Separate preferences for native and cross debugger in Insight.

Separate preferences for native and cross debugger in Insight.

[Ton van Overbeek]

On June 5 I submitted a small patch to use separate gdbtk.ini files
for native and cross debuggers.
Up till now I have not seen any reaction on the list, or received
any comments.
If I did anything wrong in submitting my patch, please tell me.
It is not clear to me if Insight follows the same rules/paper work
as gdb. If that is the reason I'll be glad to submit a form.

So, should I resubmit my patch in a different form, complete a
form, or did nobody notice it ?

Regards, and hoping for some reactions,

Ton van Overbeek
P.S. IMHO the Insight logo on the new webpages should also appear in the
     Insight about box.

Re: Separate preferences for native and cross debugger in Insight.

[Fernando Nasser]

Ton van Overbeek wrote:
> 
> On June 5 I submitted a small patch to use separate gdbtk.ini files
> for native and cross debuggers.

It is a small patch, but it represents a significant change.

I would particularly like to hear some comments from Tom Tromey about
this change.  He has spent some time thinking about the way we organize
preferences and have created our "session" facilities.

BTW, I wonder if the "sessions" code should not be enough to solve your
original problem...


> Up till now I have not seen any reaction on the list, or received
> any comments.
> If I did anything wrong in submitting my patch, please tell me.
> It is not clear to me if Insight follows the same rules/paper work
> as gdb. If that is the reason I'll be glad to submit a form.
> 
> So, should I resubmit my patch in a different form, complete a
> form, or did nobody notice it ?
> 

No, your submission was correct.  You must understand that Open Source
projects are based on volunteer work and people have day-time jobs. 
Eventualy we get hold up by work and some backlog builds up.  But do
feel free to 'ping' us if you suspect you submission may have gone
unoticed (yes, it may happen with so many e-mails).  Maintainers usually
apreciate that people do that.

And, regardless of what people decide (make this change or not), thank
you for taking the time to do this. 

> P.S. IMHO the Insight logo on the new webpages should also appear in the
>      Insight about box.

Good idea.  I will ask for someone to put together something like that
for us and post it to the list.


Regards,
Fernando


-- 
Fernando Nasser
Red Hat Canada Ltd.                     E-Mail:  fnasser@redhat.com
2323 Yonge Street, Suite #300
Toronto, Ontario   M4P 2C9

Re: Separate preferences for native and cross debugger in Insight.

[Ton van Overbeek]

At 12:34 6/10/2001 -0400, Fernando Nasser wrote:
>Ton van Overbeek wrote:
>> 
>> On June 5 I submitted a small patch to use separate gdbtk.ini files
>> for native and cross debuggers.
>
>It is a small patch, but it represents a significant change.
>
>I would particularly like to hear some comments from Tom Tromey about
>this change.  He has spent some time thinking about the way we organize
>preferences and have created our "session" facilities.
>
>BTW, I wonder if the "sessions" code should not be enough to solve your
>original problem...
>

Thanks for the pointer to the 'sessions' code. It does solve part
of my original problem. However things like window layout, which windows
open, etc. are not saved in a 'session'.
So when running (as in my case) both the native cygwin gdb and a cross
gdb/insight under cygwin there is still a mix of settings. Also a session
(= mainly exec file, search paths, etc.) for the native gdb is
garbage for the cross and vice versa.
There is no entry in the on-line help for 'sessions'. If I get to know
it a little bit better I might find to some time to propose an addition
to the srcwin help about sessions.

>No, your submission was correct.  You must understand that Open Source
>projects are based on volunteer work and people have day-time jobs. 
>Eventualy we get hold up by work and some backlog builds up.  But do
>feel free to 'ping' us if you suspect you submission may have gone
>unoticed (yes, it may happen with so many e-mails).  Maintainers usually
>apreciate that people do that.
>

My previous message was indeed intended as a 'ping'.
I do have a full-time job myself, so I certainly do appreciate all
the volunteer work you guys put in. I can imagine the problems you must
have when you are working in one specific area of gdb/insight and then a
message comes in affecting a completely different area.

>And, regardless of what people decide (make this change or not), thank
>you for taking the time to do this. 
>
Thanks and keep up the good work.

Ton van Overbeek

Re: Separate preferences for native and cross debugger in Insight.

[Tom Tromey]

>>>>> "Ton" == Ton van Overbeek <tvoverbe@cistron.nl> writes:

Ton> On June 5 I submitted a small patch to use separate gdbtk.ini files
Ton> for native and cross debuggers.

I just looked at this.

My understanding is that the long-term goal of gdb is that it be able
to handle many different targets from a single debugger.  Given that,
it seems to me that we should build Insight to make this transition
easy.  So, for instance, perhaps the relevant preference information
should be separated out inside Insight itself.

Tom

Re: Separate preferences for native and cross debugger in Insight.

[Tom Tromey]

>>>>> "Ton" == Ton van Overbeek <tvoverbe@cistron.nl> writes:

Ton> Thanks for the pointer to the 'sessions' code. It does solve part
Ton> of my original problem. However things like window layout, which
Ton> windows open, etc. are not saved in a 'session'.

Window layout is something I would like to see saved in a session (at
least optionally).  It has simply not been implemented yet.  I've done
most of the work on session saving in Insight.  However, I only rarely
actually do Insight hacking.  I'd say the session code gets a new
feature about every four months :-(.

Ton> So when running (as in my case) both the native cygwin gdb and a
Ton> cross gdb/insight under cygwin there is still a mix of
Ton> settings. Also a session (= mainly exec file, search paths, etc.)
Ton> for the native gdb is garbage for the cross and vice versa.

Good point.

One idea would be to have a different session list for each target, or
at least somehow annotate each particular session with the target
information.  That way a native debugger could know not to display a
session which was created by a cross debugger.

I believe this sort of change would be very easy to make.  It would
probably be localized entirely within session.tcl.

Ton> There is no entry in the on-line help for 'sessions'. If I get to
Ton> know it a little bit better I might find to some time to propose
Ton> an addition to the srcwin help about sessions.

I wasn't aware we had online help :-(.  Otherwise I would have updated
it.  Where is it kept?  I'd be happy to try to document sessions,
though I can't guarantee that I would get to it in the very near
future.

Tom

Re: Separate preferences for native and cross debugger in Insight.

[Fernando Nasser]

Tom Tromey wrote:
> 
> Ton> There is no entry in the on-line help for 'sessions'. If I get to
> Ton> know it a little bit better I might find to some time to propose
> Ton> an addition to the srcwin help about sessions.
> 
> I wasn't aware we had online help :-(.  Otherwise I would have updated
> it.  Where is it kept?  

It is a little bit hidden:  gdbtk/library/help



> I'd be happy to try to document sessions,
> though I can't guarantee that I would get to it in the very near
> future.
> 

We understand.  Whenever you find some time, just post some plain text to the list and we will add it to the help for you (if this helps).

Someone has posted a suggestion/request for an on-line manual.  I was thinking of making the on-line help available as an on-line manual (at least until we have something more substantial).


-- 
Fernando Nasser
Red Hat - Toronto                       E-Mail:  fnasser@redhat.com
2323 Yonge Street, Suite #300
Toronto, Ontario   M4P 2C9

Re: Separate preferences for native and cross debugger in Insight.

[Tom Tromey]

Fernando> Someone has posted a suggestion/request for an on-line
Fernando> manual.  I was thinking of making the on-line help available
Fernando> as an on-line manual (at least until we have something more
Fernando> substantial).

Sounds good to me.  I wouldn't mind if we adopted a rule that any
user-visible change must be accompanied by a patch to the manual.

Tom