Hi,

Henri Yandell wrote:
On Tue, 17 Dec 2002, Sean Schofield wrote:


I took a quick look at Joda's date parser and saw it has its own
engine.

I believe the commons one should lean onto the syntax of the java's,
such that should allow it as a bug free plug-in replacement of Sun's.

What do others think?
I agree.  I don't think we want to write a  new DateFormat from scratch.
Perhaps we can write our own class with synchronized methods that wraps
the existing SimpleDateFormat (just a thought.)  Or maybe we can rip the
source from the JDK version and add the fixes there.

Just to get it in quick. -1 to ripping source from the JDK. Licenses bite
our arses too quickly when you get near here so it's not something ASF can
do.
My point was that JDK DateFormat code is *so* broken we do not want to rip
it, nor place a synchronized warpper around it. License is another issue...

I need to take a look at Ant's code. Maybe
I'll take the leap and start a working bug free plug-in replacement
of Sun's SimpleDateFormat if no one has taken the lead...
We have already moved the ant code int the lang module of CVS (moved it
yesterday).  Right now its identical to the Ant code.  We're going to go
from there.  A thread-safe DateFormat would probably be a good next
step.  We're also trying to figure out what to do with CalendarUtils
which came from the sandbox.  (See the recent email thread concerning
DateUtils for yesterday's discussion on this.)

I'm starting to want a time sub-package as this gets larger :) Once it
hits 3 classes or so.
Either make the DateFormat an inner class to the ant DateUtil or maybe
if the date stuff grows to more than 3 classes, create a subpackage.
I don't have any string feelings on this issue of where (you all in
commons have done a good job in packaging until now!).

Hen

--
:) Christoph Reck


--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to