Re: [fuse-emulator-devel] [fuse-emulator-cvs] SF.net SVN: fuse-emulator:[4861] trunk/fuse

[Stuart B. <sd...@zu...>]

On Sat, Jan 26, 2013 at 11:48:28AM +0000, fr...@us... wrote:
> @@ -424,6 +427,8 @@
>  static libspectrum_byte printer_zxp_read(libspectrum_word port GCC_UNUSED,
>  				         int *attached)
>  {
> +if(!settings_current.printer)
> +  return(0xff);
>  if(!printer_graphics_enabled)
>    return(0xff);
>  if(plusd_available)
> @@ -470,8 +475,11 @@
>  }

I'm not sure about this, although I would accept that if this is a bug,
then it was already a bug... but it seems to me that if I load a snapshot
with the ZX printer enabled, then I shouldn't get floating bus values when
reading from the ZX printer's I/O port... but that's exactly what I would
get if the global printers option is disabled.

I notice that printer_parallel_read() and printer_parallel_write() also
seem wrong in this regard, as printer_parallel_read() probably shouldn't
return floating bus values (although it may well depend on the parallel
interface in question as to what would be returned -- I don't know)...
and printer_parallel_write() should probably update parallel_data so that
if the 'printing' option is later enabled, then when the strobe bit is
toggled, the correct data is written.  Arguably, I would imagine that
parallel_data should be saved in snapshots (in a similar vein to the way
that specdrum_dac is saved), since it's plausible that some print drivers
might not update the data register when printing sequences of the same
character...

I could raise another ticket for this, but am not sure whether it needs
fixing for 1.1 or not.

Cheers,
-- 
Stuart