Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [cdt-dev] Re: Interaction between perspectives.

Mikhail,

Ok, thanks!

Vitaly


On 10/25/06, Mikhail Khodjaiants <Mikhail.Khodjaiants@xxxxxxx> wrote:
Vitaly,

This is the tutorial presented at EclipseCon 2006 -
http://www.eclipse.org/eclipse/debug/documents/debugger_Advanced-Debug-T
utorial.pdf. The API hasn't been "officially" released, all packages
include "provisional" in their names, but it is available. For example,
the implementation of the Modules view of CDT is based on the
provisional API. There have been changes since the document was
released.

DSF is the part of DSDP project. The intention is to provide the backend
debug model with focus on the slow and remote targets.

Mikhail

-----Original Message-----
From: cdt-dev-bounces@xxxxxxxxxxx [mailto:cdt-dev-bounces@xxxxxxxxxxx]
On Behalf Of Vitaly Provodin
Sent: 25 October 2006 04:34
To: CDT General developers list.
Subject: Re: [cdt-dev] Re: Interaction between perspectives.

Mikhail,

What do you mean under _the new Debug Platform approach_?
Is it the Debugger Services Framework (DSF)
[http://dsdp.eclipse.org/help/latest/index.jsp?topic=/org.eclipse.dd.dsf
.doc/docs/dsf_white_paper.html]?

=Vitaly

On 10/24/06, Mikhail Khodjaiants <Mikhail.Khodjaiants@xxxxxxx> wrote:
> Vitaly,
>
> If your implementation communicates to the views directly, without JDT

> or CDT, you can use the new Debug Platform approach to customize the
> presentation and structure of your elements in the existing debugger
> views. There is also a proposal to allow multiple views, i.e. clone
> and pin the views, that may be useful, but I don't know the current
> status of it.
>
> Thanks,
> Mikhail
>
> -----Original Message-----
> From: cdt-dev-bounces@xxxxxxxxxxx [mailto:cdt-dev-bounces@xxxxxxxxxxx]
> On Behalf Of Vitaly Provodin
> Sent: 24 October 2006 10:37
> To: CDT General developers list.
> Subject: Re: [cdt-dev] Re: Interaction between perspectives.
>
> Mikhail,
>
> > > Do you display the stack frames for Java and C/C++ code under the
> > > same thread?
>
> Yes. There is only one debug target. So in the Debug view both kind of
> threads: Java and native are displayed under this target at the same
> time in the same view. Under the Java threads both kinds of frames:
> Java and native are displayed as well.
>
> > > And is there a need to introduce additional perspectives?
>
> An example - JNI-frame which can be displayed as either a Java frame
> or a native one. Or even another kind of frames may be introduced -
> _mixed_frame_.
> Currently, such frames are displayed as native ones. But there may be
> needs to switch between these various representations of such debug
> elements.
> To resolve this there were several suggestions: to add a new
> perspective displaying possibble representations; to add new views; to

> introduce some actions (switching representations) for existing views.
> IMHO the last one is more appropriate.
> At this moment I do not see the needs to introduce neither new
> perspectives nor new views and completely agree with Doug that minimal

> number of perspectives makes user's life easier. So we are filling
> existing view of the perspective.
>
> > > By the way, as far as I know you are using the Harmony JVM as a
> > > backend debugger for Java, what is your backend for C/C++?
>
> Both for Java and C/C++ we are using the same backend - that is
> extendable JVMTI agent which may work with any VM. But _any_VM_
> provides no native debug-info. In this case the agent ignores native
> requests and works as an ordinary JVMTI-agent. There are some add-ons
> to the Harmony VM supporting native debugging which are used by this
> agent. I'm not sure, whether these add-ons have been already
> contributed into Harmony, but I hope it will be.
>
> Thanks,
> =Vitaly
>
>
> On 10/23/06, Doug Schaefer <DSchaefer@xxxxxxx> wrote:
> > On that note, I'm currently working on the Windows debugger
> > integration with dbgeng.dll using JNI. I hate the fact I have to
> > switch between Java and CDT perspectives to get certain behaviour
> > (in
> particular the New Class Wizard).
> > Whatever we can do to keep the number of perspectives to a minimum
> > will certainly help the user from having to switch all the time.
> >
> > Doug Schaefer
> > QNX Software Systems
> > Eclipse CDT Project Lead
> > http://cdtdoug.blogspot.com
> >
> >
> > > -----Original Message-----
> > > From: cdt-dev-bounces@xxxxxxxxxxx
> > > [mailto:cdt-dev-bounces@xxxxxxxxxxx] On Behalf Of Mikhail
> > > Khodjaiants
> > > Sent: Monday, October 23, 2006 7:33 AM
> > > To: CDT General developers list.
> > > Subject: RE: [cdt-dev] Re: Interaction between perspectives.
> > >
> > > Vitaly,
> > >
> > > Do you display the stack frames for Java and C/C++ code under the
> > > same thread? And is there a need to introduce additional
> > > perspectives? The idea of the Debug perspective in Eclipse is to
> > > host all debug sessions regardless of the type and language.
> > > By the way, as far as I know you are using the Harmony JVM as a
> > > backend debugger for Java, what is your backend for C/C++?
> > >
> > > Thanks,
> > > Mikhail Khodjaiants
> > > ARM Ltd.
> > > -----Original Message-----
> > > From: cdt-dev-bounces@xxxxxxxxxxx
> > > [mailto:cdt-dev-bounces@xxxxxxxxxxx]
> > > On Behalf Of Vitaly Provodin
> > > Sent: 23 October 2006 10:28
> > > To: CDT General developers list.
> > > Subject: Re: [cdt-dev] Re: Interaction between perspectives.
> > >
> > > Ramesh,
> > >
> > > We are working on the mixed Java/JNI debugger supporting currently

> > > Java and C/C++ languages. It is being integrated into Eclipse IDE
> > > under the Debug perspective. Currently, we are able to set and
> > > stop on line breakpoints in both Java and C/C++ code. Depending on

> > > what stack frame is active, we are able to watch Java and native
> > > data - variables, their values. It works in the ordinary debug
perspective.
> > > Probably, (two or more) separate percpectives for each debug-info
> > > representation might be introduced, but at this moment there are
> > > no well-defined vision of this.
> > >
> > > Vitaly Provodin,
> > > Intel Middleware Products Division
> > >
> > >
> > > On 10/18/06, Martin Oberhuber <martin.oberhuber@xxxxxxxxxxxxx>
> wrote:
> > > > Ramesh -
> > > >
> > > > many thanks for your request. There are some initiatives in the
> > > > area of debugger inter-communications, but none is really mature
> yet.
> > > >
> > > > One problem is that when an event in debugger A should initiate
> > > > some action in debugger B, you'd typically want to do this with
> > > > very low latency. Which requires a solution sitting "very close
> > > > to
>
> > > > the target",
> > >
> > > > which typically means vendor-specific non-standard (read:
> > > > non-eclipse)
> > >
> > > > solutions.
> > > >
> > > > If low latency this is not a requirement, some Eclipse UI based
> > > > event handling could be the right solution. We thought that the
> > > > Target Management Project might be the Umbrella defining such
> > > > events, to be understood by many debuggers that want to
> > > > communicate. An empty discussion page is available under
> "Inter-Debugger Communications"
> > > > from http://wiki.eclipse.org/index.php/DSDP/TM
> > > >
> > > > I'm not sure how far the Device Debugging Project has already
> > > > progressed to support inter-debugger communications.
> > > >
> > > > Another initiative I know of, is a Java/JNI/gdb debugger
> > > > integration being investigated by intel. You might be able to
> > > > meet
>
> > > > these folks on the CDT groups.
> > > >
> > > > I'm copying the dsdp-tm-dev, dsdp-dd-dev and cdt-dev mailing
> > > > lists
>
> > > > with this answer. If there is any developer that can help you
> > > > further,
> > >
> > > > please chime in.
> > > >
> > > > Thanks
> > > > Martin
> > > >
> > > > Ramesh Ramaiyer schrieb:
> > > > > Hi all,
> > > > > We are working on a SoC with three distinct cores. Each of
> > > > > these
>
> > > > > cores have their own debuggers. Very soon they live under
> Eclipse.
> > > > > The expectation from developers is that they should be able to

> > > > > debug
> > >
> > > > > the interactions between these cores as well.
> > > > >
> > > > > This implies that interaction between Perspectives should be
> > > possible.
> > > > > Under the current implementation each debugger has its own
> > > perspective.
> > > > >  An example usecase might be - when a break point is hit on
> > > > > one core, the other two stopping (ASAP). Another might be the
> > > > > ability to
> > >
> > > > > place breakpoints on programs running on any of the cores.
> > > > >
> > > > > In my mind these usecases are similar to, some one stepping
> > > > > through a piece of java code which makes a JNI call. If the
> > > > > JNI is written in C++, is it possible to automatically invoke
> > > > > C++ perspective and
> > > switch to that?
> > > > >
> > > > > I don't know if such a support is already present. Can you
> > > > > point
>
> > > > > me in the right direction please?
> > > > >
> > > > > Any help is greatly appreciated.
> > > > > Thanks and regards
> > > > > Ramesh
> > > > > Note: I have also placed this question on Eclipse Newcomer
> group.
> > > >
> > > >
> > > > --
> > > > Martin Oberhuber
> > > > Wind River Systems, Inc.
> > > > Target Management Project Lead, DSDP PMC Member
> > > > http://www.eclipse.org/dsdp/tm
> > > > _______________________________________________
> > > > cdt-dev mailing list
> > > > cdt-dev@xxxxxxxxxxx
> > > > https://dev.eclipse.org/mailman/listinfo/cdt-dev
> > > >
> > > _______________________________________________
> > > cdt-dev mailing list
> > > cdt-dev@xxxxxxxxxxx
> > > https://dev.eclipse.org/mailman/listinfo/cdt-dev
> > >
> > > --
> > > IMPORTANT NOTICE: The contents of this email and any attachments
> > > are
>
> > > confidential and may also be privileged. If you are not the
> > > intended
>
> > > recipient, please notify the sender immediately and do not
> > > disclose the contents to any other person, use it for any purpose,

> > > or store or copy the information in any medium.  Thank you.
> > >
> > >
> > > _______________________________________________
> > > cdt-dev mailing list
> > > cdt-dev@xxxxxxxxxxx
> > > https://dev.eclipse.org/mailman/listinfo/cdt-dev
> > _______________________________________________
> > cdt-dev mailing list
> > cdt-dev@xxxxxxxxxxx
> > https://dev.eclipse.org/mailman/listinfo/cdt-dev
> >
>
>
> --
> Vitaly Provodin,
> Intel Middleware Products Division
> _______________________________________________
> cdt-dev mailing list
> cdt-dev@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/cdt-dev
> _______________________________________________
> cdt-dev mailing list
> cdt-dev@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/cdt-dev
>
_______________________________________________
cdt-dev mailing list
cdt-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/cdt-dev
_______________________________________________
cdt-dev mailing list
cdt-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/cdt-dev



--
Vitaly Provodin,
Intel Middleware Products Division


Back to the top