On Wed, Jun 8, 2011 at 10:16 AM, Wolfram Sang <[email protected]> wrote:

> Hi Jon,
>
> thanks for your patches.
>
> On Wed, Jun 08, 2011 at 09:34:02AM -0400, Jon Ringle wrote:
> > Registers sometimes have read side-effects.
> > Don't automatically readback after a write unless enabled.
>
> Can we agree on having readback enabled by default? That's to keep the
> old behaviour. I'd also think that it is the more common case.
>
> Ok.


> > Readback on write can be enabled with -r on command line
> > or 'readback 1' from memedit prompt.
> >
> > Signed-off-by: Jon Ringle <[email protected]>
> > ---
> >  memedit.c         |    4 +++-
> >  memedit.y         |   23 +++++++++++++++++------
> >  memedit_parser.c  |   18 +++++++++++++-----
> >  memedit_parser.gp |    3 ++-
> >  memedit_parser.h  |    7 ++++---
> >  y.tab.c           |   51
> +++++++++++++++++++++++++++++++--------------------
> >  6 files changed, 70 insertions(+), 36 deletions(-)
>
> You don't need to include updates to the generated files. Updates to the
> helptext/manpage would be good, though :)
>
> > +     if (strncmp (param, "readback", 8) == 0) {
> > +             readback = val;
> > +             printf ("readback = %d\n", readback);
> > +             return;
> > +     }
>
> Also toggle state if there is no parameter?
>
> The way that set_variable() is written, I don't see an easy way to tell if
no parameter was given...


> > diff --git a/memedit_parser.gp b/memedit_parser.gp
> > index 3eb75c3..edc3d5a 100644
> > --- a/memedit_parser.gp
> > +++ b/memedit_parser.gp
> > @@ -1,4 +1,5 @@
> >  /* memedit_parser.gp */
> > -#mandatory memedit
> > +#mandatory file
>
> Original line is very wrong, but file is also not mandatory; one could
> use open after memedit started.
>
> Yes. It seems to me that this is a deficiency in genparse. It really isn't
mandatory. It's only there so that the correct usage() is generated.

Jon
-- 
ptxdist mailing list
[email protected]

Reply via email to