Dave Sowerby wrote:
Hi Simon,
With regards to the 1.2.1 release you are correct that we have a
patched version of tuscany-sca-all which would work, but this however
leaves us in an awkward configuration position.
We're currently preparing a software release based around Tuscany
which is completely open to customers of our use of Tuscany, such that
we document fully how to construct services independent of our
software. As such, we do not ship any Tuscany artifacts and instead
encourage our customers to utilise the published maven repository.
Whilst requiring a patch version of one of the jars is possible; I
don't feel that this is a good representation of Tuscany - either
documenting a variant version or expecting a non-standard version of
1.2-incubating. These potential solutions are more likely to cause
issues for customers that would undermine the image of Tuscany that we
try to project.
Is anyone adamantly opposed to this release? Do you feel Tuscany
1.2.1 is still an option? I'd hope that given the potential to damage
our customer's perception of Tuscany would be enough to justify this
minor release.
Thanks for the clarifaction and explanation. It seems to me that
because we distribute Tuscany via Maven repos, which can't be patched,
this kind of situation will arise whenever a serious bug is found.
We can use patches to isolate a problem and confirm the fix, but we
generally won't be able to use them as an alternative to a release.
In a situation like this, unless a new release is imminent, the best
solution seems to be to produce a quick "bug fix" release without
incurring the overhead of a full release and testing cycle. Ant has
suggested that we could do this by applying a small set of carefully
controlled changes to the previous 1.2 release tag. I think we need
to be very strict about what changes go in, to avoid another experience
like 1.0.1. Specifically, I would suggest only including the fix
for TUSCANY-2304.
What do others think of this?
Simon
Cheers,
Dave.
On Tue, May 20, 2008 at 12:00 PM, Simon Nash <[EMAIL PROTECTED]> wrote:
Nishant Joshi wrote:
Hi All,
I have raised TUSCANY-2304 which was actually blocking me to go further
with
SCA client. So It was given high priority to resolved and fortunately Ant
has resolved it very fast, i appreciate his effortt, thanks alot Ant for
this :).
Another one was TUSCANY-2251 that was handled by Simon Nash and he has
also
done good progress on it (found from this list ). This problem came in
eclipse generated web service client (please refer it for more detail) so
this is also in high priority to get in next release. So i request to add
TUSCANY-2304 in 1.2.1 and if possible TUSCANY-2251 also.
One more thing, its very critical for us to get the next release 1.2.1
ASAP
(with 2304 and if possbile 2251 also :) ).
So I hope you can understand the effect of the TUSCANY-2304 for any
tuscany
SCA client user .
Hi Nishant,
The work to fix TUSCANY-2251 has turned out bigger than expected.
It's one of those cases where the first 80%-90% can be done quite
quickly but supporting the final 10%-20% of cases turns up many
issues, some of which require changes in other parts of the code.
I'm preparing a (large) checkin to update the new generator code
so that it handles most of the cases (perhaps 95%). This should be
enough to get the full build to run with the new code. However, I
wouldn't consider the new code to be ready to release at that point.
It will need quite a bit of further testing and a few more updates
to take care of the remaining 5% of cases. Some of these cases will
require discussion on the list to agree how they should be handled.
Also, the new code will need testing by people other than myself
with their scenarios to make sure that it does not break cases that
worked with the previous Java2WSDL generator.
For all these reasons, I think it will take about another 3 weeks
to get the new generator code to the state that I would be happy
to see it enabled in a release.
Regarding TUSCANY-2304, from other emails I see that Ant has sent
you a patched version of tuscany-sca-all-1.2-incubating.jar that
contains the fix for your problem. Can you explain why you need a
new release in addition to this patch?
Simon
