Re: bk commit into 5.0 tree (gkodinov:1.2521) BUG#29325

[Guilhem Bichot]

Hello Georgi,

On Wed, Jul 11, 2007 at 05:15:45PM +0300, Georgi Kodinov wrote:
> On 11.07.2007, at 16:58, Guilhem Bichot wrote:
> >On Wed, Jul 11, 2007 at 10:49:57AM +0300, kgeorge@stripped wrote:
> >>ChangeSet@stripped, 2007-07-11 10:49:54+03:00, gkodinov@stripped  
> >>+9 -0
> >>  Bug #29325:
> >>  By default MyISAM overwrites .MYD and .MYI files no
> >>  DATA DIRECTORY option is used. This can lead to two tables
> >>  using the same .MYD and .MYI files (that can't be dropped).
> >>
> >>  To prevent CREATE TABLE from overwriting a file a new option
> >>  is introduced : keep_files_on_create
> >>  When this is on the CREATE TABLE throws an error if either
> >>  the .MYD or .MYI exists for a MyISAM table.
> >>  The option is off by default (resulting in compatible behavior).
> >
> >>diff -Nrup a/myisam/mi_create.c b/myisam/mi_create.c
> >>--- a/myisam/mi_create.c	2007-03-13 16:05:10 +02:00
> >>+++ b/myisam/mi_create.c	2007-07-11 10:49:53 +03:00
> >>@@ -586,7 +586,8 @@ int mi_create(const char *name,uint keys
> >> 					     32 : 0));
> >>     linkname_ptr=0;
> >>     /* Replace the current file */
> >>-    create_flag=MY_DELETE_OLD;
> >>+    if (!(flags & HA_CREATE_KEEP_FILES))
> >>+      create_flag=MY_DELETE_OLD;
> >>   }
> >>
> >>   /*
> >>@@ -647,7 +648,8 @@ int mi_create(const char *name,uint keys
> >>       {
> >> 	fn_format(filename,name,"",MI_NAME_DEXT,4);
> >> 	linkname_ptr=0;
> >>-	create_flag=MY_DELETE_OLD;
> >>+        if (!(flags & HA_CREATE_KEEP_FILES))
> >>+          create_flag=MY_DELETE_OLD;
> >>       }
> >>       if ((dfile=
> >> 	   my_create_with_symlink(linkname_ptr, filename, 0, create_mode,
> >
> >ha_myisam::create() and mi_create() are also called in case of
> >TRUNCATE TABLE (destroying old data and index files as it creates new
> >empty ones). I wonder if it could fail now, with the new option on.
> 
> This option (HA_CREATE_KEEP_FILES) is never passed down for TRUNCATE  
> TABLE.

great, thanks for checking :)

> >Could you please verify, with the new option on and a simple table
> >like this:
> >
> >create table t(a int) engine=myisam;
> >insert into t values(1);
> >truncate table t;
> >
> >that TRUNCATE still works, and if it works, add this in a .test file?
> 
> Good idea. The patch is at http://lists.mysql.com/commits/30697. Can  
> you please check it out and let me know if it's OK to push ?

It is ok to push but please consider adding ENGINE=MyISAM to the
"CREATE TABLE t3", as the bug and the bugfix were about MyISAM.