MySQL Bugs: #40358: HA_KEY_HAS_PART_KEY_SEG does not properly get passed to storage engine

Bug #40358	HA_KEY_HAS_PART_KEY_SEG does not properly get passed to storage engine
Submitted:	27 Oct 2008 19:34	Modified:	9 Jul 2009 15:06
Reporter:	Zardosht Kasheff (OCA)	Email Updates:
Status:	Not a Bug	Impact on me:	None
Category:	MySQL Server	Severity:	S3 (Non-critical)
Version:	5.1.28	OS:	Any
Assigned to:	Alexander Nozdrin	CPU Architecture:	Any
Tags:	HA_KEY_HAS_PART_KEY_SEG

Description:
The flag HA_KEY_HAS_PART_KEY_SEG does not get passed to storage engines in the key->flags.

The reason is the following. Although properly set in mysql_prepare_create_table, in the function pack_keys in unireg.cc, only 2 bytes are stored for a key's flags. This causes HA_KEY_HAS_PART_KEY_SEG to be truncated, because it maps to 0x10000.

How to repeat:
Do the following:
create table foo (a varchar(10), b int, key (a(3)));

Break in mysqld!pack_keys

Notice how keyinfo[0].flags = 0x10040

Break in mysqld!ha_myisam::create

Notice how table_arg->key_info[0].flags = 0x40

Suggested fix:
4 bytes should be stored for a key's flags. KEY->flags is a ulong, so it makes sense to store more than 2 bytes.

Thank you for the bug report. Verified as described:

-	keyinfo	0x01f4efc0 {key_length=0x00000009 flags=0x00010068 key_parts=0x00000001 ...}	st_key *
	key_length	0x00000009	unsigned int
	flags	0x00010068	unsigned long
	key_parts	0x00000001	unsigned int
	extra_length	0x00000000	unsigned int
	usable_key_parts	0x00000000	unsigned int
	block_size	0x00000000	unsigned int
	algorithm	HA_KEY_ALG_UNDEF	ha_key_alg
+	parser	0x00000000	st_plugin_int * *
<cut>

-	table_arg	0x0392d8ac {s=0x0392d458 {table_category=TABLE_CATEGORY_USER name_hash={key_offset=0x00000000 key_length=0x00000000 blength=0x00000000 ...} mem_root={free=0x01f2dc50 {next=0x00000000 left=0x000000f8 size=0x000003c8 } used=0x00000000 {next=??? left=??? size=??? } pre_alloc=0x00000000 {next=??? left=??? size=??? } ...} ...} file=0x01f54a10 {file=0x00000000 {s=??? state=??? save_state={records=??? del=??? empty=??? ...} ...} int_table_flags=0x0000000d14626d91 data_file_name=0xa5a5a5a5 <Bad Ptr> ...} next=0x00000000 {s=??? file=??? next=??? ...} ...}	st_table *
<cut>
-	key_info	0x01f54c40 {key_length=0x0000000c flags=0x00000068 key_parts=0x00000001 ...}	st_key *
	key_length	0x0000000c	unsigned int
	flags	0x00000068	unsigned long
	key_parts	0x00000001	unsigned int
	extra_length	0x00000003	unsigned int
	usable_key_parts	0x00000001	unsigned int
	block_size	0x00000000	unsigned int
	algorithm	HA_KEY_ALG_UNDEF	ha_key_alg
+	parser	0x00000000	st_plugin_int * *
<cut>

I discussed this issue with Sergei Golubchik.

It is not a bug, because HA_KEY_HAS_PART_KEY_SEG flag is
internal to the MySQL server by design. It is not supposed
neither to be saved in FRM-files, nor to be passed to storage
engines (it's intended to pass information into internal
static sort_keys(KEY *, KEY *) function)

This flag can actually be calculated -- it's based on key lengths
comparison. So, a storage engine can determine it by itself.

The main motivation for filing this bug is that the fact that the .frm file only uses 2 bytes to store key flags can be a problem. The internal data structure uses 4 bytes. So, the key flags cannot be extended. I will file a separate bug on this.

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/78300

2826 Alexander Nozdrin	2009-07-09
      Add a comment in the source (Bug#40358).

Pushed into 5.4.4-alpha (revid:alik@sun.com-20090723102221-ps4uaphwbxzj8p0q) (version source revid:alik@sun.com-20090709163043-2gf8cgh583hjcbr4) (merge vers: 5.4.4-alpha) (pib:11)

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/87159

2931 Alexander Nozdrin	2009-10-16
      Backporting revision from 6.0:
      revno: 2617.56.25
      committer: Alexander Nozdrin <alik@sun.com>
      branch nick: azalea-bugfixing
      timestamp: Thu 2009-07-09 20:30:43 +0400
      message:
        Add a comment in the source (Bug#40358).

Pushed into 6.0.14-alpha (revid:alik@sun.com-20091102151658-j9o4wgro47m5v84d) (version source revid:alik@sun.com-20091023064702-2f8jdmny61bdl94u) (merge vers: 6.0.14-alpha) (pib:13)

Pushed into 5.5.0-beta (revid:alik@sun.com-20091110093229-0bh5hix780cyeicl) (version source revid:mikael@mysql.com-20091103113702-p61dlwc6ml6fxg18) (merge vers: 5.5.0-beta) (pib:13)