Prestons example suggests, that in that particular case it is possible
to be compliant with 3.1 ("optional second parameter"), while
upholding upwards compatibility to 1.0.

Or, in other words: We can have both. (In that particular case.)

In less obvious cases, one might introduce a configuration parameter
like "xquery_spec_level" with a default value of "1.0".

In general, I wouldn't want to block further development in any
direction. Although I agree with Till, that reaching older targets
should have priority over new targets. Otherwise, we'd basically play
hare and hedgehog with a moving target. Or, in other words: We'd be
the hare that cannot catch the hedgehog.

Jochen


On Tue, Feb 24, 2015 at 8:43 AM, Till Westmann <[email protected]> wrote:
> Hi,
>
> right now I’d prefer to focus on getting further along on the path to XQuery 
> 1.0 compliance.
> There’s still quite some work between us and XQuery 1.0. And XQuery 3.0 and 
> 3.1 contain a significant number of new features (both in the data model and 
> in the language).
> I think that it’d be better to strive for 1.0 compliance with a few 
> 3.0/3.1-based extensions (the group-by cause comes to mind) than to aim at 
> 3.1 immediately.
>
> My 2c,
> Till
>
>> On Feb 23, 2015, at 4:37 PM, Eldon Carman <[email protected]> wrote:
>>
>> Hi all,
>>
>> What do you think about targeting for XQuery 3.1 specification instead of
>> 1.0?
>>
>> Recently, I was debugging the round function for XQuery decimal values. As
>> a result I found that round function in 3.1 allows an optional second
>> parameter to specify precision. While many functions remain the same, a few
>> have new parameters.
>>
>> The XQuery Test Suite for 3.* specifications has updated the previous tests
>> and add new ones to cover the new features. The new test suite has ~26,000
>> tests. I create an issue to support the new XQTS 3.* catalog [1]. While
>> supporting the new testing suite is step one, I was curious to see what
>> everyone though about supporting XQuery 3.1 vs XQuery 1.0. What do you see
>> as the pros and cons?
>>
>> Thanks
>> Preston
>>
>> [1] https://issues.apache.org/jira/browse/VXQUERY-137
>



-- 
Any world that can produce the Taj Mahal, William Shakespeare,
and Stripe toothpaste can't be all bad. (C.R. MacNamara, One Two Three)

Reply via email to