Ilja Booij wrote:
Aaron Stone wrote:
Off the top of my head, there's a two step process here. First is that the top
level, messgae-oriented functions will need an extra flag. Second is that the
inner physmessage functions will need that flag, too.
db_insert_message_block() and db_insert_message_block_physmessage().
I don't think we should wait on 2.0 for this. We can tweak performance in the
point releases. We should add whatever columns we'll need to support the
functionality, though.
I'm going to release a new RC today. I'd like to add the new column
(is_header) to the messageblks table, just to be prepared for it. We can
then add the code to use it later, without having to change the database
structure.
I might make sense to also add an index right away, because we're also
going to need that later on.
INDEX messageblk_physmsg_isheader_idx(physmessage_id, isheader)
Ilja
