MI parser

MI parser

[Vladimir Prus]

Is there a parser for MI protocol, specifically for responses that gdb sends
(not for commands from frontend).

I see one is mentioned in
http://sources.redhat.com/ml/gdb/2004-08/msg00307.html

but can't find any MI parser in gdb CVS. Does it exist?

Thanks,
Volodya

Re: MI parser

[Daniel Jacobowitz]

On Wed, Sep 07, 2005 at 06:19:31PM +0400, Vladimir Prus wrote:
> 
> Is there a parser for MI protocol, specifically for responses that gdb sends
> (not for commands from frontend).
> 
> I see one is mentioned in
> http://sources.redhat.com/ml/gdb/2004-08/msg00307.html
> 
> but can't find any MI parser in gdb CVS. Does it exist?

Bob's written one.  It's not in CVS yet.

-- 
Daniel Jacobowitz
CodeSourcery, LLC

Re: MI parser

[Bob Rossi]

On Wed, Sep 07, 2005 at 10:29:33AM -0400, Daniel Jacobowitz wrote:
> On Wed, Sep 07, 2005 at 06:19:31PM +0400, Vladimir Prus wrote:
> > 
> > Is there a parser for MI protocol, specifically for responses that gdb sends
> > (not for commands from frontend).
> > 
> > I see one is mentioned in
> > http://sources.redhat.com/ml/gdb/2004-08/msg00307.html
> > 
> > but can't find any MI parser in gdb CVS. Does it exist?
> 
> Bob's written one.  It's not in CVS yet.

I've put up the current state of the parser here,
   http://brasko.net:81/bob/gdbmi_parser.tar

Basically, use at your own risk. Hopefully, in the near future, I'm
going to get the parser to an acceptable state to sumbit to GDB.

Several months ago I tested the parser against the gdb.mi testsuite
directory, and made sure that all MI output commands complied with the
grammar. They did, with a few minor changes to the grammar.

untar, run make, and with the executable made run './main b'.
That will test the front end against a few MI output commands that are
in the file called 'b'.

Thanks,
Bob Rossi

-break-list output format

[Stefan Burström]

Hello!
I am wondering why the output format of -break-list is rather different from
the rest of the mi commands. Is there anything specific that I need to take
care of when I parse the output?

regards,
Stefan Burstrom

-break-list output format

[Nick Roberts]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=unknown, Size: 349 bytes --]

Stefan Burström writes:
 > Hello!
 > I am wondering why the output format of -break-list is rather different from
 > the rest of the mi commands. Is there anything specific that I need to take
 > care of when I parse the output?

What version of GDB are you running?

How is it different?

What do you want to do with the output?

Nick

Re: -break-list output format

[Stefan Burström]

Hello Nick

On 2005-10-20, you wrote:

> What version of GDB are you running?

I am using 6.3

> How is it different?

Well, it includes a table header etc.If I want to build a table of the
output in my program I would of course just parse the output and build my
own tables with headers, labels etc.

For example, -file-list-exec-source-files just lists each file in a tuple.
But -break-list tells me the widh and height of the output etc. And I wonder
if I need to use it or if I can just skip the table header info. Finally, I
was just curious to know why the break-list output was done this way when
all other commands are more 'raw' in nature.

> What do you want to do with the output?

I am building a graphical frontend to gdb for AmigaOS4 :)

regards,
Stefan Burstrom

Re: -break-list output format

[Bob Rossi]

> Well, it includes a table header etc.If I want to build a table of the
> output in my program I would of course just parse the output and build my
> own tables with headers, labels etc.
> 
> For example, -file-list-exec-source-files just lists each file in a tuple.
> But -break-list tells me the widh and height of the output etc. And I wonder
> if I need to use it or if I can just skip the table header info. Finally, I
> was just curious to know why the break-list output was done this way when
> all other commands are more 'raw' in nature.

Well, I think it's a legacy issue. The table (with headers) has been
around for a long time. Long before MI was invented. So, I'm guessing
the writer of -break-list decided it would be good to keep that
information. You can ignore it if you choose.

Bob Rossi