You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@tuscany.apache.org by kelvin goodson <kg...@thegoodsons.org.uk> on 2010/06/01 15:14:36 UTC
releasing the maven-tuscany-plugin
Ant mentioned that the maven-tuscany-plugin should be released in
parallel with the sca-java 2.0 M5 release. Is there any reason why I
shouldn't go ahead and cut a branch for this?
Kelvin.
Re: releasing the maven-tuscany-plugin
Posted by kelvin goodson <kg...@thegoodsons.org.uk>.
Having had a discussion with ant on IRC, it seems clear the profile
approach won't work. Without the fix in maven 3.0 alpha, no matter
what order we put modules into a profile, maven's dependency
resolution mechanism is going to end up putting the cart before the
horse. So that leaves us with Hobson's choice ... the documented
process that when bumping the version number, the only way to get a
clean build first time is first to build in modules, then build the
plugins and then in samples.
However, it's still seems advantageous to bring the
maven-tuscany-plugin into the sca-java-2.x to save the pain of
multiple branches and releases. Is there any reason why we need to
retain the flexibility of being able to release that plugin
independently?
Kelvin.
On Tue, Jun 1, 2010 at 3:59 PM, kelvin goodson <kg...@thegoodsons.org.uk> wrote:
> On Tue, Jun 1, 2010 at 3:47 PM, Simon Laws <si...@googlemail.com> wrote:
>> On Tue, Jun 1, 2010 at 3:22 PM, ant elder <an...@gmail.com> wrote:
>>> On Tue, Jun 1, 2010 at 2:48 PM, Simon Laws <si...@googlemail.com> wrote:
>>>> On Tue, Jun 1, 2010 at 2:14 PM, kelvin goodson <kg...@thegoodsons.org.uk> wrote:
>>>>> Ant mentioned that the maven-tuscany-plugin should be released in
>>>>> parallel with the sca-java 2.0 M5 release. Is there any reason why I
>>>>> shouldn't go ahead and cut a branch for this?
>>>>>
>>>>> Kelvin.
>>>>>
>>>>
>>>> I don't expect so but if this plugin and Tuscany trunk depend on one
>>>> another why don't we pull this particular plugin into trunk otherwise
>>>> it'll be a problem every time we try and release. It feels like make
>>>> work.
>>>>
>>>
>>> The main reason we don't do this was because there didn't used to be a
>>> way to have the Maven reactor build the plugins before there were
>>> used. That meant that either the build would fail or you needed a
>>> separate build step( eg -Psetup) or if you got "lucky" the build would
>>> work but it would be using the previous version of the plugin so any
>>> problems didn't get found until the next build which used to cause all
>>> sorts of confusion. It looks like this has been fixed in Maven now -
>>> http://jira.codehaus.org/browse/MNG-2135 - so we could try moving it
>>> to be part of trunk. I shall have a look...
>>>
>>> ...ant
>>>
>>
>> That was marked as fixed in 3.0Alpha so that maybe a concern. If that
>> gives problems we should stick with what we have for the short term
>> and document a process (or a profile) which gets the
>> modules/plugin/samples build in the right order
>
> given that we have this relatively simple way of decoupling this
> issue, I'd prefer to avoid trying out the alpha version of maven and
> go ahead with the profile approach,
>
> Kelvin.
>
>>
>>
>> --
>> Apache Tuscany committer: tuscany.apache.org
>> Co-author of a book about Tuscany and SCA: tuscanyinaction.com
>>
>
Re: releasing the maven-tuscany-plugin
Posted by kelvin goodson <kg...@thegoodsons.org.uk>.
On Tue, Jun 1, 2010 at 3:47 PM, Simon Laws <si...@googlemail.com> wrote:
> On Tue, Jun 1, 2010 at 3:22 PM, ant elder <an...@gmail.com> wrote:
>> On Tue, Jun 1, 2010 at 2:48 PM, Simon Laws <si...@googlemail.com> wrote:
>>> On Tue, Jun 1, 2010 at 2:14 PM, kelvin goodson <kg...@thegoodsons.org.uk> wrote:
>>>> Ant mentioned that the maven-tuscany-plugin should be released in
>>>> parallel with the sca-java 2.0 M5 release. Is there any reason why I
>>>> shouldn't go ahead and cut a branch for this?
>>>>
>>>> Kelvin.
>>>>
>>>
>>> I don't expect so but if this plugin and Tuscany trunk depend on one
>>> another why don't we pull this particular plugin into trunk otherwise
>>> it'll be a problem every time we try and release. It feels like make
>>> work.
>>>
>>
>> The main reason we don't do this was because there didn't used to be a
>> way to have the Maven reactor build the plugins before there were
>> used. That meant that either the build would fail or you needed a
>> separate build step( eg -Psetup) or if you got "lucky" the build would
>> work but it would be using the previous version of the plugin so any
>> problems didn't get found until the next build which used to cause all
>> sorts of confusion. It looks like this has been fixed in Maven now -
>> http://jira.codehaus.org/browse/MNG-2135 - so we could try moving it
>> to be part of trunk. I shall have a look...
>>
>> ...ant
>>
>
> That was marked as fixed in 3.0Alpha so that maybe a concern. If that
> gives problems we should stick with what we have for the short term
> and document a process (or a profile) which gets the
> modules/plugin/samples build in the right order
given that we have this relatively simple way of decoupling this
issue, I'd prefer to avoid trying out the alpha version of maven and
go ahead with the profile approach,
Kelvin.
>
>
> --
> Apache Tuscany committer: tuscany.apache.org
> Co-author of a book about Tuscany and SCA: tuscanyinaction.com
>
Re: releasing the maven-tuscany-plugin
Posted by Simon Laws <si...@googlemail.com>.
On Tue, Jun 1, 2010 at 3:22 PM, ant elder <an...@gmail.com> wrote:
> On Tue, Jun 1, 2010 at 2:48 PM, Simon Laws <si...@googlemail.com> wrote:
>> On Tue, Jun 1, 2010 at 2:14 PM, kelvin goodson <kg...@thegoodsons.org.uk> wrote:
>>> Ant mentioned that the maven-tuscany-plugin should be released in
>>> parallel with the sca-java 2.0 M5 release. Is there any reason why I
>>> shouldn't go ahead and cut a branch for this?
>>>
>>> Kelvin.
>>>
>>
>> I don't expect so but if this plugin and Tuscany trunk depend on one
>> another why don't we pull this particular plugin into trunk otherwise
>> it'll be a problem every time we try and release. It feels like make
>> work.
>>
>
> The main reason we don't do this was because there didn't used to be a
> way to have the Maven reactor build the plugins before there were
> used. That meant that either the build would fail or you needed a
> separate build step( eg -Psetup) or if you got "lucky" the build would
> work but it would be using the previous version of the plugin so any
> problems didn't get found until the next build which used to cause all
> sorts of confusion. It looks like this has been fixed in Maven now -
> http://jira.codehaus.org/browse/MNG-2135 - so we could try moving it
> to be part of trunk. I shall have a look...
>
> ...ant
>
That was marked as fixed in 3.0Alpha so that maybe a concern. If that
gives problems we should stick with what we have for the short term
and document a process (or a profile) which gets the
modules/plugin/samples build in the right order.
build
--
Apache Tuscany committer: tuscany.apache.org
Co-author of a book about Tuscany and SCA: tuscanyinaction.com
Re: releasing the maven-tuscany-plugin
Posted by ant elder <an...@gmail.com>.
On Tue, Jun 1, 2010 at 2:48 PM, Simon Laws <si...@googlemail.com> wrote:
> On Tue, Jun 1, 2010 at 2:14 PM, kelvin goodson <kg...@thegoodsons.org.uk> wrote:
>> Ant mentioned that the maven-tuscany-plugin should be released in
>> parallel with the sca-java 2.0 M5 release. Is there any reason why I
>> shouldn't go ahead and cut a branch for this?
>>
>> Kelvin.
>>
>
> I don't expect so but if this plugin and Tuscany trunk depend on one
> another why don't we pull this particular plugin into trunk otherwise
> it'll be a problem every time we try and release. It feels like make
> work.
>
The main reason we don't do this was because there didn't used to be a
way to have the Maven reactor build the plugins before there were
used. That meant that either the build would fail or you needed a
separate build step( eg -Psetup) or if you got "lucky" the build would
work but it would be using the previous version of the plugin so any
problems didn't get found until the next build which used to cause all
sorts of confusion. It looks like this has been fixed in Maven now -
http://jira.codehaus.org/browse/MNG-2135 - so we could try moving it
to be part of trunk. I shall have a look...
...ant
Re: releasing the maven-tuscany-plugin
Posted by kelvin goodson <kg...@thegoodsons.org.uk>.
that would make life much easier; perhaps a plugins directory in the
sca-java-2.x trunk, where coreq plugins reside (leaving non-coreq
plugins where they are)
Kelvin.
On Tue, Jun 1, 2010 at 2:48 PM, Simon Laws <si...@googlemail.com> wrote:
> On Tue, Jun 1, 2010 at 2:14 PM, kelvin goodson <kg...@thegoodsons.org.uk> wrote:
>> Ant mentioned that the maven-tuscany-plugin should be released in
>> parallel with the sca-java 2.0 M5 release. Is there any reason why I
>> shouldn't go ahead and cut a branch for this?
>>
>> Kelvin.
>>
>
> I don't expect so but if this plugin and Tuscany trunk depend on one
> another why don't we pull this particular plugin into trunk otherwise
> it'll be a problem every time we try and release. It feels like make
> work.
>
> Simon
>
>
> --
> Apache Tuscany committer: tuscany.apache.org
> Co-author of a book about Tuscany and SCA: tuscanyinaction.com
>
Re: releasing the maven-tuscany-plugin
Posted by Simon Laws <si...@googlemail.com>.
On Tue, Jun 1, 2010 at 2:14 PM, kelvin goodson <kg...@thegoodsons.org.uk> wrote:
> Ant mentioned that the maven-tuscany-plugin should be released in
> parallel with the sca-java 2.0 M5 release. Is there any reason why I
> shouldn't go ahead and cut a branch for this?
>
> Kelvin.
>
I don't expect so but if this plugin and Tuscany trunk depend on one
another why don't we pull this particular plugin into trunk otherwise
it'll be a problem every time we try and release. It feels like make
work.
Simon
--
Apache Tuscany committer: tuscany.apache.org
Co-author of a book about Tuscany and SCA: tuscanyinaction.com