maria-developers team mailing list archive

[Aleksey Midenkov <midenok@xxxxxxxxx>]

Hi Sergei, Monty!

On Fri, Sep 10, 2021 at 10:02 PM Sergei Golubchik <serg@xxxxxxxxxxx> wrote:
>
> Hi, Aleksey!
>
> I presume that changes to the logging code is what Monty asked.

Not everything. I've done what I could in the current timeframe, the
less important things were postponed. Didn't get his approval for that
though, so I'm asking now whether the current state of patch is OK to
push from his POV?

> So I'll only comment on a couple of minor issues:
>
> On Sep 10, Aleksey Midenkov wrote:
> > revision-id: 818407474d6 (mariadb-10.6.1-80-g818407474d6)
> > parent(s): bb4b95975f7
> > author: Aleksey Midenkov
> > committer: Aleksey Midenkov
> > timestamp: 2021-09-10 18:47:31 +0300
> > message:
> >
> > MDEV-22166 CONVERT PARTITION: move out partition into a table
> >
> > diff --git a/sql/sql_partition.cc b/sql/sql_partition.cc
> > index db371d23ba7..cbdb40472d8 100644
> > --- a/sql/sql_partition.cc
> > +++ b/sql/sql_partition.cc
> > @@ -5371,8 +5367,12 @@ that are reorganised.
> >          tab_part_info->is_auto_partitioned= FALSE;
> >        }
> >      }
> > -    else if (alter_info->partition_flags & ALTER_PARTITION_DROP)
> > +    else if ((alter_info->partition_flags & ALTER_PARTITION_DROP) |
> > +             (alter_info->partition_flags & ALTER_PARTITION_CONVERT_OUT))
> >      {
> > +      const char * const cmd=
> > +        (alter_info->partition_flags & ALTER_PARTITION_CONVERT_OUT) ?
> > +          "EXTRACT" : "DROP";
>
> "CONVERT"

Fixed.

>
> >        /*
> >          Drop a partition from a range partition and list partitioning is
> >          always safe and can be made more or less immediate. It is necessary
> > diff --git a/sql/sql_table.h b/sql/sql_table.h
> > index aacb6c99f15..a580c7d348d 100644
> > --- a/sql/sql_table.h
> > +++ b/sql/sql_table.h
> > @@ -20,6 +20,17 @@
> >  #include <my_sys.h>                             // pthread_mutex_t
> >  #include "m_string.h"                           // LEX_CUSTRING
> >
> > +inline bool suicide()
> > +{
> > +  DBUG_SUICIDE();
> > +  return false;
> > +}
>
> as I wrote earlier, this is wrong. DBUG_SUICIDE() must always be an
> expression.

Fixed.

>
> > +
> > +#define ERROR_INJECT_CRASH(code) \
> > +  (DBUG_IF(code) && suicide())
> > +#define ERROR_INJECT_ERROR(code) \
> > +  (DBUG_IF(code) && (my_error(ER_UNKNOWN_ERROR, MYF(0)), 1))
> > +
> >  class Alter_info;
> >  class Alter_table_ctx;
> >  class Column_definition;
>
> Regards,
> Sergei
> VP of MariaDB Server Engineering
> and security@xxxxxxxxxxx



-- 
@midenok