- developers - lists.mariadb.org

[Maria-developers] High Priest of Humility
by md＠rpzdesign.com 08 Jan '13

08 Jan '13

Arjen: I would not go so far as to call it "raving" or boasting about anything. It was a private email message to you only. Not broadcast publicly. For a MariaDB captain, that is pretty bold offering a negative opinion to a potential contributor. Negative comments like that are not the best support for any level of enthusiasm for openness, collaboration, and sharing. Hint: Stop giving hints. marco On 1/7/2013 5:05 PM, Arjen Lentz wrote: > Hi Marco > >> Nice to see your response. >> >> I will be providing the code that I use to modify the MariaDB core >> daemon soon. > I presume you'll put it up in a branch on Launchpad? > > >> You will be amazed how little I am changing the core code. > Hint: modesty/humility are a good traits for an OSS developer. Let other people be amazed and impressed when you deliver code - raving about yourself beforehand is less cool... > > >> Where is the open source for Galera? (could not find it readily) > At Codership, but also in the 5.5 Galera repository of MariaDB. > > > Regards, > Arjen. > >

1 0

[Maria-developers] Missing man/mysql_plugin.1 in tar ball
by Honza Horak 07 Jan '13

07 Jan '13

Hi all, I've noticed that man page for mysql_plugin utility (man/mysql_plugin.1) is not present in mariadb-5.5.28a.tar.gz while it is present in mysql-5.5.28.tar.gz. However, I failed to find the reason why in case there is one. Or is that a mistake? Does anybody know? Regards, Honza

2 1

Re: [Maria-developers] API Documentaion for Replication Features
by Lixun Peng 31 Dec '12

31 Dec '12

Hi Monty, On Tue, Dec 25, 2012 at 8:21 PM, Michael Widenius < michael.widenius(a)gmail.com> wrote: > > Hi! > > Sorry that I missed this email before. > <cut> > Lixun> I can try to implement it. Are you doing it? > > Please try to implemement it? > You can look at the keycache code how we have done this before! > > OK, I'm trying to do it. Thanks, Lixun -- Senior MySQL Developer @ Taobao.com Mobile Phone: +86 18658156856 (Hangzhou) Gtalk: penglixun(at)gmail.com Twitter: http://www.twitter.com/plinux Blog: http://www.penglixun.com

2 1

Re: [Maria-developers] Rev 3480: Post review changes (total (really))
by Sergei Golubchik 24 Dec '12

24 Dec '12

Hi, Oleksandr! On Dec 23, Oleksandr Byelkin wrote: > 22.12.2012 21:28, Sergei Golubchik пишет: > >> @@ -100,6 +101,8 @@ struct st_dynamic_column_value > >> > >> typedef struct st_dynamic_column_value DYNAMIC_COLUMN_VALUE; > >> > >> +/* old functions (deprecated) */ > > I usually prefer code to comments. Kind of a comment that a compiler > > can make use of. That means an assert(aaa == 0) instead of /* aaa > > should be 0 here */ and instead of comment above, I'd rather have > > __attribute__((deprecated)) > > __attribute__((deprecated)) is GCC, we also use MS compiler. So ifdef > is better until we get appropriate multi-platform macro. 1. __attribute__ is defined to be empty for other compilers (somewhere in my_global.h, I guess), so it won't break the compilation. 2. I did not mean to use it instead of the #ifdef, I mean to use it instead of a /* old functions (deprecated) */ comment. Because a compiler cannot read the comment. 3. And anyway, it was just a suggestion, you can keep your current code and not change anything. As you prefer. > everything else will be fixed. Thanks! Regards, Sergei

1 0

Re: [Maria-developers] Rev 3480: Post review changes (total (really))
by Sergei Golubchik 22 Dec '12

22 Dec '12

Hi, Sanja! See below. Just a couple of small API-related comments. And one JSON bug. On Dec 21, sanja(a)askmonty.org wrote: > At file:///home/bell/maria/bzr/work-maria-10.0-cassandra/ > > ------------------------------------------------------------ > revno: 3480 > revision-id: sanja(a)askmonty.org-20121221192211-93tc13j3a6c85qea > parent: sanja(a)montyprogram.com-20121119121604-5h5tu0zn11em0sb3 > committer: sanja(a)askmonty.org > branch nick: work-maria-10.0-cassandra > timestamp: Fri 2012-12-21 21:22:11 +0200 > message: > Post review changes (total (really)). > > === modified file 'include/ma_dyncol.h' > --- a/include/ma_dyncol.h 2012-09-28 12:27:16 +0000 > +++ b/include/ma_dyncol.h 2012-12-21 19:22:11 +0000 > @@ -38,12 +38,13 @@ > how the offset are stored. > */ > #define MAX_DYNAMIC_COLUMN_LENGTH 0X1FFFFFFFL > +#define MAX_DYNAMIC_COLUMN_LENGTH_NM 0XFFFFFFFFFL Hm. Two questions. 1. Where does the number come from? 2. You've introduced a constant, but you aren't using it anywhere. What's the point? > /* > Limits of implementation > */ > -#define MAX_NAME_LENGTH 255 > #define MAX_TOTAL_NAME_LENGTH 65535 > +#define MAX_NAME_LENGTH (MAX_TOTAL_NAME_LENGTH/4) > > /* NO and OK is the same used just to show semantics */ > #define ER_DYNCOL_NO ER_DYNCOL_OK > @@ -100,6 +101,8 @@ struct st_dynamic_column_value > > typedef struct st_dynamic_column_value DYNAMIC_COLUMN_VALUE; > > +/* old functions (deprecated) */ I usually prefer code to comments. Kind of a comment that a compiler can make use of. That means an assert(aaa == 0) instead of /* aaa should be 0 here */ and instead of comment above, I'd rather have __attribute__((deprecated)) > +#ifdef MADYNCOL_DEPRECATED > enum enum_dyncol_func_result > dynamic_column_create(DYNAMIC_COLUMN *str, > uint column_nr, DYNAMIC_COLUMN_VALUE *value); > @@ -133,73 +121,105 @@ dynamic_column_update_many(DYNAMIC_COLUM ... > enum enum_dyncol_func_result > -dynamic_column_exists_fmt(DYNAMIC_COLUMN *str, void *key, my_bool string_keys); > +dynamic_column_get(DYNAMIC_COLUMN *org, uint column_nr, > + DYNAMIC_COLUMN_VALUE *store_it_here); > +#endif > > -/* List of not NULL columns */ > +/* new functions */ > enum enum_dyncol_func_result > -dynamic_column_list(DYNAMIC_COLUMN *org, DYNAMIC_ARRAY *array_of_uint); > +mariadb_dyncol_create_many(DYNAMIC_COLUMN *str, > + uint column_count, > + uint *column_numbers, > + DYNAMIC_COLUMN_VALUE *values, > + my_bool new_string); > enum enum_dyncol_func_result > -dynamic_column_list_str(DYNAMIC_COLUMN *str, DYNAMIC_ARRAY *array_of_lexstr); > +mariadb_dyncol_create_many_named(DYNAMIC_COLUMN *str, > + uint column_count, > + LEX_STRING *column_keys, > + DYNAMIC_COLUMN_VALUE *values, > + my_bool new_string); > + > + > enum enum_dyncol_func_result > -dynamic_column_list_fmt(DYNAMIC_COLUMN *str, DYNAMIC_ARRAY *array, my_bool string_keys); > +mariadb_dyncol_update_many(DYNAMIC_COLUMN *str, > + uint add_column_count, > + uint *column_keys, > + DYNAMIC_COLUMN_VALUE *values); > +enum enum_dyncol_func_result > +mariadb_dyncol_update_many_named(DYNAMIC_COLUMN *str, > + uint add_column_count, > + LEX_STRING *column_keys, > + DYNAMIC_COLUMN_VALUE *values); > + > + > +enum enum_dyncol_func_result > +mariadb_dyncol_exists(DYNAMIC_COLUMN *org, uint column_nr); > +enum enum_dyncol_func_result > +mariadb_dyncol_exists_named(DYNAMIC_COLUMN *str, LEX_STRING *name); > + > +/* List of not NULL columns */ > +enum enum_dyncol_func_result > +mariadb_dyncol_list(DYNAMIC_COLUMN *org, DYNAMIC_ARRAY *array_of_uint); Sanja, let's be constistent. If mariadb_dyncol_list_named() returns an array of LEX_STRING's, than mariadb_dyncol_list should return an array of integers, not a dynamic array. Besides you don't use DYNAMIC_ARRAY *anywhere* else in the new (not deprecated) API. > +enum enum_dyncol_func_result > +mariadb_dyncol_list_named(DYNAMIC_COLUMN *str, uint *count, LEX_STRING **names); > > /* > if the column do not exists it is NULL > */ > enum enum_dyncol_func_result > -dynamic_column_get(DYNAMIC_COLUMN *org, uint column_nr, > +mariadb_dyncol_get(DYNAMIC_COLUMN *org, uint column_nr, > DYNAMIC_COLUMN_VALUE *store_it_here); > enum enum_dyncol_func_result > -dynamic_column_get_str(DYNAMIC_COLUMN *str, LEX_STRING *name, > - DYNAMIC_COLUMN_VALUE *store_it_here); > +mariadb_dyncol_get_named(DYNAMIC_COLUMN *str, LEX_STRING *name, > + DYNAMIC_COLUMN_VALUE *store_it_here); > > -my_bool dynamic_column_has_names(DYNAMIC_COLUMN *str); > +my_bool mariadb_dyncol_has_names(DYNAMIC_COLUMN *str); > > enum enum_dyncol_func_result > -dynamic_column_check(DYNAMIC_COLUMN *str); > +mariadb_dyncol_check(DYNAMIC_COLUMN *str); > > enum enum_dyncol_func_result > -dynamic_column_json(DYNAMIC_COLUMN *str, DYNAMIC_STRING *json); > +mariadb_dyncol_json(DYNAMIC_COLUMN *str, DYNAMIC_STRING *json); > > #define dynamic_column_initialize(A) memset((A), 0, sizeof(*(A))) > #define dynamic_column_column_free(V) dynstr_free(V) > > /* conversion of values to 3 base types */ > enum enum_dyncol_func_result > -dynamic_column_val_str(DYNAMIC_STRING *str, DYNAMIC_COLUMN_VALUE *val, > +mariadb_dyncol_val_str(DYNAMIC_STRING *str, DYNAMIC_COLUMN_VALUE *val, > CHARSET_INFO *cs, my_bool quote); > enum enum_dyncol_func_result > -dynamic_column_val_long(longlong *ll, DYNAMIC_COLUMN_VALUE *val); > +mariadb_dyncol_val_long(longlong *ll, DYNAMIC_COLUMN_VALUE *val); > +enum enum_dyncol_func_result > +mariadb_dyncol_val_double(double *dbl, DYNAMIC_COLUMN_VALUE *val); > + > + > enum enum_dyncol_func_result > -dynamic_column_val_double(double *dbl, DYNAMIC_COLUMN_VALUE *val); > +mariadb_dyncol_unpack(DYNAMIC_COLUMN *str, > + uint *count, > + LEX_STRING **names, DYNAMIC_COLUMN_VALUE **vals); > > +int mariadb_dyncol_column_cmp_named(const LEX_STRING *s1, const LEX_STRING *s2); > enum enum_dyncol_func_result > -dynamic_column_vals(DYNAMIC_COLUMN *str, > - DYNAMIC_ARRAY *names, DYNAMIC_ARRAY *vals, > - char **free_names); > +mariadb_dyncol_column_count(DYNAMIC_COLUMN *str, uint *column_count); > > /*************************************************************************** > Internal functions, don't use if you don't know what you are doing... > ***************************************************************************/ > > -#define dynamic_column_reassociate(V,P,L, A) dynstr_reassociate((V),(P),(L),(A)) > +#define mariadb_dyncol_reassociate(V,P,L, A) dynstr_reassociate((V),(P),(L),(A)) I'd remove it. You can use dynstr_reassociate() directly in Item_func_dyncol_create::val_str(), if you need it, but don't encourage others to do it, by pussing this macro here. > -#define dynamic_column_value_init(V) (V)->type= DYN_COL_NULL > +#define dyncol_value_init(V) (V)->type= DYN_COL_NULL 1. you forgot the mariadb_ prefix 2. This macro seems to be unused. Perhaps you can simply delete it? > === modified file 'mysql-test/r/dyncol.result' > --- a/mysql-test/r/dyncol.result 2012-09-28 11:01:17 +0000 > +++ b/mysql-test/r/dyncol.result 2012-12-21 19:22:11 +0000 > @@ -1577,3 +1587,53 @@ COLUMN_CHECK('') > SELECT COLUMN_CHECK(NULL); > COLUMN_CHECK(NULL) > NULL > +# > +# escaping check > +# > +select column_json(column_create("string", "'\"/\\`.,whatever")),hex(column_create("string", "'\"/\\`.,whatever")); > +column_json(column_create("string", "'\"/\\`.,whatever")) hex(column_create("string", "'\"/\\`.,whatever")) > +[{"string":"'\"/\\`.,whatever"}] 040100060000000300737472696E670827222F5C602E2C7768617465766572 Please, add a test with two values. What it will look like: [{"a":1},{"b":b}] ? or { "a" : 1, "b" : 2 } ? I don't see a reason why it should be an array of one-value objects, instead of one object, with many values. But perhaps I'm missing something, why did you prefer an array? > +# > +# embedding test > +# > +select column_json(column_create("val", "val", "emb", column_create("val2", "val2"))); > +column_json(column_create("val", "val", "emb", column_create("val2", "val2"))) > +[{"emb":[{"val2":"val2"}],{"val":"val"}] ouch. invalid json, curly braces aren't properly paired. > +select column_json(column_create(1, "val", 2, column_create(3, "val2"))); > +column_json(column_create(1, "val", 2, column_create(3, "val2"))) > +[{"1":"val"},{"2":[{"3":"val2"}]] same here. > === modified file 'sql/item_func.h' > --- a/sql/item_func.h 2012-09-27 23:06:56 +0000 > +++ b/sql/item_func.h 2012-12-21 19:22:11 +0000 > @@ -58,7 +58,7 @@ public: > NOW_FUNC, TRIG_COND_FUNC, > SUSERVAR_FUNC, GUSERVAR_FUNC, COLLATE_FUNC, > EXTRACT_FUNC, CHAR_TYPECAST_FUNC, FUNC_SP, UDF_FUNC, > - NEG_FUNC, GSYSVAR_FUNC }; > + NEG_FUNC, GSYSVAR_FUNC, DYNCOL }; DYNCOL_FUNC > enum optimize_type { OPTIMIZE_NONE,OPTIMIZE_KEY,OPTIMIZE_OP, OPTIMIZE_NULL, > OPTIMIZE_EQUAL }; > enum Type type() const { return FUNC_ITEM; } Regards, Sergei

1 0

[Maria-developers] MariaDB Galera 5.5.28a Release Prep
by Daniel Bartholomew 21 Dec '12

21 Dec '12

All, looks like my first message on this never reached the list, so sending again. I've prepped the release notes and changelog pages for this release and will be activating it in a few hours (barring any show stoppers). - https://kb.askmonty.org/en/mariadb-galera-5528a-release-notes/ - https://kb.askmonty.org/en/mariadb-galera-5528a-changelog/ I've updated the "What is MariaDB Galera Cluster?" and "Getting Started..." pages. If other updates are needed, let me know. https://kb.askmonty.org/en/what-is-mariadb-galera-cluster/ https://kb.askmonty.org/en/getting-started-with-mariadb-galera-cluster/ Also please let me know if updates are needed on the "About" and "Known Limitations" pages https://kb.askmonty.org/en/about-galera-replication/ https://kb.askmonty.org/en/mariadb-galera-cluster-known-limitations/ Thanks. -- Daniel Bartholomew Google+ - http://gplus.to/dbart Twitter - http://twitter.com/daniel_bart MariaDB: An Enhanced Drop-in Replacement for MySQL Website - http://mariadb.org Twitter - http://twitter.com/mariadb Google+ - http://google.com/+mariadb Facebook - http://fb.com/MariaDB.dbms Knowledgebase - http://kb.askmonty.org Monty Program - http://montyprogram.com

1 0

Re: [Maria-developers] [Commits] Rev 3484: postreview fixes (part 5)
by Sergei Golubchik 21 Dec '12

21 Dec '12

Hi, Sanja! On Dec 21, sanja(a)askmonty.org wrote: > At file:///home/bell/maria/bzr/work-maria-10.0-cassandra/ > ------------------------------------------------------------ > revno: 3484 > revision-id: sanja(a)askmonty.org-20121221010609-cnr2z3fg99kpxzma > committer: sanja(a)askmonty.org > branch nick: work-maria-10.0-cassandra > timestamp: Fri 2012-12-21 03:06:09 +0200 > message: > postreview fixes (part 5) > === modified file 'include/ma_dyncol.h' > --- a/include/ma_dyncol.h 2012-12-20 23:35:38 +0000 > +++ b/include/ma_dyncol.h 2012-12-21 01:06:09 +0000 > @@ -169,7 +169,8 @@ madyncol_exists_named(DYNAMIC_COLUMN *st > enum enum_dyncol_func_result > madyncol_list(DYNAMIC_COLUMN *org, DYNAMIC_ARRAY *array_of_uint); > enum enum_dyncol_func_result > -madyncol_list_named(DYNAMIC_COLUMN *str, DYNAMIC_ARRAY *array_of_lexstr); > +madyncol_list_named(DYNAMIC_COLUMN *str, DYNAMIC_ARRAY *array_of_lexstr, > + void **memory_to_free); Same here. You can put this memory in the dynarray and don't burden the caller with these nuisanses. Or you can use a normal array, not a dynarray. > /* > if the column do not exists it is NULL > > === modified file 'mysql-test/r/dyncol.result' > --- a/mysql-test/r/dyncol.result 2012-12-20 23:48:20 +0000 > +++ b/mysql-test/r/dyncol.result 2012-12-21 01:06:09 +0000 > @@ -1587,3 +1587,7 @@ COLUMN_CHECK('') > SELECT COLUMN_CHECK(NULL); > COLUMN_CHECK(NULL) > NULL > +# quating check "quoting" But in fact, it's not quoting, it's escaping. > +select column_json(column_create("string", "'\"/\\`.,whatever")),hex(column_create("string", "'\"/\\`.,whatever")); > +column_json(column_create("string", "'\"/\\`.,whatever")) hex(column_create("string", "'\"/\\`.,whatever")) > +[{"string":"'\"/\\`.,whatever"}] 040100060000000300737472696E670827222F5C602E2C7768617465766572 > Regards, Sergei

1 0

[Maria-developers] Odd Error in Logs
by Gordan Bobic 21 Dec '12

21 Dec '12

Hi, I recently switched from MySQL 5.6.5 to MariaDB 10.0, and I've been seeing errors like the following in the logs since: 121221 13:22:34 [ERROR] Master 'master1': Found index PRIMARY whose column info does not match that of MySQL. 121221 13:22:34 [ERROR] Master 'master1': Build InnoDB index translation table for Table ./data/aggregate#P#p201006 failed I didn't dump/restore the data during the upgrade. What exactly is the cause of this? Is it dangerous? Everything seems to work fine, I just want to make sure that something subtle isn't going to cause silent data corruption. TIA. Gordan

1 0

Re: [Maria-developers] [Commits] Rev 3482: post review changes - interface (part 3)
by Sergei Golubchik 21 Dec '12

21 Dec '12

Hi, Sanja! On Dec 21, sanja(a)askmonty.org wrote: > At file:///home/bell/maria/bzr/work-maria-10.0-cassandra/ > > ------------------------------------------------------------ > revno: 3482 > revision-id: sanja(a)askmonty.org-20121220233538-b171yhb9isogvd5h > parent: sanja(a)askmonty.org-20121220204008-fox73otsp1lviqda > committer: sanja(a)askmonty.org > branch nick: work-maria-10.0-cassandra > timestamp: Fri 2012-12-21 01:35:38 +0200 > message: > post review changes - interface (part 3) > === modified file 'include/ma_dyncol.h' > --- a/include/ma_dyncol.h 2012-12-20 20:40:08 +0000 > +++ b/include/ma_dyncol.h 2012-12-20 23:35:38 +0000 > @@ -100,6 +100,8 @@ struct st_dynamic_column_value > > typedef struct st_dynamic_column_value DYNAMIC_COLUMN_VALUE; > > +/* old functions (depricated) */ > +#ifdef MADYNCOL_DEPRICATED "deprEcated" > enum enum_dyncol_func_result > dynamic_column_create(DYNAMIC_COLUMN *str, > uint column_nr, DYNAMIC_COLUMN_VALUE *value); > @@ -133,75 +120,105 @@ dynamic_column_update_many(DYNAMIC_COLUM > uint *column_numbers, > DYNAMIC_COLUMN_VALUE *values); > enum enum_dyncol_func_result > -dynamic_column_update_many_fmt(DYNAMIC_COLUMN *str, > - uint add_column_count, > - void *column_keys, > - DYNAMIC_COLUMN_VALUE *values, > - my_bool string_keys); > - > -enum enum_dyncol_func_result > dynamic_column_delete(DYNAMIC_COLUMN *org, uint column_nr); > > enum enum_dyncol_func_result > dynamic_column_exists(DYNAMIC_COLUMN *org, uint column_nr); > + > enum enum_dyncol_func_result > -dynamic_column_exists_named(DYNAMIC_COLUMN *str, LEX_STRING *name); > +dynamic_column_list(DYNAMIC_COLUMN *org, DYNAMIC_ARRAY *array_of_uint); > + > enum enum_dyncol_func_result > -dynamic_column_exists_fmt(DYNAMIC_COLUMN *str, void *key, my_bool string_keys); > +dynamic_column_get(DYNAMIC_COLUMN *org, uint column_nr, > + DYNAMIC_COLUMN_VALUE *store_it_here); > +#endif > > -/* List of not NULL columns */ > +/* new functions */ > enum enum_dyncol_func_result > -dynamic_column_list(DYNAMIC_COLUMN *org, DYNAMIC_ARRAY *array_of_uint); > +madyncol_create_many(DYNAMIC_COLUMN *str, "madyncol" looks silly :) let's agree on one namespace prefix and use it everywhere. either it's ma_ like in ma_open ma_dyncol_create_many ma_alloc or it's maria_ like in maria_open maria_dyncol_create_many maria_alloc > + uint column_count, > + uint *column_numbers, > + DYNAMIC_COLUMN_VALUE *values, > + my_bool new_string); Regards, Sergei

2 1

Re: [Maria-developers] review of the MDEV-377 Name support for dynamic columns
by Sergei Golubchik 21 Dec '12

21 Dec '12

Hi, Oleksandr! On Dec 21, Oleksandr Byelkin wrote: > It looks like you reviewed some old sources in 10.0-cassandra this is > fixed as we discussed it. I've reviewed the 5.5-cassandra tree. And yes, I thought that was already fixed, so I was a bit susprised to find it unfixed :) Ok, wrong tree, then. Thanks! > [skip] > > > > - > > -int dynamic_column_update(DYNAMIC_COLUMN *str, uint column_nr, > > - DYNAMIC_COLUMN_VALUE *value) > > +enum enum_dyncol_func_result > > +dynamic_column_vals(DYNAMIC_COLUMN *str, > > + DYNAMIC_ARRAY *names, DYNAMIC_ARRAY *vals, > > + char **free_names) > > { > > - return dynamic_column_update_many(str, 1, &column_nr, value); > > + DYN_HEADER header; > > + char *nm; > > + uint i; > > + enum enum_dyncol_func_result rc; > > + > > + *free_names= 0; > > + bzero(names, sizeof(DYNAMIC_ARRAY)); /* In case of errors */ > > + bzero(vals, sizeof(DYNAMIC_ARRAY)); /* In case of errors */ > > + if (str->length == 0) > > + return ER_DYNCOL_OK; /* no columns */ > > + > > + if ((rc= init_read_hdr(&header, str)) < 0) > > + return rc; > > + > > + if (header.entry_size * header.column_count + FIXED_HEADER_SIZE > > > + str->length) > > + return ER_DYNCOL_FORMAT; > > + > > + if (init_dynamic_array(names, sizeof(LEX_STRING), > > + header.column_count, 0) || > > + init_dynamic_array(vals, sizeof(DYNAMIC_COLUMN_VALUE), > > + header.column_count, 0) || > > + (header.format == DYNCOL_FMT_NUM && > > + !(*free_names= (char *)malloc(DYNCOL_NUM_CHAR * header.column_count)))) > > why do you need a special malloc() for names, you can put them in > > the buffer of 'names' dynarray. it'll be a lot more convenient for > > the caller (no free_names to care about). > dynamic array also should be freed, so I do not see difference, I can > rename it to memory_to_free or memory_should_be_freed. The difference is - you pass a separate pointer to the caller, that a caller should take care of. If you put this memory in dynarray, than the caller doesn't have to bother. He eventually delete_dynamic() and that does all the necessary cleanup. It's easier to use for the caller - and for API it's extremely important. More important than simple implementation. Regards, Sergei

1 0