Details
-
Technical task
-
Status: In Review (View Workflow)
-
Major
-
Resolution: Unresolved
-
10.5
-
None
Description
MDEV-21978 found many erroneous my_snprintf formattings when it reenabled GCC printf checks.
Patches to these errors should apply to the oldest applicable version in maintenance and up.
In MariaDB 10.5 and up, the following strings may display their fields corrupted depending on the compiler and platform:
- …,<server host>,<username>,<host>,<connection ID>,<query ID>,<operation>
- …,<database>,,<event status>
- …,<database>,<table>,
- …,<database>,<table>|<new database>.<new table>,
- …at offset <RegExp error offset>
- The slave is applying a ROW event on behalf of a<event> statement on table <table> and is currently taking a considerable amount of time (<difference> seconds). This is due to the fact that it is scanning <something> while looking up records to be processed.
- The current group of events starts with a non-GTID <type> event; the last seen GTID is <GTID>
- <±><seconds>.<microseconds>
- table->s->reclength <length> table->s->fields <count>
- table->s->key_info[<index>] flags & HA_NOSAME = <integer>
- table->field[<index>]->offset = %<difference> and the chunk that follows
- repl.max_ws_size=<size>
InnoDB: invalid PAGE_COMPRESSION_LEVEL = <level>. Valid values are [1, 2, 3, 4, 5, 6, 7, 8, 9](big-endian CPUs only)- Got error from move_object(<from> -> <to>): <error ID> % (% should be an error message)
- normalized data truncated for multiple column index: … column-name: <<name>> data: <<data>> (known: Visual Studio)
- Found block with wrong recordlength: <length> at <position>
- Deleted block with impossible length <length> at <position>
- Found block with wrong recordlength: <length> at <position>
- Can't read indexpage from filepos: <integer>