maria-developers team mailing list archive

[Oleksandr Byelkin <sanja@xxxxxxxxxxxx>]

Hi!

22 янв. 2010, в 15:31, Michael Widenius написал(а):

> Hi!
>
> > > > > > "sanja" == sanja  <sanja@xxxxxxxxxxxx> writes:
>
> sanja> At file:///Users/bell/maria/bzr/work-maria-5.2-engine/
> sanja> ------------------------------------------------------------
> sanja> revno: 2734
> sanja> revision-id: sanja@xxxxxxxxxxxx-20091204114937-cfkvax0g36d3nq3j
> sanja> parent: psergey@xxxxxxxxxxxx-20091202142609-18bp41q8mejxl47t
> sanja> committer: sanja@xxxxxxxxxxxx
> sanja> branch nick: work-maria-5.2-engine
> sanja> timestamp: Fri 2009-12-04 13:49:37 +0200
> sanja> message:
> sanja>   Maria WL#61
>
> <cut>
>
>
> > +++ b/CMakeLists.txt	2009-12-04 11:49:37 +0000
> > @@ -251,6 +251,7 @@
> >
> >         IF (ENGINE_BUILD_TYPE STREQUAL "STATIC")
> >           SET (mysql_plugin_defs  "${mysql_plugin_defs},builtin_$ 
> > {PLUGIN_NAME}_plugin")
> > +          SET (mariaext_plugin_defs  "$ 
> > {mariaext_plugin_defs},bltnmext_${PLUGIN_NAME}_plugin")
> >           SET (MYSQLD_STATIC_ENGINE_LIBS $ 
> > {MYSQLD_STATIC_ENGINE_LIBS} ${PLUGIN_NAME})
> >           SET (STORAGE_ENGINE_DEFS "${STORAGE_ENGINE_DEFS} -DWITH_$ 
> > {ENGINE}_STORAGE_ENGINE")
> >           SET (WITH_${ENGINE}_STORAGE_ENGINE TRUE)
> > @@ -269,6 +270,7 @@
> > IF(NOT WITHOUT_PARTITION_STORAGE_ENGINE)
> >   SET (STORAGE_ENGINE_DEFS "${STORAGE_ENGINE_DEFS} - 
> > DWITH_PARTITION_STORAGE_ENGINE")
> >   SET (mysql_plugin_defs  "$ 
> > {mysql_plugin_defs},builtin_partition_plugin")
> > +  SET (mariaext_plugin_defs  "$ 
> > {mariaext_plugin_defs},bltnmext_partition_plugin")
> > ENDIF(NOT WITHOUT_PARTITION_STORAGE_ENGINE)
>
> Probably stupid questions, but why mariaext instead of just mariadb?

MARIAdbEXTension

> What does 'bltnmext_partition_plugin' stand for ?
> Can we use a more understandable name, like  
> builtin_mariadb_partition_plugin ?

yes

> <cut>
>
> > --- a/include/mysql/plugin.h	2009-09-07 20:50:10 +0000
>
> <cut>
>
> > +/*
> > +  MariaDB extension for plugins declaration structure.
> > +*/
> > +
> > +struct st_mariaext_plugin
> > +{
> > +  const char *sversion;  /* plugin version string */
>
> sversion -> version ?

actually it was string version (we have already numeric one).

> > +  int maturity;          /* HA_PLUGIN_MATURITY_XXX */
> > +};
> > +
> > / 
> > *************************************************************************
> >   API for Full-text parser plugin. (MYSQL_FTPARSER_PLUGIN)
> > */
> >
> > === modified file 'sql/sql_plugin.cc'
> > --- a/sql/sql_plugin.cc	2009-11-12 04:31:28 +0000
> > +++ b/sql/sql_plugin.cc	2009-12-04 11:49:37 +0000
[skip]
> > +    /* extensions should be the same numbers as static plugins) */
> > +    DBUG_ASSERT(*bltnmexts);
> > +    for (plugin= *builtins, ext= *bltnmexts;
> > +         plugin->info;
> > +         plugin++, ext++)
> >     {
> > +
> > +      /* in case if plugin describe less extensions then plugins */
> > +      mariaext= ext;
> > +      if (!ext->sversion)
> > +      {
> > +         mariaext= empty_mariaext;
> > +         ext--;
> > +      }
> > +
>
> I am a bit unsure about the above code;  If we allow one to describe
> less extensions, then it's not sure that when one adds a new plugin
> with extension information, one remembers to also add all missing
> extensions.
>
> I think that for staticly compiled plugins we should require that
> they have also extension information. What do you think about this?

It is for dynamically linked plugins. Of course statically linked will  
have everything but they will be included in built in plugins array  
(it is other code).

[skip]