Re: [Simulavr-devel] simulavr Tcl main framework

[Joel Sherrill]

Weddington, Eric wrote:
>  
>
>   
> > -----Original Message-----
> > From: 
> > address@hidden 
> > [mailto:address@hidden
> > u.org] On Behalf Of Joel Sherrill
> > Sent: Wednesday, March 25, 2009 2:22 PM
> > To: address@hidden
> > Subject: [Simulavr-devel] simulavr Tcl main framework
> >
> > Hi,
> >
> > I just got anacomp's check.tcl and checkdebug.tcl
> > capabilties to run using the Tcl main framework and
> > an anacomp specific extension file.
> >
> > As I convert the examples from a duplicated set of
> > Tcl files, is it OK to start committing new ones
> > based upon the framework?
> >
> > Am I in a corner of the world where no one cares
> > what I am doing? :D
> >     
>
> No, we really do care. It's just that at this point, you're about as expert in 
> this project as any of us (or more so). ;-)
>
> I think that as long we stick to the overall goals of the project (the main one 
> being speed of simulation), then what you're doing is probably all good. If you 
> don't hear any objections, then by all means use your best judgement. :-)
>   
OK.  This work is an improvement from a Tcl viewpoint because
you won't have to hand craft the Tcl start scripts.  And there
will not be as much duplication.  It also begins to address the
question of locating files from Tcl after installation (e.g. libsimulavr.so,
kbd.xbm, etc)

Would there be any objection to be changing the -T/-B argument
interpretation slightly.?  If you use  -T name, the behavior
will be the same.  But if you pass in -T "name1 name2", it will
register name1 and name2.  I only get one shot at a command
line argument in Tcl and it would be nice to have nearly
identical command line arguments.

All this in anticipation of setting up enough to begin
writing a simulation of an SPI attached device. ;)

Thanks.

--joel