Re: GDB and Guile

Re: GDB and Guile

[Andrey Smirnov]

On Sun, Jan 8, 2012 at 7:23 PM, Joel Brobecker <brobecker@adacore.com> wrote:
> Hello,
>
> Richard Stallman asked us to post the following announcement:
>
> GDB ought to support extensibility using Guile, the GNU extensibility
> package (an implementation of Scheme).  We are looking for people to
> write the code to interface the two.  Please contact us at gdb-patches
> AT sourceware DOT org if you are interested.
>

Since it sounds like an interesting project, I guess I can try to do
this, but:
 1. I have only superficial knowledge of GDB code base.
 2. The situation with my Guile embedding knowledge is the same.
 3. My lisp experience is limited to a single small patch for Magit (and
 it is in Elisp instead of Scheme).
 4. I'll be doing it in my free time which on average would amount to
 10-12 hours/week.
 5. I still have to settle Copyright Assignment with my new employer.
 (Which I hope will go well, but one never knows)
 6. I still have some -Wshadow patches to send.
I guess this makes me the worst possible candidate, but if, for some
unfathomable reason, there are no better candidates, feel free to nudge
me.

P. S. Don't want to create separate thread, so I'll just ask here: Is it
OK to apply approved -Wshadow patches, provided of course #5 up above is
resolved, or should I wait more?

Re: GDB and Guile

[Tom Tromey]

Andrey>  5. I still have to settle Copyright Assignment with my new employer.
Andrey>  (Which I hope will go well, but one never knows)

Andrey> P. S. Don't want to create separate thread, so I'll just ask here: Is it
Andrey> OK to apply approved -Wshadow patches, provided of course #5 up above is
Andrey> resolved, or should I wait more?

If you changed employers, and you are not sure whether the old
assignment holds, then you must contact the FSF for guidance.
Please let us know when it is all resolved.

Tom

Re: GDB and Guile

[Joel Brobecker]

> > I guess this makes me the worst possible candidate, but if, for some
> > unfathomable reason, there are no better candidates, feel free to nudge
> > me.
[...]
> Even if I'm not very familiar with either GDB core or Guile, I want to
> participate.  Is it OK?

The project does not need to be tackled by a single person. I think
it is perfectly OK if both of you coordinate together. If other
candidates come along, they should be integrated in the group.
And even if they were more familiar with Guile, it would be a great
asset to the group, but that does not mean that you should feel
compelled to stay away from the project because you would be "less
qualified".

-- 
Joel

Re: GDB and Guile

[Jiang Jilin]

On Tue, Jan 10, 2012 at 11:55 AM, Joel Brobecker <brobecker@adacore.com> wrote:
>
> > > I guess this makes me the worst possible candidate, but if, for some
> > > unfathomable reason, there are no better candidates, feel free to nudge
> > > me.
> [...]
> > Even if I'm not very familiar with either GDB core or Guile, I want to
> > participate.  Is it OK?
>
> The project does not need to be tackled by a single person. I think
> it is perfectly OK if both of you coordinate together. If other
> candidates come along, they should be integrated in the group.
> And even if they were more familiar with Guile, it would be a great
> asset to the group, but that does not mean that you should feel
> compelled to stay away from the project because you would be "less
> qualified".
>

Got it, Thank you!

> --
> Joel




--
Jiang

Re: GDB and Guile

[Mike Frysinger]

[-- Attachment #1: Type: Text/Plain, Size: 465 bytes --]

On Sunday 08 January 2012 22:23:50 Joel Brobecker wrote:
> GDB ought to support extensibility using Guile, the GNU extensibility
> package (an implementation of Scheme).  We are looking for people to
> write the code to interface the two.  Please contact us at gdb-patches
> AT sourceware DOT org if you are interested.

i know the request/announce doesn't say it would, but i hope that gaining 
Guile support wouldn't affect existing python support
-mike

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: GDB and Guile

[Joel Brobecker]

> i know the request/announce doesn't say it would, but i hope that
> gaining Guile support wouldn't affect existing python support

I can confirm that Python support will not be affected.

-- 
Joel

Re: GDB and Guile

[Mark Kettenis]

> > i know the request/announce doesn't say it would, but i hope that
> > gaining Guile support wouldn't affect existing python support
>
>  I can confirm that Python support will not be affected.

But I'd say that, given the limited resources we have, supporting two
extension languages is undesirable.  Let's face it, Guile support is
something
that Stallman wants, but probably not a lot of other people.  Now if somebody
steps up to do the work and is motivated enough to build a community to
support it (like what happened with Python support) we should of course
seriously consider integrating that work.  But if somebody wants to do
this just because Stallman says so, I think we shouldn't encourage
him/her.

GDB and Guile

[Tom Emerson]

Greetings,

I see on the GDB site that you were looking for people who might be
interested in integrating GDB with Guile. I'm definitely interested in
contributing to the project if there's room.

Thanks for your time,

    -tree

-- 
Tom Emerson
tremerson@gmail.com
http://www.dreamersrealm.net/tree

Re: GDB and Guile

[Doug Evans]

On Thu, Jun 28, 2012 at 8:08 AM, Tom Emerson <tremerson@gmail.com> wrote:
> Greetings,
>
> I see on the GDB site that you were looking for people who might be
> interested in integrating GDB with Guile. I'm definitely interested in
> contributing to the project if there's room.
>
> Thanks for your time,
>
>    -tree

Hi.  There is indeed room.

The thing to do is to become familiar with libguile and with gdb's
Python interface, and come up with something sensible to provide in
gdb+guile what gdb provides today in gdb+python.  Keeping the gdb side
of things (as opposed to the Python or Guile side of things) clean and
maintainable (given that we'd be now supporting multiple external
scripting languages) is of the utmost importance.  [which isn't to say
the Python or Guile side are of lesser value of course :-)]

Have fun, and thanks!

Re: GDB and Guile

[Ludovic Courtès]

Hello,

Please don’t hesitate to ask for feedback on guile-user@gnu.org, or on
#guile on Freenode.  We’d be happy to support you in this endeavor.  :-)

Thanks,
Ludo’.

Re: GDB and Guile

[Tom Emerson]

On Mon, Jul 9, 2012 at 12:19 PM, Ludovic Courtès <ludo@gnu.org> wrote:
> Please don’t hesitate to ask for feedback on guile-user@gnu.org, or on
> #guile on Freenode.  We’d be happy to support you in this endeavor.  :-)

Thanks Ludo. At this point I've been reading through the current
Python integration in GDB to get a feel for where the common touch
points would be. I'm also looking at the various Python examples and
re-imagining them in Scheme. What, if anything, has been done related
to Guile/GDB integration heretofore? Has anyone collected a wish list
or something similar? Parity with the Python functionality is
obviously the first order of business, which in turn speaks to
generalizing the interface between GDB and Python so that it is more
(extension language) neutral.

Peace,

    -tree

-- 
Tom Emerson
tremerson@gmail.com
http://www.dreamersrealm.net/tree

Re: GDB and Guile

[Tom Tromey]

Tom> Parity with the Python functionality is obviously the first order
Tom> of business, which in turn speaks to generalizing the interface
Tom> between GDB and Python so that it is more (extension language)
Tom> neutral.

It depends on what exactly you mean here, but to I don't think we want
any kind of generic extension API.  I think the way that gdb is exposed
to Python has to be ad hoc, and for good reason -- we want control over
API evolution as seen by Python to a degree that we actively don't want
internally.

If you can give more details on what you plan I can give better
feedback.  I just didn't want you to set off down a bad path.

Tom

Re: GDB and Guile

[Tom Emerson]

On Mon, Jul 9, 2012 at 5:09 PM, Tom Tromey <tromey@redhat.com> wrote:
> It depends on what exactly you mean here, but to I don't think we want
> any kind of generic extension API.  I think the way that gdb is exposed
> to Python has to be ad hoc, and for good reason -- we want control over
> API evolution as seen by Python to a degree that we actively don't want
> internally.

I've been working on the assumption (I've only been looking at this in
my spare time for a week or so) that many of the places Guile would
interact with the rest of GDB would be the same as those that Python
uses, which would in turn lead to additional conditionalized or
duplicated functionality that could be abstracted. The suggestion that
there be a 'generic' extension API was motivated by that. If that
isn't a desirable or realistic approach then it is definitely better
not to even consider that path.

Thanks!

    -tree

-- 
Tom Emerson
tremerson@gmail.com
http://www.dreamersrealm.net/tree