Enabled Maven snapshot releases for Woodstox 4.1 branch

[Cowtowncoder]

Quick note: I added support (in build-maven-deploy.xml) for doing
Maven snapshot builds, and also built a 4.1.0-SNAPSHOT (available from
Codehaus Maven2 snapshot repo, see
[http://xircles.codehaus.org/projects/woodstox] for details).

This should help testing pre-release versions of Woodstox, especially
4.1 when new functionality is added.

What new functionality? Current plans include:

(a) Add support for collecting (and exposing) bit more info on
attributes (exact Location, when enabled; type of quote used)
(b) Improve W3C Schema handling, by using features new MSV versions
expose; specifically to allow loading in and using multiple-namespace
(and thereby, multi-file) schemas: this is needed for supporting
common WSDL instances.

Anyway: please let me know if there are problems accessing these
builds; for what it's worth, directories look fine for me,

-+ Tatu +-

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email

[bimargulies]

Tatu,

I published snapshots the other week after puzzling out what you already had in there. I'll have to go see what you changed.

--benson

On Wed, Jun 17, 2009 at 10:16 PM, Tatu Saloranta <tsaloranta@...> wrote:

Quick note: I added support (in build-maven-deploy.xml) for doing
Maven snapshot builds, and also built a 4.1.0-SNAPSHOT (available from
Codehaus Maven2 snapshot repo, see
[http://xircles.codehaus.org/projects/woodstox] for details).

This should help testing pre-release versions of Woodstox, especially
4.1 when new functionality is added.

What new functionality? Current plans include:

(a) Add support for collecting (and exposing) bit more info on
attributes (exact Location, when enabled; type of quote used)
(b) Improve W3C Schema handling, by using features new MSV versions
expose; specifically to allow loading in and using multiple-namespace
(and thereby, multi-file) schemas: this is needed for supporting
common WSDL instances.

Anyway: please let me know if there are problems accessing these
builds; for what it's worth, directories look fine for me,

-+ Tatu +-

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

   http://xircles.codehaus.org/manage_email

[Cowtowncoder]

On Thu, Jun 18, 2009 at 4:33 AM, Benson Margulies<bimargulies@...> wrote:
> Tatu,
>
> I published snapshots the other week after puzzling out what you already had
> in there. I'll have to go see what you changed.
>

That explains the stuff that was already there -- I didn't connect the
dots last night (had bit of sleep deprivation :) ).

But what had not been working was:

- 'SNAPSHOT' not being included in version unless build.xml explicitly
defined it
- Unique timestamp not being added as part of deployment

but those are now resolved; figure out how to do it for Jackson, then
ported to Woodstox.

I am really looking forward to getting to work on 5.0 tho, since
that's regular Maven build, not Ant-wrapped one. :-)
(StaxMate is using Maven, and will change Jackson in due time too).

-+ Tatu +-

ps. Once again I have spent more time on Jackson than planned: this
time to get 1.1 out. But it might be worth it -- it'll support JAXB
annotations, and I suspect it might be useful for CXF too, for some
level of json support.

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email

[bimargulies]

Ah, I had turned off the unique timestamps because they haven't worked for me in the past at we don't use them at CXF. However, if they behave for you, no problem.

On Thu, Jun 18, 2009 at 1:08 PM, Tatu Saloranta <tsaloranta@...> wrote:

On Thu, Jun 18, 2009 at 4:33 AM, Benson Margulies<bimargulies@...> wrote:
> Tatu,
>
> I published snapshots the other week after puzzling out what you already had
> in there. I'll have to go see what you changed.
>

That explains the stuff that was already there -- I didn't connect the
dots last night (had bit of sleep deprivation :) ).

But what had not been working was:

- 'SNAPSHOT' not being included in version unless build.xml explicitly
defined it
- Unique timestamp not being added as part of deployment

but those are now resolved; figure out how to do it for Jackson, then
ported to Woodstox.

I am really looking forward to getting to work on 5.0 tho, since
that's regular Maven build, not Ant-wrapped one. :-)
(StaxMate is using Maven, and will change Jackson in due time too).

-+ Tatu +-

ps. Once again I have spent more time on Jackson than planned: this
time to get 1.1 out. But it might be worth it -- it'll support JAXB
annotations, and I suspect it might be useful for CXF too, for some
level of json support.

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

   http://xircles.codehaus.org/manage_email

[Cowtowncoder]

On Thu, Jun 18, 2009 at 10:30 AM, Benson Margulies<bimargulies@...> wrote:
> Ah, I had turned off the unique timestamps because they haven't worked for
> me in the past at we don't use them at CXF. However, if they behave for you,
> no problem.

Ok. I better test that then: I haven't actually tried using snapshot
versions, with or without unique timestamps. I assumed they are needed
to work around caching issues (assumed immutability of maven artifacts
etc).

If you do have problems I have no problems disabling addition of timestamps.

-+ Tatu +-

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email

[bimargulies]

In all the projects around CXF, we leave them turned off. mvn
automatically revalidates timestamps on snapshots.

On Thu, Jun 18, 2009 at 4:32 PM, Tatu Saloranta<tsaloranta@...> wrote:

> On Thu, Jun 18, 2009 at 10:30 AM, Benson Margulies<bimargulies@...> wrote:
>> Ah, I had turned off the unique timestamps because they haven't worked for
>> me in the past at we don't use them at CXF. However, if they behave for you,
>> no problem.
>
> Ok. I better test that then: I haven't actually tried using snapshot
> versions, with or without unique timestamps. I assumed they are needed
> to work around caching issues (assumed immutability of maven artifacts
> etc).
>
> If you do have problems I have no problems disabling addition of timestamps.
>
> -+ Tatu +-
>
> ---------------------------------------------------------------------
> To unsubscribe from this list, please visit:
>
>    http://xircles.codehaus.org/manage_email
>
>
>

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email

[Cowtowncoder]

On Thu, Jun 18, 2009 at 3:16 PM, Benson Margulies<bimargulies@...> wrote:
> In all the projects around CXF, we leave them turned off. mvn
> automatically revalidates timestamps on snapshots.

So that it doesn't cache blindly, assuming artifact never changes?
That is, making timestamps obsolete except if someone really wanted to
link against very specific build? (which is kind of silly for
snapshots anyway).

Put another way: leaving out timestamp should not cause problems when
rebuilding snapshots for given version?

-+ Tatu +-

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email

[bimargulies]

The model seems to be this: if you want to find it ever again, release it.

Anything that declares a dependency on a snapshot is willing to accept
that snapshot being 'improved' at any moment.

On Thu, Jun 18, 2009 at 7:34 PM, Tatu Saloranta<tsaloranta@...> wrote:

> On Thu, Jun 18, 2009 at 3:16 PM, Benson Margulies<bimargulies@...> wrote:
>> In all the projects around CXF, we leave them turned off. mvn
>> automatically revalidates timestamps on snapshots.
>
> So that it doesn't cache blindly, assuming artifact never changes?
> That is, making timestamps obsolete except if someone really wanted to
> link against very specific build? (which is kind of silly for
> snapshots anyway).
>
> Put another way: leaving out timestamp should not cause problems when
> rebuilding snapshots for given version?
>
> -+ Tatu +-
>
> ---------------------------------------------------------------------
> To unsubscribe from this list, please visit:
>
>    http://xircles.codehaus.org/manage_email
>
>
>

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email

[Cowtowncoder]

On Thu, Jun 18, 2009 at 4:42 PM, Benson Margulies<bimargulies@...> wrote:
> The model seems to be this: if you want to find it ever again, release it.
>
> Anything that declares a dependency on a snapshot is willing to accept
> that snapshot being 'improved' at any moment.
>

Yes, definitely, I was not arguing you should be able to do it (I
agree, snapshot is just a transient thing) -- just that that would be
the only reason to use a timestamp. That is, unless timestamp was not
needed to force Maven to re-check for status of the snapshot. I'm just
trying to understand intended usage of unique timestamp. Maybe I
should just google for comments by others users. :)

-+ Tatu +-

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email