transactions - end of transaction actions

[John Arbash Meinel]

Aaron Bentley wrote:
> Robert Collins wrote:
> | accordance with the above ordering rule]. One decision we need to make
> | is whether writes can be globbed together - i.e. if I ask for something
> | to be written that affects files X, then Y, then X again, can the system
> | write both setsof data to X, and then the one to Y? I think that this
> | violates the domain ordering constraints in the background section
> | above, and we should say that the transaction must not allow that to
> | happen.
> 
> I would add, "absent specific grouping instructions."  For example, if
> you delayed writes to inventory storage until all text storage updates
> were complete, you'd be able to do the text storage updates in any order
> you liked.
> 
> Not that we should necessarily do this, but it does seem possible.
> 
> Aaron

I agree it is possible. I'm curious what it would do for performance. It
also would be interesting if you could have 'superceeders'. For example,
if you update the inventory.weave more than one time, probably only the
last one needs to be written out.

But I think the way Robert described it is nice and straightforward. It
is just a FIFO queue.

John
=:->
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 256 bytes
Desc: OpenPGP digital signature
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20060124/d32a0482/attachment.pgp