Rob Atkinson ha scritto:
> Justin's solution seems the most elegant - i.e. this is simply an 
> encoding issue at both the source and XML output end.

I agree it's elegant not too sure I'd call it a pure xml encoding issue 
thought. People are going to use those Calendar objects
to build applications (since that's what we're going to have inside
the Features). Using the java.util.Date hierarchy has the
advantage of being familiar to most java programmers (more familiar
than Calendar, that is).

I also have the feeling that doing resultSet.getObject(xxx) on a
time column will return a java.sql.Time, if it does, then the original
timezone information is lost anyways on databases.
This hibernate article seems to suggest so: 
http://www.hibernate.org/100.html
That is, the impression is that you can do things so that you read
date/time/timestamp in whatever timezone you want, but you can't
get to know what the db timezone is. I'll investigate further during
the weekend.

Even if this cuts off databases, keeping around the timezone information
can be worthy in the case of cascaded WFS, where we get to parse the
XML and thus retain the original timezone info if using Calendar objects.

> For myself, I would suggest we live with the ugly limitations of date in 
> 2.4,  and would push for a proper solution on trunk rather than a lot of 
> effort on 2.4 which will just delay and derail the real progress required. 
> 
> i.e. perhaps Andreas solution #3 will work for 2.4 - its a system wide 
> config, for expert users, with a proper solution in the pipeline.

I believe it is. Yet, it seems I failed to convey the difference between
a short term solution that can be done now, and a long term solution
that I fear won't be applied at all unless there is specific funding
for it. There is no sponsoring behind this topic, it's just something
I felt was important because it's one of the basics we don't get right.

Then again, the general consensus seems to be to go for the clean way.
I hope we'll have the resources (and will) to address it in the not too 
distant future, in other words, that it does not become another GeoAPI 
filter thing (everybody agrees we should clean up the code base from
the old filters, yet nobody is doing anything about it because it's too
much work).

> I think we would be kidding ourselves that a lot of work on a specific 
> case is not going to result in making the next step less stable.

Not sure I understand? :)
Cheers
Andrea

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Geotools-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/geotools-devel

Reply via email to