Re: Scope Checking Patch

["Rob Quill" <rob.quill@gmail.com>]

On 11/06/07, Markus Deuling <deuling@de.ibm.com> wrote:
> Hi Rob,
>
> I haven't gone through your patch, but I have a suggestion. If I do a patch for GDB I do it like that:
>
> a) Checkout current head
> b) Create a copy of that directory e.g. gdb.new
> c) Do my changes in gdb.new
> d) Create patch by: diff -urN gdb/ gdb.new/ > diff-file

Thanks. I tried this, but I ended up with a 214MB diff file. I'm not
sure why this is, as I did a make distclean on both copies, I think it
may be due to the testsuite results not having been cleaned or
something. Does anyone have any ideas what to do?

Thanks,

Rob

> Then you the two diff files in one file and its better readable. Also its easier to apply just one file instead of two.
>
> Regards,
> Markus
>
> Rob Quill wrote:
> > Hi all,
> >
> > This is the first patch I have ever submitted to an open source
> > project, so I'm a little bit unsure of the process. The patch adds the
> > ability to check if a variable is in scope, as descibed here:
> >
> > http://sourceware.org/ml/gdb/2006-11/msg00149.html
> >
> > I have attached the diffs for the two files I've changed. However, I
> > am seeing some regressions against the current cvs, which I can't
> > understand, so I was wondering if a) the regressions happen for anyone
> > esle? (which presumably it does), b) if anyone could offer any
> > suggestions as to the cause, and c) give thier opinions on the patch.
> >
> > The patch allows the scope of constants, variables and variables in
> > classes/structures, by using $in_scope(variable_name) as an
> > expression, with value 1 if variable_name is in scope and 0 if it is
> > not.
> >
> > Any help and thoughts you can offer is much appreciated.
> >
> > Thanks,
> >
> > Rob
> >
>
>
> --
> Markus Deuling
> GNU Toolchain for Linux on Cell BE
> deuling@de.ibm.com
>
>