[news.eclipse.tools.buckminster] Re: Materializing target platform from

[Thomas Hallgren <thomas@xxxxxxx>]

Hi Roland,

On 09/17/2009 10:01 AM, Roland Tepp wrote:
> 14.09.2009 13:04, Thomas Hallgren kirjutas:
> > You would execute two commands.
> >
> > buckminster importtargetdefinition my-platform.target
> > buckminster import my.mspec OR my.cquery
> Am I right to assume that these commands are available for headless
> build (I can "materialize" the target platform easily enough from the
> target definition from Eclipse UI, so my first impulse to ask how do I
> do it from Eclipse IDE is moot)
Yes, these commands (and all other commands I might add) are avaliable in headless mode. That has been one of the core 
objects for Buckminster. Everything that you can do with it in the IDE must also be available for headless execution.

> Other than that - if I want to execute some Buckminster commands from
> Eclipse UI, I can do that only from within a component?
In your IDE, you will find things that corresponds to the headless commands. The 'import' for instance, is reached using 
the standard context sensitive menu in your package explorer and then choosing Import. The wizard that pops up has a 
Buckminster entry.

Alternatively, you right click on an MSPEC or CQUERY file, choose Buckminster and then Import.

The 'importtargetdefinition' command is essentially doing the same thing as the PDE preference settings.

A 'build' is a normal workspace build.

A 'perform' is the same as right-clicking on a project, choose 'Buckminster' -> 'Invoke Action'

etc.

> > This setup will might require some redundant information though, since
> > what you require for your TP can be derived from the features that you
> > materialize.
> >
> > Regards,
> > Thomas Hallgren
>
> I would guess, that in the case of workspace materialization, this is
> quite okay, as in the development scenario I might need access to much
> wider set of target platform plug-ins to be easily able to add new
> dependencies on stuff I had before.
>
> Another use case I see this might be useful for is that when I have a
> target platform that is composed of multiple features (RCP + ECF + EMF
> for example) - then I could import the entire target platform and then
> in the second step, use the newly imported target platform as one of the
> workspace/product materialization sources...
>
> Does it make sanse? Can aI use it this way?
Yes. That makes a lot of sense. You can also slam it together to one single import step and control what ends up where 
by use of the mspec. You may have several such mspecs and if they share target platform artifacts, they will not be 
materialized more then once since the actual TP is an artifact repository managed by P2.

Regards,
Thomas Hallgren