Bryce Denney wrote:
> 
> On Fri, 8 Jun 2001, Martin wrote:
> > I am in favor of keeping older snapshots.
> > It is easy that way to find when some bug was introduced without
> > reconstructing an older version with CVS.
> > Also it is easy that way to track where and when bochs and plex86 grew
> > out of eachother.
> > For example: I am now implementing the diskd feature (use two disk
> > images) which was build into bochs a few months ago. This uses features
> > in bochs bios 1.3. This bios is quite different from plex86 bios version
> > 1.3. Only 1.1 is the same for both.
> 
> Martin,
> 
> The rombios is an example of code that can truly be shared between bochs
> and plex86.  I don't think there are any bochs-specific or plex-specific
> hacks in either of our rombios files (if there are, they are minor and a
> #define could easily take care of it).  It's already on my to-do list to
> integrate the plex86 changes to rombios, and I'm happy to try to explain
> any changes we've made that you are interested in.
> 
> Let me know.  I hate for the two projects to re-invent things that could
> easily be shared.

Agreed fully.

Up to now, I've tried to keep some things static, so that developing
plex86 would be easier.  Now that plex86 is coming along nicely, I've also
had it in mind to synchronize with the bochs project on the BIOS front,
as well as some others like device models.

For device emualation, if both projects can agree on a device model
plugin interface or such, I think we could easily share devices,
both current and future ones, like a new video card emulation,
network stuff etc.

Though, for device models the news that perhaps not everybody
wants to hear is that they should be in C, so that they could
be compiled into the kernel module of plex86 for better performance,
loaded dynamically as a plugin in user space, or static in user space,
depending on the situation and performance requirements.

-Kevin

Reply via email to