I'm afraid we will need to remove circular dependencies our of our code :(
Without it the build if failed :(
I'll make the structure consistent and will commit it


On Fri, Apr 4, 2014 at 4:10 PM, seba.wag...@gmail.com <seba.wag...@gmail.com
> wrote:

> It is quite complex. For instance the openlaszlo compilation is not a
> widespread maven target. I think we also have a bit a learning curve to get
> comfortable with the new options.
> I think a migration phase is quite acceptable for everybody.
>
> Sebastian
> On Apr 4, 2014 10:04 PM, "Maxim Solodovnik" <solomax...@gmail.com> wrote:
>
>> OK
>> Thanks!
>>
>> I'll checking working parts and things will live "in parallel" for some
>> time
>>
>>
>> On Fri, Apr 4, 2014 at 4:02 PM, seba.wag...@gmail.com <
>> seba.wag...@gmail.com> wrote:
>>
>>> We would like to go away from a highly customized build process to a
>>> standardize process to build, develop, release and test software.
>>>
>>> While Ant is tool which is good to compile software we are already need
>>> to extend it with Ivy to get dependency management. While Ivy is also using
>>> Maven repositories in the end. Maven will integrate that in one tool
>>> without hacking our own way.
>>>
>>> Maven further has standardized hooks for instance to release software.
>>> For example this entire naming convention "-SNAPSHOT" that we simulate
>>> manually, actually its root is from Maven, where this is just the way Maven
>>> calls the packages. And Maven supplies a build target to create a release,
>>> commit the tag to the SVN and update the main branch to with the version
>>> name in one go.
>>> Same for Testing, we simulate Maven functionality while Maven has a
>>> build in target that would nicely integrate with Jenkins to generate our
>>> test reports. Building the test suite is part of the Maven release and
>>> build process. So every build will automatically include the regression
>>> test. Instead of us manually somehow hook some hand coded ant builds in
>>> some order, Maven would do that.
>>> In other words: It is not just a tool to compile something, it is
>>> framework for the entire software development life cycle , build, develop,
>>> test and release.
>>>
>>> Besides that it makes it easier for us to build components that are less
>>> coupled and can life on their own. We can build multiple Maven plugins.
>>> While we do not see use in some of our JARs and classes, others might be.
>>> The more accessible we make our project the easier it will be for 3rd party
>>> to hook into our application and contribute something.
>>>
>>> Our ANT script is pretty much an organic grown monster. I started it
>>> with 10 lines of build script. Now it is thousands. For anybody beyond you
>>> and me this is pretty much un-maintainable. And with Maven we do not only
>>> solve that but also can get rid of some of the custom build script as it is
>>> already build into Maven.
>>>
>>> So from my point of view that would be a very desirable goal to migrate
>>> to Maven. It is probably not as straightforward as switching a couple of
>>> flags, but the longer we wait the more difficult it will be to maintain
>>> what we have.
>>>
>>> Sebastian
>>> On Apr 4, 2014 9:26 PM, "Maxim Solodovnik" <solomax...@gmail.com> wrote:
>>>
>>>> Hello Sebastian,
>>>>
>>>> I have started to migrate our build system to maven.
>>>> Could you please remind me why do we need such migration?
>>>>
>>>> --
>>>> WBR
>>>> Maxim aka solomax
>>>>
>>>
>>
>>
>> --
>> WBR
>> Maxim aka solomax
>>
>


-- 
WBR
Maxim aka solomax

Reply via email to