Kazutoshi Satoda wrote:
>> That's why the client should be fixed first, so that the server can be
>> fixed (in a fully compliant way) when all supported client versions
>> are known to work.
>
> In that way, server can not be fixed before 1.3.0 becomes unsupported.
> Then issue 2151 will be left for, I think, at least 6 months. With my
> proposal, issue 2151 can be fixed in 1.4.0. It also can be fixed in
> 1.3.1 with some luck. I strongly prefer the performance fix comes first.
I didn't propose not to do other changes. I merely was saying that if
making the server fully RFC3253 compliant (on this matter) means that
all relevant clients need to be upgraded first, then this should be done
ASAP
>>> Am I missing something?
>>
>> Yes. If you don't fix the client, you can't make the server fully
>> compliant at a later point of time (right?).
>
> That's right. But issue 2151 is a performance issue, not a compliance
> issue. Making the server fully compliant is no required here. I know
> the fixed server won't be a fully compliant DeltaV server, but that is
> not a regression. Does my proposal prevent the compliance issue?
Not really. But it seems to be a workaround for a problem caused by
non-compliance, thus the *long-term* fix should be to get rid of the
compliance issue, instead of wiring-in a workaround forever, right?
Best regards, Julian
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Feb 20 20:17:57 2006