Hi,
* Alexander Nozdrin <alik@stripped> [08/01/28 16:49]:
> ChangeSet@stripped, 2008-01-28 16:20:15+03:00, anozdrin@quad. +14 -0
> Fix for Bug#30217: Views: changes in metadata behaviour
> between 5.0 and 5.1.
> diff -Nrup a/mysql-test/include/ddl_i18n.check_views.inc
> b/mysql-test/include/ddl_i18n.check_views.inc
> --- a/mysql-test/include/ddl_i18n.check_views.inc 2007-06-28 21:34:50 +04:00
> +++ b/mysql-test/include/ddl_i18n.check_views.inc 2008-01-28 16:20:13 +03:00
> @@ -9,6 +9,10 @@ SHOW CREATE VIEW v1|
>
> SHOW CREATE VIEW v2|
>
> +--echo
> +
> +SHOW CREATE VIEW v3|
> +
Please add a test case that shows the output of SHOW CREATE
VIEW/SELECT * FROM INFORMATION_SCHEMA after the view was restored
from backup.
> void Item_string::print(String *str)
> {
> - str->append('_');
> - str->append(collation.collation->csname);
> + THD *thd= current_thd;
> + bool dump_utf8_query= thd->get_dump_utf8_query();
Please do not use a global variable in THD. Either add a new
virtual method (::print_utf8) or a parameter to the existing
hierarchy of print() methods.
> + init(); // TODO: why call it here, why not in init_for_queries()?
> +
If you would like to push this or other not directly related
comments (below), let's discuss them and add in a separate patch.
> +/**
> + This operation just contains the code common for THD-instance
> + initialization for a new connection and for COM_CHANGE_USER
> + implementation.
> */
> +/**
> + Initialize THD instance for connection processing.
>
> -/*
> - Init THD for query processing.
> - This has to be called once before we call mysql_parse.
> - See also comments in sql_class.h.
> + Initialize memory roots necessary for query processing and (!)
> + pre-allocate memory for it. We can't do that in THD constructor because
> + there are use cases (acl_init, delayed inserts, watcher threads, killing
> + mysqld) where it's vital to not allocate excessive and not used memory.
> +
> + TODO: is it really the case? It seems, THD::init_for_queries() is always
> + called for THD instance. It seems, the real reason is that THD instance
> + can be cached, i.e. one THD instance can serve several connections. So,
> + this operation pre-allocates memory for the new connection. Similarly,
> + cleanup() frees pre-allocated memory for the connection.
> +
> + This operation must be be called once before query parsing (i.e. before
> + calling mysql_parse()).
> +
> + @note This operation does not return error. If preallocation fails, it
> + will be noticed at the first call to alloc_root().
> */
> +/**
> + Change user context (implement COM_CHANGE_USER).
>
> - SYNOPSIS
> - change_user()
> -
> - IMPLEMENTATION
> - Reset all resources that are connection specific
> + Basically change_user() resets all resources that are connection
> + specific.
> */
>
> +/**
> + Cleanup THD instance after connection.
> +
> + Reverse operation for THD::init_for_queries(). THD::cleanup() frees all
> + the connection-specific memory so that THD instance can be placed back
> + into the cache and then used for handling another connection.
> +*/
>
> +/**
> + Cleanup THD instance after query execution.
> + This function is used to reset thread data to its default state. It is
> + called once after processing one atomic query.
> + Unfortunately there is no distinguish operation to be called before
> + processing atomic query -- this logic is distributed in the code. For
> + example, one can consider sql_parse.cc (dispatch_command(), alloc_root(),
> + mysql_parse() and so on).
> +
> + @note This function is not suitable for setting thread data to some
> + non-default values, as there is only one replication thread, so different
> + master threads may overwrite data of each other on slave.
> */
--
-- Konstantin Osipov Software Developer, Moscow, Russia
-- MySQL AB, www.mysql.com The best DATABASE COMPANY in the GALAXY
