Re: xAPdb ready for testing....

[mark_harrison_uk2]

Thanks to the people who've replied off-list volunteering to test.

I thought that it would be worth documenting the tables used... all
of which are assumed to be in a database called xap.

TABLE message_header
header_id type int(11), not null, default 0
date_entered type datetime, not null, default 0000-00-00 00:00
xap_version type int(11), not null, default 0
hop_count type int(10)unsigned, not null, default 0
unique_identifier type varchar(8), not null
class type varchar (255), not null
source type varchar (255), not null
target type varchar (255)

TABLE message_block
header_id type int(10)unsigned, not null, default 0
block_id type int(10)unsigned, not null, default 0
block_name type varchar(255), not null

TABLE block_contents
header_id type int(10)unsigned, not null, default 0
block_id type int(10)unsigned, not null, default 0
content_id type int(10)unsigned, not null, default 0
keyname type varchar(255), not null
value type varchar(255), not null
is_hex type char(1)

The logic for having IDs, which are automatically generated by the
logger program, and rise monotonically, is to enable the EXACT re-
creation of the xAP message stream, even in areas where multiple
message blocks have identical names within the same message, and
multiple name-value pairs have identical names within the same
message block.

As you might imagine:

- header_id forms an index on message_header
- header_id, block_id forms an index on message_header
- header_id, block_id, content_id forms an index on message_header

The semantics of the fields should be fairly obvious :-)

Regards,

Mark