database/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2025-10-27 05:56:07 +03:00
Code Activity

branches/innodb+: Merge revisions 6949:6970 from branches/zip,

Browse Source 
skipping the revisions that do not make sense in branches/innodb+:
  ------------------------------------------------------------------------
  r6957 | vdimov | 2010-04-06 22:05:54 +0300 (Tue, 06 Apr 2010) | 2 lines
  Changed paths:
     M /branches/zip/trx/trx0trx.c

  branches/zip: Whitespace fixup
  ------------------------------------------------------------------------
  r6962 | mmakela | 2010-04-07 15:03:54 +0300 (Wed, 07 Apr 2010) | 2 lines
  Changed paths:
     M /branches/zip/handler/ha_innodb.cc

  branches/zip: innobase_init(): Correct the error message
  about wrong innodb_change_buffering value.  Reported by Ranger.
  ------------------------------------------------------------------------
  r6967 | vdimov | 2010-04-07 22:40:38 +0300 (Wed, 07 Apr 2010) | 2 lines
  Changed paths:
     M /branches/zip/include/univ.i

  branches/zip: Whitespace fixup in univ.i
  ------------------------------------------------------------------------
This commit is contained in:
mmakela
2010-04-22 09:48:57 +00:00
parent 18a68a24b9
commit eda8d678a5
3 changed files with 3 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
handler/ha_innodb.cc
View File

@@ -2281,7 +2281,7 @@ mem_free_and_error:
}
sql_print_error("InnoDB: invalid value "
"innodb_file_format_check=%s",
"innodb_change_buffering=%s",
innobase_change_buffering);
goto mem_free_and_error;
}

2
include/univ.i
View File

@@ -115,7 +115,7 @@ if we are compiling on Windows. */
/* Include <sys/stat.h> to get S_I... macros defined for os0file.c */
# include <sys/stat.h>
# if !defined(__NETWARE__) && !defined(__WIN__)
# if !defined(__NETWARE__) && !defined(__WIN__)
# include <sys/mman.h> /* mmap() for os0proc.c */
# endif

2
trx/trx0trx.c
View File

@@ -847,7 +847,7 @@ trx_commit_off_kernel(
recovery i.e.: back ground rollback thread is still active
then there is a chance that the rollback thread may see
this trx as COMMITTED_IN_MEMORY and goes adhead to clean it
up calling trx_cleanup_at_db_startup(). This can happen
up calling trx_cleanup_at_db_startup(). This can happen
in the case we are committing a trx here that is left in
PREPARED state during the crash. Note that commit of the
rollback of a PREPARED trx happens in the recovery thread