1. Why is e4 still using an (empty) Activator in almost every project?
Shouldn't it be a good OSGi citizen and use the
FramworkUtil.getBundle(Class) whenever the current Bundle/
BundleContext
is needed? Also the Activator.ID is error prone to refactoring, thats
why I never use it in my own projects.
A similar problem has the NLS feature of Eclipse PDE, when it
creates the Messages class and puts the BUNDLE into a string rather
to have
NLS.initialize(Messages.class.getPackage().getName() + ".messages",
Messages.class)
Yes, setting up good OSGi programming practices from the beginning
would be great. IMHO this actually means NOT using OSGi types
wherever possible. Through DS and other means we should be able to
avoid referencing BundeContexts in normal Eclipse code yet still use
OSGi services to structure things.
The use of string constants for getting bundle ids does have the
brittleness mentioned and IMHO should be avoided. The frameworkUtil
method you mention is a fine way to do this. Not that if you really
want to use constants (for this and other purposes) there is no need
to use an Activator for that. BundleActivators are OSGi specific and
are there for lifecycle. Historically the Plugin class was the place
where we dumped all manner of stuff but best practice would have your
random domain constants etc wherever you would normally put them if
not using OSGi and just putting the OSGi-related code (what? why?) in
OSGi-related types.
Jeff
_______________________________________________
e4-dev mailing list
[email protected]
https://dev.eclipse.org/mailman/listinfo/e4-dev
