Re: Apple version gdb-309

Re: Apple version gdb-309

[Jim Ingham]

Nick,


> Does the version:
>
> GNU gdb 5.3-20030128 (Apple version gdb-309) (Thu Dec  4 15:41:30 GMT 
> 2003)
> Copyright 2003 Free Software Foundation, Inc.
> GDB is free software...
> ...
> This GDB was configured as "powerpc-apple-darwin".
>
> come out of the FSF repository or Apple's? I realise it says
> Apple version gdb-309 but Mandrake has:
>
> GNU gdb 5.2.1-2mdk (Mandrake Linux)
>
> and this is standard gdb AFAICT.

Not quite sure what the Mandrake version has to do with this, but a gdb 
that says "Apple version gdb-xxx" on a Mac OS X system is most likely 
from the Apple CVS repository.  The rest of the numbering: 
"5.3-20030128" is the gdb version this roughly maps onto (5.3 in this 
case) and the date of the last merge from the FSF to our tree (Jan. 
28th, 2003 in this case).

>
> I've had a report of the following transaction:
>
> (gdb) server interpreter mi "-var-update *"
> Switching to interpreter "console"
> (gdb)
>
>
> AFAIK gdb 5.3 in the FSF repository (at sources.redhat.com) doesn't 
> know
> about the command "interpreter mi".
>


This snippet doesn't work in either gdb.  Is there a typo somewhere 
here?  Maybe this was

(gdb) interpreter execute mi "-var-update *"

That would work in our gdb and in the FSF version of gdb that has the 
interpreter code.  The interpreter subsystem was added to the FSF gdb 
around 6.0 (IIRC) by Keith Seitz, Elena and Andrew, from a version 
originally done at Apple.  The FSF version and the Apple gdb-309 
version are somewhat different.  We are working to merge the FSF TOT 
back to our codebase, though the interpreters bit will continue to be a 
little different when we finish this ('cause there are a few things 
about the way I did it that I like better :-)  But I think what I have 
will be a superset, and mi or cli code that uses the interpreters 
commands from the FSF version should still work on our version.


> If these are two different versions where do they overlap? (or, 
> conversely,
> where do they differ?)

There are substantial differences between the Apple gdb and the FSF 
version.  But we don't keep a comprehensive list.

>
> More specifically will the future versions for Mac OS
> ( == powerpc-apple-darwin? )  have the changes for -stack-list-locals 
> and
> -var-list-children that I committed to the FSF repository on 
> 2004-01-20?

Yes.  Both these commands are different in our gdb from the version in 
the FSF repository, but as I merge the differences, I will try to keep 
the behavior you added working.  It doesn't look too hard, though in 
the case of stack-list-locals, your change uses argument values that we 
had already taken (I think Jason mentioned this at the time you were 
adding it).  So I have to get Xcode to change as well.  Fortunately for 
now we keep the Xcode and gdb versions in sync, so this won't be a 
problem...

Jim
--
Jim Ingham                                   jingham@apple.com
Developer Tools
Apple Computer

Re: Apple version gdb-309

[Nick Roberts]

 > > More specifically will the future versions for Mac OS
 > > ( == powerpc-apple-darwin? )  have the changes for -stack-list-locals 
 > > and
 > > -var-list-children that I committed to the FSF repository on 
 > > 2004-01-20?
 > 
 > Yes.  Both these commands are different in our gdb from the version in 
 > the FSF repository, but as I merge the differences, I will try to keep 
 > the behavior you added working.  It doesn't look too hard, though in 
 > the case of stack-list-locals, your change uses argument values that we 
 > had already taken (I think Jason mentioned this at the time you were 
 > adding it).  So I have to get Xcode to change as well.  Fortunately for 
 > now we keep the Xcode and gdb versions in sync, so this won't be a 
 > problem...

Jim,

I did say to Jason (12 Dec 2003):

Me> I don't see why both our changes can't be accommodated. This patch uses a
Me> switch statement for each value of print_values. If for some reason Apple
Me> need -stack-list-locals 2, I dont mind using another value.

He didn't seem enthusiastic but perhaps I missed his point. All my changes to
Emacs and GDB are still in CVS, so, as far as I'm concerned, you're welcome to
change my values if this makes things easier for your merge.

Thanks for the info,

Nick

Re: Apple version gdb-309

[Jim Ingham]

On Mar 29, 2004, at 3:22 PM, Nick Roberts wrote:

>
>>> More specifically will the future versions for Mac OS
>>> ( == powerpc-apple-darwin? )  have the changes for -stack-list-locals
>>> and
>>> -var-list-children that I committed to the FSF repository on
>>> 2004-01-20?
>>
>> Yes.  Both these commands are different in our gdb from the version in
>> the FSF repository, but as I merge the differences, I will try to keep
>> the behavior you added working.  It doesn't look too hard, though in
>> the case of stack-list-locals, your change uses argument values that 
>> we
>> had already taken (I think Jason mentioned this at the time you were
>> adding it).  So I have to get Xcode to change as well.  Fortunately 
>> for
>> now we keep the Xcode and gdb versions in sync, so this won't be a
>> problem...
>
> Jim,
>
> I did say to Jason (12 Dec 2003):
>
> Me> I don't see why both our changes can't be accommodated. This patch 
> uses a
> Me> switch statement for each value of print_values. If for some 
> reason Apple
> Me> need -stack-list-locals 2, I dont mind using another value.


>
> He didn't seem enthusiastic but perhaps I missed his point. All my 
> changes to
> Emacs and GDB are still in CVS, so, as far as I'm concerned, you're 
> welcome to
> change my values if this makes things easier for your merge.

I think Jason's point was that as long as we were just using numbers, 
we were going to keep stepping on each other's toes, and it would be 
better to use string enumerators.  That would make code that users of 
the mi would write easier as well.  So there is no reason to go out of 
our way fixing something that was not a very good solution in the first 
place...

And it won't be hard to get the Xcode folks here to adjust to this, so 
I'll just fix it here for now.  But then we should cook up a set on 
name strings to stand in the place of the {0,1,2} arguments that we 
currently use.

Jim
--
Jim Ingham                                   jingham@apple.com
Developer Tools
Apple Computer

Re: Apple version gdb-309

[Nick Roberts]

Jim,

As Emacs release draws closer, I would like my mode (gdb-ui.el) to work with
Tiger. I've looked over some old e-mails and (Mon, 29 Mar 2004) you have said:

 > > I've had a report of the following transaction:
 > >
 > > (gdb) server interpreter mi "-var-update *"
 > > Switching to interpreter "console"
 > > (gdb)
 > >
 > >
 > > AFAIK gdb 5.3 in the FSF repository (at sources.redhat.com) doesn't 
 > > know
 > > about the command "interpreter mi".
 > >
 > 
 > 
 > This snippet doesn't work in either gdb.  Is there a typo somewhere 
 > here?  Maybe this was
 > 
 > (gdb) interpreter execute mi "-var-update *"

On FSF gdb (GNU gdb 6.3.50.20050502-cvs) "interpreter mi" expands to
"interpreter-exec mi" and this is still the command that I use in gdb-ui.el:

usage: interpreter-exec <interpreter> [ <command> ... ]

If I try your example, I get:

(gdb) interpreter execute mi "-var-update *"
Could not find interpreter "execute".

Tiger ( == powerpc-apple-darwin8.0? ) uses a version of GDB based on 6.0 right? 

What should I use for this case, "interpreter mi" or "interpreter execute mi"?

Thanks,

Nick

Apple version gdb-309

[Nick Roberts]

Does the version:

GNU gdb 5.3-20030128 (Apple version gdb-309) (Thu Dec  4 15:41:30 GMT 2003)
Copyright 2003 Free Software Foundation, Inc.
GDB is free software...
...
This GDB was configured as "powerpc-apple-darwin".

come out of the FSF repository or Apple's? I realise it says 
Apple version gdb-309 but Mandrake has:

GNU gdb 5.2.1-2mdk (Mandrake Linux)

and this is standard gdb AFAICT.

I've had a report of the following transaction:

(gdb) server interpreter mi "-var-update *"
Switching to interpreter "console"
(gdb)


AFAIK gdb 5.3 in the FSF repository (at sources.redhat.com) doesn't know
about the command "interpreter mi".

If these are two different versions where do they overlap? (or, conversely,
where do they differ?)

More specifically will the future versions for Mac OS 
( == powerpc-apple-darwin? )  have the changes for -stack-list-locals and
-var-list-children that I committed to the FSF repository on 2004-01-20?

Thanks for any help,

Nick