Patches for RDI stop code and general features

Patches for RDI stop code and general features

[Tim Chick]

A while ago this was disscussed, and Fernando said he would
put in the RDI stop code. This hasn't happened as of the 10/10
snapshot, I'm not moaning, just maybe a reminder :-)

Since getting insight to work with my Jeeni, I have noticed
a few problems:

1. If you specify a symbol file on the command line, and
   it can't be opened, there is no warning that the file
   could not be opened.

2. I can't delete my breakpoints with the GUI, I have to
   use the console. If I try to delete a breakpoint the gui
   gives a popup:
   Error: No breakpoint at <sourcefile>.s:<line>

   This only occurs in breakpoints in assembly code.

3. Disasembly and mixed source and assembly view. On
   the ARM using the Jeeni it sometimes gets the disassembly
   wrong.

If I were to try and fix these, how would I supply patches for
approval? What program should I use to produce these patch files?

Thanks,
Tim


----- Original Message -----
From: "Fernando Nasser" <fnasser@cygnus.com>
To: "Grant Edwards" <grante@visi.com>
Cc: <insight@sourceware.cygnus.com>
Sent: Tuesday, September 12, 2000 10:04 PM
Subject: Re: Patches for source path and SIGINT warning


> Grant,
>
> I would like to fix the path thing using the libgui call instead of
> modifying Tcl.  I don't know when I will be given time to work on that but
> I will take note of it.  Of course, you already got your problem solved
and
> anyone else who needs it in a hurry can use your patch.  Thanks for dooing
it.
>
> Of course, as you say, we must not incorporate the SIGINT ignore patch.
Again,
> people who are doing exclusively RDI debugging can use your patch.
> I am tempted to get rid of that window and printing the signal received
message
> in the status line...
>
> Regards,
> Fernando
>
> P.S.: I think I will add your RDI stop code regardless.  Is it OK?
>
>
> Grant Edwards wrote:
> >
> > Attached are two patches against Insight-5.0.
> >
> > The first fixes the problem where the source window couldn't
> > open the source file because the path in the debugging info is
> > cygwin-style.
> >
> > Disclaimer: Somebody argued that this is the wrong way to fix
> >             it, but I didn't see any suggested right way, so
> >             those of us with work to do to will just have to
> >             suffer under the philosphical burden.  ;)
> >
> > The second disables the pop-up warning dialog about "target
> > received SIGINT" that happens when you press the "stop" button
> > to stop execution of an RDI target.
> >
> > Warning: it disables the SIGINT dialog completely, regardless
> >          of the target.  For the RDI target, the only source of
> >          SIGINT is a user-requested stop, so it's no big deal.
> >          For other targets, disabling this dialog may result in
> >          lossage of useful information.
> >
> >          So, there needs to be a more elegant solution, but I don't
> >          know enough Tcl to figure one out.
> >
> > Both of these patches (along with the gdb patches to enable the
> > "stop" button) are available at
> >
> >   ftp://ftp.visi.com/users/grante/gdb-rdi-patches/5.0
> >
>
> --
> Fernando Nasser
> Red Hat - Toronto                       E-Mail:  fnasser@cygnus.com
> 2323 Yonge Street, Suite #300           Tel:  416-482-2661 ext. 311
> Toronto, Ontario   M4P 2C9              Fax:  416-482-6299
>

Re: Patches for RDI stop code and general features

[Fernando Nasser]

Tim Chick wrote:
> 
> A while ago this was disscussed, and Fernando said he would
> put in the RDI stop code. This hasn't happened as of the 10/10
> snapshot, I'm not moaning, just maybe a reminder :-)
> 
It was not forgotten, just in a long queue of things to do.
It is on the works now and will be in pretty soon.


> Since getting insight to work with my Jeeni, I have noticed
> a few problems:
> 
> 1. If you specify a symbol file on the command line, and
>    it can't be opened, there is no warning that the file
>    could not be opened.
> 
You are right.  This is a bug but I don't know how to solve it.
This does not happen at GUI's command so I wonder how to get this
information back.

If you register this as a PR (problem report) in the insight bug database
you'll make sure I won't forget :-)

      http://source.redhat.com/insight
Follow the Bug link and from there to the dababase.

> 2. I can't delete my breakpoints with the GUI, I have to
>    use the console. If I try to delete a breakpoint the gui
>    gives a popup:
>    Error: No breakpoint at <sourcefile>.s:<line>
> 
>    This only occurs in breakpoints in assembly code.
> 
All right.  Another one.  Good catch, thanks!


> 3. Disasembly and mixed source and assembly view. On
>    the ARM using the Jeeni it sometimes gets the disassembly
>    wrong.
> 

This would be a binutils bug, probably.



> If I were to try and fix these, how would I supply patches for
> approval? What program should I use to produce these patch files?
> 

There are instructions on how to contribute somewhere but here are some basics:

1) If it is a small fix you don't need it, but if you are adding any functionality
   you will need an asignment of your work to the FSF (see fsf.org)

2) Use cvs or diff to produce the patch

There is a paragraph in the contribute file:

http://sources.redhat.com/cgi-bin/cvsweb.cgi/~checkout~/src/gdb/CONTRIBUTE?content-type=text/x-cvsweb-markup&cvsroot=src

that says:

        The patch itself. If you are accessing the CVS repository at:
        Cygnus, use "cvs update; cvs diff -c3p"; else, use "diff -c3p
        OLD NEW" or "diff -up OLD NEW". If your version of diff does
        not support these options, then get the latest version of GNU
        diff.

3) Don't forget the ChangeLog entry

4) Don't forget documentation changes

5) If adding a new feature or fixing a bug add a testcase or more if necessary.

The two last itens I am temporarely taking care of myself.



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

Re: Patches for source path and SIGINT warning

[Grant Edwards]

On Tue, Sep 12, 2000 at 05:04:38PM -0400, Fernando Nasser wrote:

> Grant Edwards wrote:
> > 
> > Attached are two patches against Insight-5.0.
> > 
> > The first fixes the problem where the source window couldn't
> > open the source file because the path in the debugging info is
> > cygwin-style.
> > 
> > Disclaimer: Somebody argued that this is the wrong way to fix
> >             it, but I didn't see any suggested right way, so
> >             those of us with work to do to will just have to
> >             suffer under the philosphical burden.  ;)
>
> I would like to fix the path thing using the libgui call
> instead of modifying Tcl.  I don't know when I will be given
> time to work on that but I will take note of it.  Of course,
> you already got your problem solved and anyone else who needs
> it in a hurry can use your patch.  Thanks for dooing it.

Did this problem get fixed?  

I'm about to bundle up a weekly snapshot for use by my
customers, and I don't know if I need to include my patch. I
suspect that the problem addressed by my original patch to
Tcl_TranslateFileName has been fixed elsewhere, -- can anybody
confirm this?  I don't currently have a Cygwin system to try it
on.

The patch added a call to cygwin_conv_to_win32_path() at the
beginning of Tcl_TranslateFileName().

I don't see anything in the libgui changelog, but I don't
really know what I'm lookig for...

Thanks,

-- 
Grant Edwards
grante@visi.com

Re: Patches for source path and SIGINT warning

[Keith Seitz]

On Mon, 13 Aug 2001, Grant Edwards wrote:

> Did this problem get fixed?

I believe it was.

> I'm about to bundle up a weekly snapshot for use by my
> customers, [...]

Can I have some customers? :-)

Keith

Re: Patches for source path and SIGINT warning

[Grant Edwards]

On Mon, Aug 13, 2001 at 10:28:21AM -0700, Keith Seitz wrote:
> On Mon, 13 Aug 2001, Grant Edwards wrote:
> 
> > Did this problem get fixed?
> 
> I believe it was.
> 
> > I'm about to bundle up a weekly snapshot for use by my
> > customers, [...]
> 
> Can I have some customers? :-)

You don't want them -- they cause all sorts of problems with
their entirely unreasonable demands that bugs be fixed and
suchlike.

-- 
Grant Edwards
grante@visi.com