Paul,
see inlined. :-)
Markus KARG wrote:
Well, this depends on what Paul and Marc want to make out of Jersey: If they
want to make it a collection of JAX-RS utilities ontop of a JAX-RS
implementation, then I do not see that a split is needed. Actually it was a
standalone project before (internally hosted by us, not publicly available)
and I asked whether I shall make a standalone project or contribute to
Jersey. It was Paul that said that he likes it to be part of Jersey. So I
have to wait for his decision how he will deal with the fact that it can run
with other JAX-RS implementations and has its own version schema and
livecycle.
My personal feelings about this -- if a module lives in a Jersey namespace
(Java package name and/or Maven group and artifact identifiers), ships with
Jersey (presumably a goal at some point in the development process), and has
its source somewhere in the Jersey trunk (presumably would happen with a
merge to trunk from the current branch), then that module should carry the
same version number (and release timelines) as the rest of Jersey. Further,
I would think the package names, as well as other conventions such as JDK
dependencies and common build plugins, should also match.
The fact that the module may or may not have any dependencies on Jersey
internals is an implementation detail (and that of course gets reflected in
the <dependencies> section of its POM) that may or may not change in the
future. That by itself should not be a deciding factor on versioning, or
for that matter naming.
If the module wants the "marketing" benefits :-) of being part of Jersey,
instead of being standalone, then it should really *be* part of Jersey.
Even if (at the moment) it will actually run on any JAX-RS implemenetation
(and therefore declare only the API jar in its dependencies).
This particular codebase is off to a very promising start, and is something
I would personally like to see in Jersey (with, as usual for me, some love
for the client side too :-). But I'd prefer to see it "all in" rather than
being perceived as straddling a fence.
I agree with Craig.
I would like to see this as part of Jersey under the same constraints as
other modules in the contribs. I assumed that this would be the case, but i
did not communicate this assumption before hand, but sometimes these
discussions take time to reveal themselves and solidify :-) as a result i
think we should change the contribution page to detail the constraints for
modules.
It was not clear that "contribution to jersey" means "contributing MODULES
to jersey that give up their indepency" (sorry for assuming that my code
could stay independent -- I am living in a federal nation so I just assumed
this principle. My fault.).
Markus, i think you have to decide whether you want the WebDAV module to be
part of Jersey or as a separate project. Obviously you know which decision i
would prefer but whatever decision you make you have my support in terms of
helping out.
I also want the project to stay part of the Jersey project, but not
necessarily of the Jersey product, but I do not agree with what this means
in the end for users of other JAX-RS implementations. Beeing "a module"
means to give up the idea that other JAX-RS implementations can run my
WebDAV code, or at least, it would be harder to identify what JARs are
needed and what version number contains which changes. I do not want that
WebDAV is "a non-standalone module", but it is and shall stay ONTOP of the
jersey core (why not? what would be the benefit of doing that for the
user?).
So actually for me this reads like "don't be easily useable with other
JAX-RS implementations or we will kick you out", which would be a rather
strange decision for an open project (or maybe this is not an open project
but just a project producing open source? Sorry, my fault.).
Why can't we just add a folder to the SVN tree that contains JAX-RS related,
standalone projects which can run with any JAX-RS implementation but are not
forced to have the Java Version and project version be in line with Jersey's
core? That would be an easy, simple solution with which we all could be
live, and I would be happy if WebDAV would not be the only one -- ATOM
support would be absolutely correctly located there, too, and a lot of other
JAX-RS implementation users would be happy to download jersey-atom.jar and
jersey-webdav.jar and use it with any implementation they like. :-)
What do you think about this?
Regards
Markus
Paul.
Craig
Regards
Markus
From: Edelson, Justin [mailto:Justin.Edelson_at_mtvstaff.com]
Sent: Samstag, 24. Januar 2009 22:45
To: dev_at_jersey.dev.java.net
Subject: RE: [Jersey] WebDAV pom.xml
Just my 2 cents (and feel free to ignore), but it seems like this needs to
be separated from Jersey and made a standalone project at java.net (or
somewhere else).
Justin
_____
From: Markus KARG [mailto:markus.karg_at_gmx.net]
Sent: Sat 1/24/2009 11:58 AM
To: dev_at_jersey.dev.java.net
Subject: RE: [Jersey] WebDAV pom.xml
Paul,
first of all, thank you for your ideas and sorry for not answering earlier.
Once more I catched a cold and so I did not find the power to spend my
evenings with complex issues. More inline.
> Every time we release Jersey we do the following:
> - make a tag of the trunk;
> - change the version in the trunk of all modules from "xxx-
> SNAPSHOT" to "yyy-SNAPSHOT"; and
> - change the version in the tag of all modules to "xxx" and deploy
> the maven artifacts to the repo.
> It makes it a lot easier to manage releases if all module versions are
> in sync and we release at the same time. Generally all Jersey modules
> tend to have a coupling of Jersey APIs where we also tend to, at least
> currently, have a fast update between components to fix issues and
> support features. So i would prefer to retain this approach.
> Do you envisage having a different release cycle to that of Jersey as
> well as a different version number?
> Maybe the contribs area is not really suitable in this respect and we
> need another area separate from Jersey for components that can be
> versioned and released independently? e.g.
> trunk/components
> I think that may better suit your requirements based on what you say
> above and below. i.e. i don't want to unduly constrain your
> development and what you depend on. We can set up separate Hudson jobs
> to build components.
In fact, WebDAV stuff is not related to Jersey in any way. The project was
build solely on JAX-RS 1.0, and does not know about any Jersey specific
thing. So from a technical aspect, it is not a Jersey component in the
narrow sense, but more a usage of Jersey, just as it could be a usage of any
other JAX-RS 1.0 implementation. This is the reason why the release politics
makes sense to stay different from Jersey's. WebDAV just has no technical
need to keep anything in sync with Jersey. The problem is that it seems
there is no really good place to put it. Actually WebDAV is neither a
contrib to the Jersey engine while it IS one to the Jersey project as a pool
of RESTful technology projects, and it neither is a component of the Jersey
engine. I wouldn't say that it would be best to put it into its own top
level project, but maybe it would be good to have a new folder containing
projects like mine which just USES JAX-RS but not directly Jersey, and which
could be downloaded separately? I'd like people of other JAX-RS
implementations being able to use WebDAV, and that would be the best way for
that. So Jersey could be two things: A JAX-RS implementation with additional
features PLUS a set of JAX-RS components independent of the JAX-RS
implementation. This would also be a good place to put Daniel Manzke's
"Microsoft Interoperability" stuff. What do you think about that?
> Note that JAX-RS does require support on SE 5, so the additional
> Jersey modules require it as well. In IDEs (at least in NetBeans) you
> can set a project to use SE 6 but compile SE 5 constrained source to
> catch errors.
Actually what your wrote into the spec is that SE 5 OR LATER is needed, so
it is not a constraint to exactly match SE 5 in all projects that are USING
JAX-RS -- and my code just USES Jersey. My WebDAV code targets in Java EE 6
contained JAX-RS, so SE 6 is what I develop and test upon (and what I can
afford -- I have no time to explicitly test drive on SE 5). Again, WebDAV is
nothing that is inside of Jersey, and it sits ontop of it, so if somebody
wants to use WebDAV he must use SE 6 (which is not a problem in times of SE
6 everywhere and EE 6 published soon). I do not see the actual problem right
now. Can you elaborate on this so I could understand (anyways I have no time
to check SE 5, and my Eclipse will not know what APIs are not existing in SE
5 unless I install SE 5 in addition to SE 6, which I just will not due since
I do not understand the need)?
> Jakub will know more, he is our Maven guru :-)
Did not hear anything from him so far, so it seems it was OK what I did.
> > As "1.0" effectively means
> > "[1.0,)" I
> > do not see why it is better than my explicit "[1.0,)"...?!
> Oh! i guess i do not understand maven version declarations :-) i want
> to be clear under what conditions Jersey has been tested against but
> did not want to necessarily restrict developers to using other
> versions.
Then you should change from your "1.0" (i. e. "1.0, or later than 1.0") to
my "[1.0;2.0)" (i. e. "anything in the range from 1.0 to but not including
2.0"), which more clearly says that API changes (2.x, 3.x) are not accepted,
while bug fixes are.
Have Fun
Markus
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_jersey.dev.java.net
For additional commands, e-mail: dev-help_at_jersey.dev.java.net