transport lock methods

[John A Meinel]

Robert Collins wrote:
> On Fri, 2006-02-17 at 10:39 -0600, John A Meinel wrote:
> ..
>> I agree that we don't really want to make 0.8 a watershed. But how would
>> you recommend doing it properly. If you are planning to stay cooperative
>> with bzr-0.7, then we need to leave locks the way they are.
> 
> We can move them around in the code base though.
> 
> For instance we can fix a bug in 0.7 by:
> os locking where we can
> locking branch-lock.write-lock *as well* irrespective of transport as
> long as its writable.
> 
> Rob
> 

I suppose that is good enough. You still have a double lock, and what do
you do if one exists but the other doesn't, etc. I suppose we just need
to always unlock in order, and unlock in the opposite order. Though we
still have race conditions against clients or transports that don't
handle both locks.

John
=:->


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 249 bytes
Desc: OpenPGP digital signature
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20060217/24b0377e/attachment.pgp