On Thu, Jun 20, 2013 at 11:50 AM, Ivan Zhakov <ivan_at_visualsvn.com> wrote:
> On Thu, Jun 20, 2013 at 5:41 AM, Johan Corveleyn <jcorvel_at_gmail.com> wrote:
>> On Thu, Jun 20, 2013 at 3:27 AM, Johan Corveleyn <jcorvel_at_gmail.com> wrote:
>>> On Thu, Jun 20, 2013 at 3:23 AM, Mark Phippard <markphip_at_gmail.com> wrote:
>>>> On Jun 19, 2013, at 9:09 PM, jcorvel_at_apache.org wrote:
>>>>
>>>>> Author: jcorvel
>>>>> Date: Thu Jun 20 01:09:20 2013
>>>>> New Revision: 1494829
>>>>>
>>>>> URL: http://svn.apache.org/r1494829
>>>>> Log:
>>>>> Add information about serf skelta mode and the configuration knobs involved
>>>>> to the 1.8 release notes.
>>>>>
>>>>> Patch by: lgo
>>>>> (tweaked by me)
>>>>>
>>>>> Modified:
>>>>> subversion/site/publish/docs/release-notes/1.8.html
>>>>>
>>>>>
>>>>>
>>>>> +<div class="notice">
>>>>> +<p>For 1.8 servers, set <strong>SVNAllowBulkUpdates</strong> to
>>>>> +<strong>Prefer</strong> if your server configuration does not allow an
>>>>> +intermediate caching server - on your side or on the client's side - to
>>>>> +successfully cache server responses.</p>
>>>>> +</div>
>>>>>
>>>>
>>>> I do not understand why we would make this recommendation. It seems overly strong.
>>>>
>>>> This whole section paints Serf in a really negative way as it makes it sound like skelta mode has little value.
>>>>
>>>
>>> Agreed, let's yank that notice. As I said in my other reply (which I
>>> just sent before reading yours :-), this ignores the client-side
>>> optimization of not requesting pristines that are already at the
>>> client side.
>>
>> Okay, I removed it.
>>
>> As I said, I'll leave it to someone else to advertise the
>> pristine-fetching benefit.
>>
> I don't like such big last minute changes to release notes and will be
> glad if all change will be reverted until we find best wording for
> this details.
Fine for me. I actually don't care about this personally (at our
workplace we're still lightyears away from upgrading, both client and
server side). I just found it to be an important omission from our
release notes (that actually should have been included way before
release), and thought the information good enough, and at least better
than nothing. But if people object, no problem.
Feel free to revert (and drive the effort to find the best wording).
BTW: apparently the pristine-fetching optimization is actually present
in the release notes, in the Enhancements and Bugfixes section:
http://subversion.apache.org/docs/release-notes/1.8.html#pristines
(though that section doesn't say it's dependent on serf's
skelta-mode). But I think we should advertise this a bit in the "serf"
section as an important benefit (perhaps referencing the other
section).
--
Johan
Received on 2013-06-20 12:56:45 CEST