Re: [Simulavr-devel] Introducement / Status request

[Theodore A. Roth]

On Mon, 2 Sep 2002, qmax wrote:

:) Now i'm curious about status of current developement process -
:) to do not dublicate someone's else work if i take some,
:) and also about some CVS usage/patchposting policies and conventions,
:) other then mentioned in CVS FAQ, ...if any.

Don't worry about duplicating work. There's not much going on right now
since I'm pretty busy with a lot of different projects right now.

As for cvs/patching, there's now set in stone rules, only my preferrence.
If you have a cvs sandbox, it's really easy to generate a patch as such
(from root of sandox):

  $ cvs diff -u > <some_descriptive_name>.diff

Then I can review them and apply the patch with:

  $ patch -p0 --dry-run < <some_descriptive_name>.diff
  $ patch -p0 < <some_descriptive_name>.diff

If the dry-run fails miserably, chances are pretty good I'll ignore the
patch.

Patches should be posted to the simulavr-dev list so I won't lose them in
my inbox (which has been happening far too often lately ;-).

:)
:) Currently im working on 16 bit Timer1 with COM and PWM modes (almost
:) complete and amazing). The OC and PWM require their output to be
:) displayed (I've made display_io_port in ext_port_wr), and this is
:) rather question of external i/o.

Cool. I'd be interrested in seeing a patch once you get it working. Be
sure to include a simple test_asm file which exercises the functionality.

:)
:) ***
:)
:) I've grepped simulavr-devel archive and found no discussions about
:) virtual hardware,mplementation,  other then in TODO.
:)
:) My first-shot suggestion is to implement PIN[A-D] read/write facilities
:) in communication proto with external process, similar to, or the same
:) as disp-proto. This simply requires latching/updating these addresses
:) in main.c: ext_port_* instead of "writing 0x%02x to 0x%04x" and "Enter
:) a byte of data to read into 0x%04x" (While not dealing with external
:) signal source).

I don't particularly like the ext port io being done in main.c as it is
now. It was just an ugly hack to so I could see what's going on.

I'd really like to see a plug-in system, where simulavr just checks a
directory for modules to plugin which provide whatever interface you want
for a port. This would give the user more flexibility in tweaking simulavr
to their needs without having to dig into the code so much.

:)
:) Im (up | trying) to write a tcl-powered GUI display program (displavr)
:) to introduce widgets representing ports statuses in human-viewable form
:) /* mostly those blinkenlighten; attempt to make oscillogramme (egg, to
:) watch PWM) leads to requirement of proto packets time/clock
:) synchronization */. Adding ability to write to PIN[A-D] addresses will
:) make it possible to write tcl scripts implementing hardware enviroment
:) logic or simple input widgets.

My only concern with using tcl/tk is that it will be very slow, but if you
are hooking in to the disp interface, you are free to do that. Which is
why the disp interface uses a coprocess.


Ted Roth