Trunk build broken

[Jean-Baptiste Onofré <jb...@nanthrax.net>]

Hi all,

I raised the KARAF-451 this morning to track the fact that the trunk 
build is broken.

After digging a little more, I can see that the 2.2.x branch and trunk 
don't use at all the same assembly structure:
- the 2.2.x branch uses "new" assemblies structure (containing 
distribution and features)
- the trunk uses "old" assembly only with distribution and another 
assembly in the features repo.

I'm not very pleased with that because it means that changes on the 
2.2.x branch have not been merged to the trunk.

I think it's important that all changes are first performed on trunk and 
merge into a target branch, but not the opposite.

If a kind of POC branch is required to test some enhancement before 
applying on trunk, it's not a problem, but the name should show that 
it's a POC branch (why not create a branch name containing the 
corresponding Jira id ?).

Now, it's not easy to know if the "correct" structure is the branch 
2.2.x or the trunk.

I would like to fix the trunk build and I have two ways to do it:
1/ only update the assembly by deleting war deployer and manual dependencies
2/ apply the new assemblies structure as we have on 2.2.x branch

Could you light my way ?

Thanks
Regards
JB

Re: Trunk build broken

[Jean-Baptiste Onofré <jb...@nanthrax.net>]

Indeed, I've just made a cleanup around Karaf deployer.war references 
(in POMs).

Regards
JB

On 02/09/2011 09:33 AM, Guillaume Nodet wrote:
> I think we already use it, it's just that we still have references to
> the karaf one, even if we have deleted that module.
>
> On Wed, Feb 9, 2011 at 09:19, Achim Nierbeck<bc...@googlemail.com>  wrote:
>> Please do so, the pax-web deployer is much more improved :)
>>
>> Regards, Achim
>>
>> 2011/2/9 Jean-Baptiste Onofré<jb...@nanthrax.net>:
>>> Just another thing that I'm gonna change: the war deployer is no more
>>> available (replaced by Pax Web), but both trunk and 2.2.x branch refer it in
>>> assembly and POMs.
>>> I'm gonna remove war deployer references.
>>>
>>> Regards
>>> JB
>>>
>>> On 02/09/2011 08:32 AM, Jean-Baptiste Onofré wrote:
>>>>
>>>> It's not a big deal :)
>>>> Just not a "good" situation :)
>>>>
>>>> I propose the following:
>>>> 1/ I fix the trunk as it's (with the "old" assembly)
>>>> 2/ If we state that the "good" assembly structure if the new one, I
>>>> apply on the trunk as it has been performed on 2.2.x branch.
>>>>
>>>> Anyway, the broken build on trunk is not directly linked to this, it's
>>>> just digging around assembly that I saw that :)
>>>>
>>>> Regards
>>>> JB
>>>>
>>>> On 02/09/2011 08:27 AM, Guillaume Nodet wrote:
>>>>>
>>>>> Bad Guillaume!
>>>>>
>>>>> I agree with you on the concept of branch to test features, but that's
>>>>> not really the case here, it's just that I forgot to apply the same
>>>>> changes on both trunk.
>>>>> I did the changes on 2.2.x branch and planned to backport them to
>>>>> trunk, but haven't gone through it, though I don't really see how/why
>>>>> it's related to trunk being broken as I haven't done any modifications
>>>>> in trunk. I think the new structure as we've discussed on the mailing
>>>>> list makes more sense, so I'll work on that today.
>>>>>
>>>>>
>>>>> On Wed, Feb 9, 2011 at 08:14, Jean-Baptiste Onofré<jb...@nanthrax.net>
>>>>> wrote:
>>>>>>
>>>>>> Hi all,
>>>>>>
>>>>>> I raised the KARAF-451 this morning to track the fact that the trunk
>>>>>> build
>>>>>> is broken.
>>>>>>
>>>>>> After digging a little more, I can see that the 2.2.x branch and
>>>>>> trunk don't
>>>>>> use at all the same assembly structure:
>>>>>> - the 2.2.x branch uses "new" assemblies structure (containing
>>>>>> distribution
>>>>>> and features)
>>>>>> - the trunk uses "old" assembly only with distribution and another
>>>>>> assembly
>>>>>> in the features repo.
>>>>>>
>>>>>> I'm not very pleased with that because it means that changes on the
>>>>>> 2.2.x
>>>>>> branch have not been merged to the trunk.
>>>>>>
>>>>>> I think it's important that all changes are first performed on trunk and
>>>>>> merge into a target branch, but not the opposite.
>>>>>>
>>>>>> If a kind of POC branch is required to test some enhancement before
>>>>>> applying
>>>>>> on trunk, it's not a problem, but the name should show that it's a POC
>>>>>> branch (why not create a branch name containing the corresponding
>>>>>> Jira id
>>>>>> ?).
>>>>>>
>>>>>> Now, it's not easy to know if the "correct" structure is the branch
>>>>>> 2.2.x or
>>>>>> the trunk.
>>>>>>
>>>>>> I would like to fix the trunk build and I have two ways to do it:
>>>>>> 1/ only update the assembly by deleting war deployer and manual
>>>>>> dependencies
>>>>>> 2/ apply the new assemblies structure as we have on 2.2.x branch
>>>>>>
>>>>>> Could you light my way ?
>>>>>>
>>>>>> Thanks
>>>>>> Regards
>>>>>> JB
>>>>>>
>>>>>
>>>>>
>>>>>
>>>
>>
>
>
>

Re: Trunk build broken

[Guillaume Nodet <gn...@gmail.com>]

I think we already use it, it's just that we still have references to
the karaf one, even if we have deleted that module.

On Wed, Feb 9, 2011 at 09:19, Achim Nierbeck <bc...@googlemail.com> wrote:
> Please do so, the pax-web deployer is much more improved :)
>
> Regards, Achim
>
> 2011/2/9 Jean-Baptiste Onofré <jb...@nanthrax.net>:
>> Just another thing that I'm gonna change: the war deployer is no more
>> available (replaced by Pax Web), but both trunk and 2.2.x branch refer it in
>> assembly and POMs.
>> I'm gonna remove war deployer references.
>>
>> Regards
>> JB
>>
>> On 02/09/2011 08:32 AM, Jean-Baptiste Onofré wrote:
>>>
>>> It's not a big deal :)
>>> Just not a "good" situation :)
>>>
>>> I propose the following:
>>> 1/ I fix the trunk as it's (with the "old" assembly)
>>> 2/ If we state that the "good" assembly structure if the new one, I
>>> apply on the trunk as it has been performed on 2.2.x branch.
>>>
>>> Anyway, the broken build on trunk is not directly linked to this, it's
>>> just digging around assembly that I saw that :)
>>>
>>> Regards
>>> JB
>>>
>>> On 02/09/2011 08:27 AM, Guillaume Nodet wrote:
>>>>
>>>> Bad Guillaume!
>>>>
>>>> I agree with you on the concept of branch to test features, but that's
>>>> not really the case here, it's just that I forgot to apply the same
>>>> changes on both trunk.
>>>> I did the changes on 2.2.x branch and planned to backport them to
>>>> trunk, but haven't gone through it, though I don't really see how/why
>>>> it's related to trunk being broken as I haven't done any modifications
>>>> in trunk. I think the new structure as we've discussed on the mailing
>>>> list makes more sense, so I'll work on that today.
>>>>
>>>>
>>>> On Wed, Feb 9, 2011 at 08:14, Jean-Baptiste Onofré<jb...@nanthrax.net>
>>>> wrote:
>>>>>
>>>>> Hi all,
>>>>>
>>>>> I raised the KARAF-451 this morning to track the fact that the trunk
>>>>> build
>>>>> is broken.
>>>>>
>>>>> After digging a little more, I can see that the 2.2.x branch and
>>>>> trunk don't
>>>>> use at all the same assembly structure:
>>>>> - the 2.2.x branch uses "new" assemblies structure (containing
>>>>> distribution
>>>>> and features)
>>>>> - the trunk uses "old" assembly only with distribution and another
>>>>> assembly
>>>>> in the features repo.
>>>>>
>>>>> I'm not very pleased with that because it means that changes on the
>>>>> 2.2.x
>>>>> branch have not been merged to the trunk.
>>>>>
>>>>> I think it's important that all changes are first performed on trunk and
>>>>> merge into a target branch, but not the opposite.
>>>>>
>>>>> If a kind of POC branch is required to test some enhancement before
>>>>> applying
>>>>> on trunk, it's not a problem, but the name should show that it's a POC
>>>>> branch (why not create a branch name containing the corresponding
>>>>> Jira id
>>>>> ?).
>>>>>
>>>>> Now, it's not easy to know if the "correct" structure is the branch
>>>>> 2.2.x or
>>>>> the trunk.
>>>>>
>>>>> I would like to fix the trunk build and I have two ways to do it:
>>>>> 1/ only update the assembly by deleting war deployer and manual
>>>>> dependencies
>>>>> 2/ apply the new assemblies structure as we have on 2.2.x branch
>>>>>
>>>>> Could you light my way ?
>>>>>
>>>>> Thanks
>>>>> Regards
>>>>> JB
>>>>>
>>>>
>>>>
>>>>
>>
>



-- 
Cheers,
Guillaume Nodet
------------------------
Blog: http://gnodet.blogspot.com/
------------------------
Open Source SOA
http://fusesource.com

Re: Trunk build broken

[Achim Nierbeck <bc...@googlemail.com>]

Please do so, the pax-web deployer is much more improved :)

Regards, Achim

2011/2/9 Jean-Baptiste Onofré <jb...@nanthrax.net>:
> Just another thing that I'm gonna change: the war deployer is no more
> available (replaced by Pax Web), but both trunk and 2.2.x branch refer it in
> assembly and POMs.
> I'm gonna remove war deployer references.
>
> Regards
> JB
>
> On 02/09/2011 08:32 AM, Jean-Baptiste Onofré wrote:
>>
>> It's not a big deal :)
>> Just not a "good" situation :)
>>
>> I propose the following:
>> 1/ I fix the trunk as it's (with the "old" assembly)
>> 2/ If we state that the "good" assembly structure if the new one, I
>> apply on the trunk as it has been performed on 2.2.x branch.
>>
>> Anyway, the broken build on trunk is not directly linked to this, it's
>> just digging around assembly that I saw that :)
>>
>> Regards
>> JB
>>
>> On 02/09/2011 08:27 AM, Guillaume Nodet wrote:
>>>
>>> Bad Guillaume!
>>>
>>> I agree with you on the concept of branch to test features, but that's
>>> not really the case here, it's just that I forgot to apply the same
>>> changes on both trunk.
>>> I did the changes on 2.2.x branch and planned to backport them to
>>> trunk, but haven't gone through it, though I don't really see how/why
>>> it's related to trunk being broken as I haven't done any modifications
>>> in trunk. I think the new structure as we've discussed on the mailing
>>> list makes more sense, so I'll work on that today.
>>>
>>>
>>> On Wed, Feb 9, 2011 at 08:14, Jean-Baptiste Onofré<jb...@nanthrax.net>
>>> wrote:
>>>>
>>>> Hi all,
>>>>
>>>> I raised the KARAF-451 this morning to track the fact that the trunk
>>>> build
>>>> is broken.
>>>>
>>>> After digging a little more, I can see that the 2.2.x branch and
>>>> trunk don't
>>>> use at all the same assembly structure:
>>>> - the 2.2.x branch uses "new" assemblies structure (containing
>>>> distribution
>>>> and features)
>>>> - the trunk uses "old" assembly only with distribution and another
>>>> assembly
>>>> in the features repo.
>>>>
>>>> I'm not very pleased with that because it means that changes on the
>>>> 2.2.x
>>>> branch have not been merged to the trunk.
>>>>
>>>> I think it's important that all changes are first performed on trunk and
>>>> merge into a target branch, but not the opposite.
>>>>
>>>> If a kind of POC branch is required to test some enhancement before
>>>> applying
>>>> on trunk, it's not a problem, but the name should show that it's a POC
>>>> branch (why not create a branch name containing the corresponding
>>>> Jira id
>>>> ?).
>>>>
>>>> Now, it's not easy to know if the "correct" structure is the branch
>>>> 2.2.x or
>>>> the trunk.
>>>>
>>>> I would like to fix the trunk build and I have two ways to do it:
>>>> 1/ only update the assembly by deleting war deployer and manual
>>>> dependencies
>>>> 2/ apply the new assemblies structure as we have on 2.2.x branch
>>>>
>>>> Could you light my way ?
>>>>
>>>> Thanks
>>>> Regards
>>>> JB
>>>>
>>>
>>>
>>>
>

Re: Trunk build broken

[Jean-Baptiste Onofré <jb...@nanthrax.net>]

Just another thing that I'm gonna change: the war deployer is no more 
available (replaced by Pax Web), but both trunk and 2.2.x branch refer 
it in assembly and POMs.
I'm gonna remove war deployer references.

Regards
JB

On 02/09/2011 08:32 AM, Jean-Baptiste Onofré wrote:
> It's not a big deal :)
> Just not a "good" situation :)
>
> I propose the following:
> 1/ I fix the trunk as it's (with the "old" assembly)
> 2/ If we state that the "good" assembly structure if the new one, I
> apply on the trunk as it has been performed on 2.2.x branch.
>
> Anyway, the broken build on trunk is not directly linked to this, it's
> just digging around assembly that I saw that :)
>
> Regards
> JB
>
> On 02/09/2011 08:27 AM, Guillaume Nodet wrote:
>> Bad Guillaume!
>>
>> I agree with you on the concept of branch to test features, but that's
>> not really the case here, it's just that I forgot to apply the same
>> changes on both trunk.
>> I did the changes on 2.2.x branch and planned to backport them to
>> trunk, but haven't gone through it, though I don't really see how/why
>> it's related to trunk being broken as I haven't done any modifications
>> in trunk. I think the new structure as we've discussed on the mailing
>> list makes more sense, so I'll work on that today.
>>
>>
>> On Wed, Feb 9, 2011 at 08:14, Jean-Baptiste Onofré<jb...@nanthrax.net>
>> wrote:
>>> Hi all,
>>>
>>> I raised the KARAF-451 this morning to track the fact that the trunk
>>> build
>>> is broken.
>>>
>>> After digging a little more, I can see that the 2.2.x branch and
>>> trunk don't
>>> use at all the same assembly structure:
>>> - the 2.2.x branch uses "new" assemblies structure (containing
>>> distribution
>>> and features)
>>> - the trunk uses "old" assembly only with distribution and another
>>> assembly
>>> in the features repo.
>>>
>>> I'm not very pleased with that because it means that changes on the
>>> 2.2.x
>>> branch have not been merged to the trunk.
>>>
>>> I think it's important that all changes are first performed on trunk and
>>> merge into a target branch, but not the opposite.
>>>
>>> If a kind of POC branch is required to test some enhancement before
>>> applying
>>> on trunk, it's not a problem, but the name should show that it's a POC
>>> branch (why not create a branch name containing the corresponding
>>> Jira id
>>> ?).
>>>
>>> Now, it's not easy to know if the "correct" structure is the branch
>>> 2.2.x or
>>> the trunk.
>>>
>>> I would like to fix the trunk build and I have two ways to do it:
>>> 1/ only update the assembly by deleting war deployer and manual
>>> dependencies
>>> 2/ apply the new assemblies structure as we have on 2.2.x branch
>>>
>>> Could you light my way ?
>>>
>>> Thanks
>>> Regards
>>> JB
>>>
>>
>>
>>

Re: Trunk build broken

[Guillaume Nodet <gn...@gmail.com>]

On Wed, Feb 9, 2011 at 08:32, Jean-Baptiste Onofré <jb...@nanthrax.net> wrote:
> It's not a big deal :)
> Just not a "good" situation :)
>
> I propose the following:
> 1/ I fix the trunk as it's (with the "old" assembly)

Sounds good.  I think both things are really unrelated.

> 2/ If we state that the "good" assembly structure if the new one, I apply on
> the trunk as it has been performed on 2.2.x branch.

Yeah, I didn't backported it because I have no time to do it.  The
problem is that it's not a simple 'git cherry-pick' thing, as on
2.2.x, i've removed the new artifacts that use the new maven
packagings, which we definitely want to keep in trunk.

>
> Anyway, the broken build on trunk is not directly linked to this, it's just
> digging around assembly that I saw that :)
>
> Regards
> JB
>
> On 02/09/2011 08:27 AM, Guillaume Nodet wrote:
>>
>> Bad Guillaume!
>>
>> I agree with you on the concept of branch to test features, but that's
>> not really the case here, it's just that I forgot to apply the same
>> changes on both trunk.
>> I did the changes on 2.2.x branch and planned to backport them to
>> trunk, but haven't gone through it, though I don't really see how/why
>> it's related to trunk being broken as I haven't done any modifications
>> in trunk.  I think the new structure as we've discussed on the mailing
>> list makes more sense, so I'll work on that today.
>>
>>
>> On Wed, Feb 9, 2011 at 08:14, Jean-Baptiste Onofré<jb...@nanthrax.net>
>>  wrote:
>>>
>>> Hi all,
>>>
>>> I raised the KARAF-451 this morning to track the fact that the trunk
>>> build
>>> is broken.
>>>
>>> After digging a little more, I can see that the 2.2.x branch and trunk
>>> don't
>>> use at all the same assembly structure:
>>> - the 2.2.x branch uses "new" assemblies structure (containing
>>> distribution
>>> and features)
>>> - the trunk uses "old" assembly only with distribution and another
>>> assembly
>>> in the features repo.
>>>
>>> I'm not very pleased with that because it means that changes on the 2.2.x
>>> branch have not been merged to the trunk.
>>>
>>> I think it's important that all changes are first performed on trunk and
>>> merge into a target branch, but not the opposite.
>>>
>>> If a kind of POC branch is required to test some enhancement before
>>> applying
>>> on trunk, it's not a problem, but the name should show that it's a POC
>>> branch (why not create a branch name containing the corresponding Jira id
>>> ?).
>>>
>>> Now, it's not easy to know if the "correct" structure is the branch 2.2.x
>>> or
>>> the trunk.
>>>
>>> I would like to fix the trunk build and I have two ways to do it:
>>> 1/ only update the assembly by deleting war deployer and manual
>>> dependencies
>>> 2/ apply the new assemblies structure as we have on 2.2.x branch
>>>
>>> Could you light my way ?
>>>
>>> Thanks
>>> Regards
>>> JB
>>>
>>
>>
>>
>



-- 
Cheers,
Guillaume Nodet
------------------------
Blog: http://gnodet.blogspot.com/
------------------------
Open Source SOA
http://fusesource.com

Re: Trunk build broken

[Jean-Baptiste Onofré <jb...@nanthrax.net>]

It's not a big deal :)
Just not a "good" situation :)

I propose the following:
1/ I fix the trunk as it's (with the "old" assembly)
2/ If we state that the "good" assembly structure if the new one, I 
apply on the trunk as it has been performed on 2.2.x branch.

Anyway, the broken build on trunk is not directly linked to this, it's 
just digging around assembly that I saw that :)

Regards
JB

On 02/09/2011 08:27 AM, Guillaume Nodet wrote:
> Bad Guillaume!
>
> I agree with you on the concept of branch to test features, but that's
> not really the case here, it's just that I forgot to apply the same
> changes on both trunk.
> I did the changes on 2.2.x branch and planned to backport them to
> trunk, but haven't gone through it, though I don't really see how/why
> it's related to trunk being broken as I haven't done any modifications
> in trunk.  I think the new structure as we've discussed on the mailing
> list makes more sense, so I'll work on that today.
>
>
> On Wed, Feb 9, 2011 at 08:14, Jean-Baptiste Onofré<jb...@nanthrax.net>  wrote:
>> Hi all,
>>
>> I raised the KARAF-451 this morning to track the fact that the trunk build
>> is broken.
>>
>> After digging a little more, I can see that the 2.2.x branch and trunk don't
>> use at all the same assembly structure:
>> - the 2.2.x branch uses "new" assemblies structure (containing distribution
>> and features)
>> - the trunk uses "old" assembly only with distribution and another assembly
>> in the features repo.
>>
>> I'm not very pleased with that because it means that changes on the 2.2.x
>> branch have not been merged to the trunk.
>>
>> I think it's important that all changes are first performed on trunk and
>> merge into a target branch, but not the opposite.
>>
>> If a kind of POC branch is required to test some enhancement before applying
>> on trunk, it's not a problem, but the name should show that it's a POC
>> branch (why not create a branch name containing the corresponding Jira id
>> ?).
>>
>> Now, it's not easy to know if the "correct" structure is the branch 2.2.x or
>> the trunk.
>>
>> I would like to fix the trunk build and I have two ways to do it:
>> 1/ only update the assembly by deleting war deployer and manual dependencies
>> 2/ apply the new assemblies structure as we have on 2.2.x branch
>>
>> Could you light my way ?
>>
>> Thanks
>> Regards
>> JB
>>
>
>
>

Re: Trunk build broken

[Guillaume Nodet <gn...@gmail.com>]

Bad Guillaume!

I agree with you on the concept of branch to test features, but that's
not really the case here, it's just that I forgot to apply the same
changes on both trunk.
I did the changes on 2.2.x branch and planned to backport them to
trunk, but haven't gone through it, though I don't really see how/why
it's related to trunk being broken as I haven't done any modifications
in trunk.  I think the new structure as we've discussed on the mailing
list makes more sense, so I'll work on that today.


On Wed, Feb 9, 2011 at 08:14, Jean-Baptiste Onofré <jb...@nanthrax.net> wrote:
> Hi all,
>
> I raised the KARAF-451 this morning to track the fact that the trunk build
> is broken.
>
> After digging a little more, I can see that the 2.2.x branch and trunk don't
> use at all the same assembly structure:
> - the 2.2.x branch uses "new" assemblies structure (containing distribution
> and features)
> - the trunk uses "old" assembly only with distribution and another assembly
> in the features repo.
>
> I'm not very pleased with that because it means that changes on the 2.2.x
> branch have not been merged to the trunk.
>
> I think it's important that all changes are first performed on trunk and
> merge into a target branch, but not the opposite.
>
> If a kind of POC branch is required to test some enhancement before applying
> on trunk, it's not a problem, but the name should show that it's a POC
> branch (why not create a branch name containing the corresponding Jira id
> ?).
>
> Now, it's not easy to know if the "correct" structure is the branch 2.2.x or
> the trunk.
>
> I would like to fix the trunk build and I have two ways to do it:
> 1/ only update the assembly by deleting war deployer and manual dependencies
> 2/ apply the new assemblies structure as we have on 2.2.x branch
>
> Could you light my way ?
>
> Thanks
> Regards
> JB
>



-- 
Cheers,
Guillaume Nodet
------------------------
Blog: http://gnodet.blogspot.com/
------------------------
Open Source SOA
http://fusesource.com