[news.eclipse.tools.buckminster] Re: Best way to include launchers in an

["Tas Frangoullides" <tfrangoullides@xxxxxxxxxxxxxxxxxx>]

Hi Thomas,

I've been looking for a suitable workaround and failed. Instead I've 
uncovered more differences between buckminster and PDE. The most difficult 
is that products installed from sites created using buckminster don't run. 
The following error is found in the log.

!ENTRY org.eclipse.osgi 4 0 2009-08-12 16:46:39.875
!MESSAGE Application error
!STACK 1
java.lang.IllegalStateException: Unable to acquire application service. 
Ensure that the org.eclipse.core.runtime bundle is resolved and started (see 
config.ini).
at 
org.eclipse.core.runtime.internal.adaptor.EclipseAppLauncher.start(EclipseAppLauncher.java:74)
at 
org.eclipse.core.runtime.adaptor.EclipseStarter.run(EclipseStarter.java:368)
at 
org.eclipse.core.runtime.adaptor.EclipseStarter.run(EclipseStarter.java:179)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(Unknown Source)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(Unknown Source)
at java.lang.reflect.Method.invoke(Unknown Source)
at org.eclipse.equinox.launcher.Main.invokeFramework(Main.java:559)
at org.eclipse.equinox.launcher.Main.basicRun(Main.java:514)
at org.eclipse.equinox.launcher.Main.run(Main.java:1311)

If I add the application plug-in to the configuration tab of the product 
configuration the product works. What's happening here? Is PDE adding some 
configuration by default?

Also, looking at the issue of having to include the executables feature, it 
has another unpleasant side effect of making the feature available in the 
resulting p2 repository which I don't want.

I really want to have a buckminster build which can produce similar results 
from the same source currently processed by a PDE build. I think many people 
migrating from PDE Build to buckminster will want this.

To workaround these issues I am having to consider using buckminster only 
for materializing the target platform and fetching the source and then 
handing things over to PDE Build for to create the p2 repository. However I 
would lose the revision based qualifier substitution which I find very 
useful. Any suggestions as to how I can work around would be greatly 
appreciate. Maybe it's easier to help resolve these problem in buckminster?

Thanks,
Tas


"Tas Frangoullides" <tfrangoullides@xxxxxxxxxxxxxxxxxx> wrote in message 
news:h5rie3$p6f$1@xxxxxxxxxxxxxxxxxxxx
> Hi Thomas,
>
> > > Following the mailapp tutorial I've noticed that the 
> > > org.eclipse.equinox.executable needs to be included in the product 
> > > feature so that launchers are included when the product is created. You 
> > > don't need to do this when using the PDE build
> >
> > That's actually a modified truth. If your objective is to create a 
> > repository from with the application can be installed onto more then one 
> > platform, then you must have the executable feature. If you don't, you 
> > can only create the application for the same platform as the one you're 
> > currently building from.
>
> This is what I really like about the approach supported by buckminster of 
> creating a multi-platform site first. Is there a way to have the feature 
> included in the site but not the installed product. I tried creating a 
> site feature and moved the executables feature into that but this didn't 
> work.
>
> > > and you actually get a different product in terms of installed features. 
> > > Also there is the bug relating to icons. Is there a way to arrange the 
> > > features or build so that both PDE and Buckminster produce identical 
> > > results?
> > Not at present. There's a bug in how the launcher is branded and my guess 
> > (unverified) is that the PDE build somehow cheats at the end to work 
> > around that bug. I haven't found any way to create the correct meta-data 
> > just yet. At present, you'll get two launchers every time. We're working 
> > on that.
>
> Fair enough. I'll just find ways to work around this in my build, maybe 
> looking into how PDE gets round them is a good starting point.
>
> Thanks,
> Tas