>>> Then, if we could
find someone that has
made a lot of extensions to the existing CDT model who could give a
presentation on it
I can give a
presentation that shares our experience
at TI for the past two years.
We started
implementing CDI debugger and
over time accumulated multiple changes in different layers (CDT,
Platform) to satisfy
our requirements.
Do you think
that will have value for the rest
of the community?
May be at
the end we can converge into
some kind of table that can help others decide which layer is most
appropriate for
their debugger integration.
- If your
debugger has these and these requirements
your best bet is to integrate against that layer.
To be
honest, I kind of expect DSF to have
this kind of guideline document.
- if your
debugger has these and these features
you can integrate against these and these services in DSF.
Regards
Dobrin
Hi Mikhail and all,
I agree with Doug, the CDT summit seems like the more appropriate time
to
discuss this topic. I would be glad to give an overview of what is
involved in creating a non-MI based debugger integration using DSF and
I think
that there is a wider audience for this topic than the folks that would
like to
spend a day in a coding camp. I also would also love to hear about the
approach that you (Mikhail) took in implementing a custom model and the
flexible hierarchy adapters for it. Then, if we could find someone
that
has made a lot of extensions to the existing CDT model who could give a
presentation on it, then we could have a balanced in-depth discussion
on the
three approaches that everyone would benefit from :-)
Cheers,
Pawel
Mikhail Khodjaiants wrote:
Doug,
We have
compared CDI and DSF many times. I
believe those who implemented a CDI based integration understand its
limitations. The new flexible hierarchy API has been added lately to
the
Platform. DSF is based on this API and takes advantage of it. CDI can
be ported
to it but it is not a simple task especially if we want to keep it
compatible with the existing implementations.
I am
currently integrating our debug
kernel into Eclipse and using the flexible hierarchy API directly. So
far
it is going quite well and I haven't had serious problems. But I
would like to understand what would I gain if I use DSF.
I don't mind
to discuss it at the CDT
summit, but it is not related directly to CDT.
From: dsdp-dd-dev-bounces@xxxxxxxxxxx
[mailto:dsdp-dd-dev-bounces@xxxxxxxxxxx]
On Behalf Of Doug
Schaefer
Sent: 22 August 2007
15:43
To: Device Debugging
developer
discussions
Subject: RE:
[dsdp-dd-dev] DSF
Coding Camp: September 24th in Ottawa
This is
definitely a discussion I’d
like to have at the CDT summit. Pawel can correct me, but I believe the
code
camp was for people currently working on DSF integrations. During our
Debug day
at the summit, we can do a Q&A and compare CDI with DSF and build
up a
checklist to help integrators decide which to use.
Hi
Pawel,
I
also would like to attend.
I think
Mikhail expressed exactly our main
concern at TI.
And I
believe it is a concern for other
people that have a non GDB version of their debuggers.
Currently
we integrated our
debugger through CDI.
Going
forward we’d like to know
which of the options makes most sense considering out debugger
features.
- stay with CDI
- go to the Platform Flexible Hierarchy.
- use all or most of DSF
- use
parts of DSF that fit most.
I
know that it is not a “yes or no answer”, but information that will
help us make the optimal decision will be most useful for us and I
believe
others that struggle with the same question.
Regards
Dobrin
Hi
Pawel,
What
kind of a DSF-related
discussion is planned at the CDT meeting?
Personally I
would like to evaluate
the pros and contras of using DSF instead of implementing the
new flexible-hierarchy API directly for non-gdb based backends. Also it
would be very useful for developers if you present a list of high level
tasks that are necessary to integrate a backend debugger into DSF.
None of
these topics is directly related
to CDT. At the same time people who write code for the gdb/mi
implementation
may not be interested. What do you think, which of the two meetings is
more
appropriate to discuss this type of questions?
From: dsdp-dd-dev-bounces@xxxxxxxxxxx
[mailto:dsdp-dd-dev-bounces@xxxxxxxxxxx]
On Behalf Of Pawel
Piech
Sent: 22 August 2007
00:37
To: Device
Debugging developer discussions
Subject: [dsdp-dd-dev]
DSF Coding
Camp: September 24th in Ottawa
Hi All,
Thanks Dominique for the reminder. To clean up the list of meetings on
DD
Wiki, I moved the long list of GDB/MI related meetings to a new GDB sub-group page.
But I
also just added a link to the upcoming
face to face meeting on the DD
Hope page.
I haven't put anything in the agenda section because I think it depends
on who
will attend. If we have a lot of folks who would like to get a first
on-hands experience with DSF we should concentrate on small-scale
exercises. If it's mostly contributors, we should concentrate on
upcoming
design issues.
The meeting is open to anyone, please add your name to the
"Attendees" list if you would like to participate. And if you
have specific questions about DSF that you would like to have answered
please
add them to the Agenda as well. I hope that by the September 6th
general
meeting we should be able to get a reliable head count and have a clear
idea in
which direction to take the coding camp.
Cheers
Pawel
Dominique Toupin (QA/EMC) wrote:
Could
you put the info (location, attendees, prerequisites, etc.) about the
face to
face meeting of Sept 24 on the WIKI and add a link on the main DD WIKI (http://wiki.eclipse.org/DSDP/DD).
It could
also be nice to
add this link on the CDT Summit page and the CDT Summit link on
the Sept 24 meeting page.
I had
people who are
not able to find info, the link for the Sept 24 meeting doesn't seem to
be
on any page.
If you send
the link to
DD dev list (maybe CDT mailing list) it would be great.
Sorry
to bother you with this I know you don't like the administrative part
of being
project lead.
--
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.
_______________________________________________
dsdp-dd-dev mailing list
dsdp-dd-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/dsdp-dd-dev