 0; > +#endif try not to overuse #ifdef's, (*) prefer __attribute__((unused)) instead. it is less error-prone and results in a more maintainable and readable code. > struct st_connection* con_slot; > > static DYNAMIC_STRING ds_connection_name; > === modified file 'mysys/ma_dyncol.c' > --- a/mysys/ma_dyncol.c 2011-09-22 09:04:00 +0000 > +++ b/mysys/ma_dyncol.c 2011-10-17 19:38:32 +0000 > @@ -1916,16 +1915,18 @@ dynamic_column_update_many(DYNAMIC_COLUM > } > > /* if (new_offset_size != offset_size) then we have to rewrite header */ > + /* > + Now we only copy. TODO: fix > + > header_delta_sign= new_offset_size - offset_size; > data_delta_sign= 0; > - copy= FALSE; > for (i= 0; i < add_column_count; i++) > { > - /* This is the check for increasing/decreasing */ > + // This is the check for increasing/decreasing > DELTA_CHECK(header_delta_sign, plan[i].hdelta, copy); > DELTA_CHECK(data_delta_sign, plan[i].ddelta, copy); > } > - > + */ why did you comment this out? > calc_param(&new_entry_size, &new_header_size, > new_offset_size, new_column_count); > > === modified file 'sql/sql_show.cc' > --- a/sql/sql_show.cc 2011-10-11 10:55:42 +0000 > +++ b/sql/sql_show.cc 2011-10-17 19:38:32 +0000 > @@ -1274,7 +1274,9 @@ int store_create_info(THD *thd, TABLE_LI > handler *file= table->file; > TABLE_SHARE *share= table->s; > HA_CREATE_INFO create_info; > +#ifdef WITH_PARTITION_STORAGE_ENGINE > bool show_table_options= FALSE; > +#endif same here, and below. > bool foreign_db_mode= (thd->variables.sql_mode & (MODE_POSTGRESQL | > MODE_ORACLE | > MODE_MSSQL | > @@ -1507,7 +1509,9 @@ int store_create_info(THD *thd, TABLE_LI > packet->append(STRING_WITH_LEN("\n)")); > if (!(thd->variables.sql_mode & MODE_NO_TABLE_OPTIONS) && !foreign_db_mode) > { > +#ifdef WITH_PARTITION_STORAGE_ENGINE > show_table_options= TRUE; > +#endif > /* > Get possible table space definitions and append them > to the CREATE TABLE statement > > === modified file 'sql/table.cc' > --- a/sql/table.cc 2011-09-02 12:10:10 +0000 > +++ b/sql/table.cc 2011-10-17 19:38:32 +0000 > @@ -5689,7 +5689,8 @@ int update_virtual_fields(THD *thd, TABL > { > DBUG_ENTER("update_virtual_fields"); > Field **vfield_ptr, *vfield; > - int error= 0; > + int error __attribute__ ((unused)); > + error= 0; FYI, you can still use initializer and __attribute__ together: int error __attribute__ ((unused)) = 0; there is no need to change the above, either way is ok. > if (!table || !table->vfield) > DBUG_RETURN(0); > > === modified file 'storage/xtradb/buf/buf0buf.c' > --- a/storage/xtradb/buf/buf0buf.c 2011-05-10 15:17:43 +0000 > +++ b/storage/xtradb/buf/buf0buf.c 2011-10-17 19:38:32 +0000 > @@ -3895,7 +3895,7 @@ buf_page_io_complete( > enum buf_io_fix io_type; > const ibool uncompressed = (buf_page_get_state(bpage) > == BUF_BLOCK_FILE_PAGE); > - enum buf_flush flush_type; > + //enum buf_flush flush_type; please send all your XtraDB changes to Percona > mutex_t* block_mutex; > Regards, Sergei (*) see for example http://www.google.com/search?q=ifdef+considered+harmful

2 2