Skip to main content


Eclipse Community Forums
Forum Search:

Search      Help    Register    Login    Home
Home » Archived » B3 » Status of b3 building tool
Status of b3 building tool [message #633587] Mon, 18 October 2010 13:06 Go to next message
Roland Tepp is currently offline Roland TeppFriend
Messages: 336
Registered: July 2009
Senior Member
Hello,

It has been relatively quiet of late and I am wondering, what is the
progress of b3 based build tooling (as opposed to b3 aggregator which
seems to be working and being used based on the latest traffic in this
newsgroup)

I understand that model and language part is relatively ready.

What is the status of the engine implementation?
There was some talk of build debugging support - any progress there?

As this seems to be a modelling subproject, maybe I am looking in the
wrong direction?

--
Roland Tepp
Re: Status of b3 building tool [message #633605 is a reply to message #633587] Mon, 18 October 2010 14:15 Go to previous messageGo to next message
Henrik Lindberg is currently offline Henrik LindbergFriend
Messages: 2509
Registered: July 2009
Senior Member
Hi Roland,
The reason for things being quiet for some time is that we have been
engaged using b3 in an application. As a result we have handled many
issues regarding packaging b3 as an RCP application, and fixing lots of
small bugs found when starting to use the b3 DSL to solve a real world
problem.

As you can see on the home page, there was a dip in commits in
september, but other than that there is quite a lot of activity:
http://www.eclipse.org/modeling/emft/b3/

Many issues in the area of type inference, code completion, and
validation have been found and corrected over the last couple of weeks.

The b3 model and language are quite stable, there are a few outstanding
issues that I expect will have minor impact on model/grammar.

Regarding build debugging - we have not yet started on this. The initial
idea was to try to reuse the JDT debugger, and let b3 generate
bytecodes. In practice this turned out to be more problematic than first
anticipated as it is too hard to present meaningful context information.
Also, when debugging a build there are other questions that you as
developer would like to have answered than the typical "what is the
value of this variable", and "how did I get here" - I am thinking about
queries regarding components, queries to repositories, debugging the
resolution process etc. So instead of the JDT debugger, the general
debugging support looks promising and should not be very difficult to
implement. (Help in this area is much appreciated).

For building, we still recommend using Buckminster as it is production
quality.

Regards
- henrik

On 10/18/10 3:06 PM, Roland Tepp wrote:
> Hello,
>
> It has been relatively quiet of late and I am wondering, what is the
> progress of b3 based build tooling (as opposed to b3 aggregator which
> seems to be working and being used based on the latest traffic in this
> newsgroup)
>
> I understand that model and language part is relatively ready.
>
> What is the status of the engine implementation?
> There was some talk of build debugging support - any progress there?
>
> As this seems to be a modelling subproject, maybe I am looking in the
> wrong direction?
>
Re: Status of b3 building tool [message #633862 is a reply to message #633605] Tue, 19 October 2010 16:01 Go to previous messageGo to next message
Roland Tepp is currently offline Roland TeppFriend
Messages: 336
Registered: July 2009
Senior Member
18.10.2010 17:15, Henrik Lindberg kirjutas:
> Hi Roland,
> The reason for things being quiet for some time is that we have been
> engaged using b3 in an application. As a result we have handled many
> issues regarding packaging b3 as an RCP application, and fixing lots of
> small bugs found when starting to use the b3 DSL to solve a real world
> problem.
>
Nice to hear, that b3 solves a real problems in a real world as well as
theoretical ones :)

Any chance of disclosing a name of the product/application?



> Regarding build debugging - we have not yet started on this. The initial
>
I did not really have too high hopes for that yet, but it is still good
to hear that the issue is being ... discussed :)

As an idea on the debugging ... maybe also as an alternative to run time
debugging an offline debugging of the already executed build process
would also make sense -- if nothing else then simply to detect where
something went wrong and trace back the steps to find the root of the
problem...

> For building, we still recommend using Buckminster as it is production
> quality.
>
Okay. So do I understand it correctly that I could use b3 as a
model/language for the build and use Buckminster as an engine running
the b3 build?

Or are you recommending to use Buckminster all the way?

--
Roland Tepp
Re: Status of b3 building tool [message #633948 is a reply to message #633862] Wed, 20 October 2010 00:55 Go to previous messageGo to next message
Henrik Lindberg is currently offline Henrik LindbergFriend
Messages: 2509
Registered: July 2009
Senior Member
On 10/19/10 6:01 PM, Roland Tepp wrote:
> 18.10.2010 17:15, Henrik Lindberg kirjutas:
> Nice to hear, that b3 solves a real problems in a real world as well as
> theoretical ones :)
>
> Any chance of disclosing a name of the product/application?
>
Will probably be disclosed around ESE (November 2-4)

> As an idea on the debugging ... maybe also as an alternative to run time
> debugging an offline debugging of the already executed build process
> would also make sense -- if nothing else then simply to detect where
> something went wrong and trace back the steps to find the root of the
> problem...
>
That was one of the reasons I initially wanted to use the JDT debugger
as it comes with remote debugging support out of the box. It is not
enough to debug the build when running locally as you do not find the
problems caused by environmental differences.

So, remote debugging is very much wanted.

>> For building, we still recommend using Buckminster as it is production
>> quality.
>>
> Okay. So do I understand it correctly that I could use b3 as a
> model/language for the build and use Buckminster as an engine running
> the b3 build?
>
That is the idea, but not ready.

> Or are you recommending to use Buckminster all the way?
>
For now yes.

Regards
- henrik
Re: Status of b3 building tool [message #634376 is a reply to message #633948] Thu, 21 October 2010 12:38 Go to previous messageGo to next message
Roland Tepp is currently offline Roland TeppFriend
Messages: 336
Registered: July 2009
Senior Member
20.10.2010 3:55, Henrik Lindberg kirjutas:
>> Any chance of disclosing a name of the product/application?
>>
> Will probably be disclosed around ESE (November 2-4)

I'll be looking for that then :)


20.10.2010 3:55, Henrik Lindberg kirjutas:
>> Okay. So do I understand it correctly that I could use b3 as a
>> model/language for the build and use Buckminster as an engine running
>> the b3 build?
>>
> That is the idea, but not ready.
>
>> Or are you recommending to use Buckminster all the way?
>>
> For now yes.

Okay. I can accept that :)

Another question.
I realize I should most likely ask this from PDE guys, but have there
been any plans or talks of moving the whole PDE build on top of b3 model?

--
Roland Tepp
Re: Status of b3 building tool [message #634430 is a reply to message #634376] Thu, 21 October 2010 16:31 Go to previous messageGo to next message
Henrik Lindberg is currently offline Henrik LindbergFriend
Messages: 2509
Registered: July 2009
Senior Member
On 10/21/10 2:38 PM, Roland Tepp wrote:
> Another question.
> I realize I should most likely ask this from PDE guys, but have there
> been any plans or talks of moving the whole PDE build on top of b3 model?
>
That was how the ideas around b3 started. PDE/Build consists of two
parts - the actual building, and the headless generated ANT scripts. The
part of PDE/Build that many do not like is the generated ANT script. The
idea for b3 is to use the same approach as in Buckminster - i.e. to call
the underlying PDE logic that performs the actual building.
That underlying logic will probably never be replaced, and the PDE/Build
(the headless part) will in any scenario be around for quite some time
as many depend on it.

An additional idea for b3 is to not be tied to one particular way of
building and allowing different parts of a system to be built by
different build technologies.

Regards
- henrik
Re: Status of b3 building tool [message #638419 is a reply to message #634430] Thu, 11 November 2010 09:23 Go to previous message
Roland Tepp is currently offline Roland TeppFriend
Messages: 336
Registered: July 2009
Senior Member
21.10.2010 19:31, Henrik Lindberg kirjutas:
> That was how the ideas around b3 started. PDE/Build consists of two
> parts - the actual building, and the headless generated ANT scripts. The
> part of PDE/Build that many do not like is the generated ANT script. The
> idea for b3 is to use the same approach as in Buckminster - i.e. to call
> the underlying PDE logic that performs the actual building.
> That underlying logic will probably never be replaced, and the PDE/Build
> (the headless part) will in any scenario be around for quite some time
> as many depend on it.

So, I assume, the short answer is "No" there are currently no plans to
replace headless (generated ANT scripts) way of building with b3?

--
Roland Tepp
Previous Topic:missing requirement while aggregation the polarion update site whith b3 aggregator
Next Topic:Cannot install b3 Aggregator into 3.7 M3.
Goto Forum:
  


Current Time: Tue Mar 19 08:24:56 GMT 2024

Powered by FUDForum. Page generated in 0.03698 seconds
.:: Contact :: Home ::.

Powered by: FUDforum 3.0.2.
Copyright ©2001-2010 FUDforum Bulletin Board Software

Back to the top