[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
Re: [gmf-dev] GMF runtime complexity makes it hard for clients to benefit from its power
|
Hi Mindaugas,
what you suggest here with your "activation explanation" is pretty much
what I had in mind.
@GMF-runtime staff: is something like that already existing or is it
feasible to implement it?
Thanks,
Henrik
Mindaugas Idzelis wrote:
>
> I also had to go through a long ramp-up period to understand all of
> GMF. The best way to learn, IMHO is through trial and error,
> persistence, and heavy use of the debugger :-) Of course, that's the
> way I've learned most of swt, gef, platform, jdt, etc. I don't think
> this ramp-up period is unique to GMF, but to all large software bases.
>
> One thing that I think will be extremely valuable is to have some kind
> of "activation explaination" built into the Service infrastructure. It
> would be great if there were debug switches available that for a given
> service will trace every provider that is consulted and why or not it
> was chosen. There are many layers that services use... execution
> strategies, enablement, policies, and it is pretty daunting at first
> to debug. Luckily, I've figured out strategic places to put
> breakpoints ;-) Nevertheless, some debug support for Service
> activation would go a long way in understanding the various types of
> extensibility.
>
>
> Thanks,
>
> Mindaugas Idzelis
> IBM Rational Middleware
>
>
> *Steven Shaw <steveshaw@xxxxxxxxxx>*
> Sent by: gmf-dev-bounces@xxxxxxxxxxx
>
> 11/08/2005 09:56 AM
> Please respond to
> "GMF Project developer discussions."
>
>
>
> To
> "GMF Project developer discussions." <gmf-dev@xxxxxxxxxxx>
> cc
>
> Subject
> RE: [gmf-dev] GMF runtime complexity makes it hard for clients
> to benefit from its power
>
>
>
>
>
>
>
>
>
> Note: In case people weren't aware, there are a number of tutorials
> currently available in the SDK and a "How-to" document that are fairly
> comprehensive.
>
> See org.eclipse.gmf.runtime.doc.isv / html / howto
> /
> tutorials
>
> Regards,
>
> Steve.
>
> /Tuesday, November 08, 2005 9:45 AM
> To: "GMF Project developer discussions." <gmf-dev@xxxxxxxxxxx>
> cc:
> From: Doug Schaefer <DSchaefer@xxxxxxx>
> Subject: RE: [gmf-dev] GMF runtime complexity makes it hard for
> clients to benefit from its power/
>
> Thanks Steve, I think a document would be immensely helpful for people
> trying to get their feet wet. The barrier to entry is pretty high and
> I keep bouncing off. J.
>
> Cheers,
> Doug
>
>
> ------------------------------------------------------------------------
>
> *From:* gmf-dev-bounces@xxxxxxxxxxx
> [mailto:gmf-dev-bounces@xxxxxxxxxxx] *On Behalf Of *Steven Shaw*
> Sent:* Tuesday, November 08, 2005 9:13 AM*
> To:* GMF Project developer discussions.*
> Cc:* GMF Project developer discussions.; gmf-dev-bounces@xxxxxxxxxxx*
> Subject:* Re: [gmf-dev] GMF runtime complexity makes it hard for
> clients to benefit from its power
>
>
> Hello Henrik,
>
> Thanks for your feedback. The complexity certainly can be
> overwhelming due to the power of the infrastructure for extensibility
> and as a platform across domains. There is a programmers guide for
> the diagram plugins under construction that hasn't been published yet
> in the SDK. This would include some basic interaction diagrams
> describing how the creation process invokes the different services
> (View, EditPart, EditPolicy etc.). This should be available in M4.
>
> There is also some planned investigation on simplifying the service
> schema to avoid some of the complex XML tags that are needed to define
> providers. These are currently necessary to avoid loading a plugin
> simply to determine enablement or provides criteria for a provider.
> With the introduction of the OSGI bundles some of this complexity
> could be potentially eliminated allowing the logic to be centralized
> in the java source. (This needs to be investigated further...)
>
> Re: log --> There is advanced tracing option available that can be
> activated that should let you see trace data logged by services. In
> the Debug dialog, if you navigate to the "Tracing" tab and turn on
> "Enable tracing for the selected plug-ins".
>
> Regards,
>
> Steve.
> ________________________________________
> Steven R. Shaw
> Rational Modeling Platform Diagram Layer Lead
> Rational Software | IBM Software Group
> 770 Palladium Drive, Kanata, ON, Canada, K2V 1C8
> tel: 613.591.7979
> steveshaw@xxxxxxxxxx
>
> *Henrik Rentz-Reichert <hrr@xxxxxxxxx>*
> Sent by: gmf-dev-bounces@xxxxxxxxxxx
>
> 11/08/2005 04:46 AM
>
>
> Please respond to
> "GMF Project developer discussions."
>
>
>
> To
> "GMF Project developer discussions." <gmf-dev@xxxxxxxxxxx>
> cc
>
> Subject
> [gmf-dev] GMF runtime complexity makes it hard for clients to benefit
> from its power
>
>
>
>
>
>
>
>
>
>
>
>
> Hi all,
>
> currently I'm implementing a simple example using the GMF runtime.
>
> I tried to follow the lines of the logic example (which is fairly
> complex though). Obviously the framework is extremely powerful. But
> its complexity also puts a high burden on its clients. My experience
> up to now is that it is very difficult to trace problems back to their
> origin. This is partly due to the complex interactions of the
> framework with the providers defined by the client. Especially
> registering the clients via extension points is very error prone and
> problems are hard to detect.
>
> Some examples:
>
> * it took me quite a long time to realize that the DiagramKind of
> the Wizard page has to be in sync with the corresponding entry
> in the ViewProvider
> * I missed to derive an EditHelper from MSLEditHelper which led to
> an exception in a context that was extremely hard to analyze
> * in general declarations in the plugin.xml interact via services
> and service providers with user written code. There are many
> places where strings here and there have to coincide literally
> which can lead to subtle errors which are typically very hard to
> detect for a GMF runtime novice
>
> Now my questions:
>
> * Is there a document (sequence diagrams would be great) that
> illustrates the interaction of framework instances with user
> defined code?
> * Is a log file written with registry information of the GMF
> runtime services? Or alternatively: can I get information of
> this kind using certain tracing switches? This kind of
> information could be extremely helpful
>
>
> Kind regards,
>
> Henrik
>
> --
> ..................................................
>
> Dr. Henrik Rentz-Reichert _mailto://hrr@protos.de_
> <mailto://hrr@xxxxxxxxx/>
>
> Hafnerstr. 1 fon +49-7533-9342-43
> D-78476 Allensbach fax -44
> _______________________________________________
> gmf-dev mailing list
> gmf-dev@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/gmf-dev_______________________________________________
> gmf-dev mailing list
> gmf-dev@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/gmf-dev
>
>------------------------------------------------------------------------
>
>_______________________________________________
>gmf-dev mailing list
>gmf-dev@xxxxxxxxxxx
>https://dev.eclipse.org/mailman/listinfo/gmf-dev
>
>
--
..................................................
Dr. Henrik Rentz-Reichert mailto://hrr@xxxxxxxxx
Hafnerstr. 1 fon +49-7533-9342-43
D-78476 Allensbach fax -44