Re: [PATCH] Clarify the server behaviour requirements for file locking

[Greg Hudson <gh...@MIT.EDU>]

On Tue, 2005-01-18 at 09:14, Julian Foad wrote:
> The locking functional specification says that the server has to check the 
> locks as it receives the transaction, before checking them again at commit 
> finalisation time.  But that is really an implementation-specific optimisation. 
>   Do you agree with this clarification and, if so, may I commit it?

An "implementation-specific optimization"?  We aren't the IETF; we only
have one implementation.  And locking-functional-spec.txt is not a
standards document, just some notes we wrote for ourselves.


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org

Re: [PATCH] Clarify the server behaviour requirements for file locking

[Julian Foad <ju...@btopenworld.com>]

Greg Hudson wrote:
> On Tue, 2005-01-18 at 09:14, Julian Foad wrote:
>>The locking functional specification says that the server has to check the 
>>locks as it receives the transaction, before checking them again at commit 
>>finalisation time.  But that is really an implementation-specific optimisation. 
>>  Do you agree with this clarification and, if so, may I commit it?
> 
> An "implementation-specific optimization"?  We aren't the IETF; we only
> have one implementation.  And locking-functional-spec.txt is not a
> standards document, just some notes we wrote for ourselves.

You're right - I worded that rationale a bit too strongly.  I'll re-phrase it:

That is really a performance optimisation, albeit an important one, rather than 
a functional requirement.

- Julian

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org