[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Wine



Christopher C Chimelis <chris@beezer.med.miami.edu> writes:

> On 15 Apr 1999, Brederlow wrote:
> 
> > Don't tell me, I'm using bochs on my alpha to emulate a complete 386.
> 
> Hehehehe...sorry :-)
...
> This is assuming we have a Wine lib that works on Alphas.  I assumed the

If it doesn't work, its time to fix it. :)

> regular discussion was just using em86 and a Wine compiled on i386.  If
> that assumption was wrong, then, yes, we would save ALOT of steps, but I
> still think Win API calls will be kinda creepy.

If you dont't emulate the api calls, why bother with wine? The native
windows runs just as fast or faster and is much more complete. Just
use bochs. I don't quite see the point in using wine if its not native.

> Also, another thought keeps popping up in my head.  The absolute best idea
> would be to somehow have Wine and em86 merged code-wise to allow em86
> handle i386 calls and Wine handle just emulating Windows and Win API
> calls (even those from Alpha binaries...assuming the supporting libs are
> ok).  I'm thinking that this kind of merge would allow us to run 386 and
> Alpha Win binaries under the same Wine setup (just conjecture...alot of

Why? em86-wine would be dynamically linked against wine-lib-alpha and
alpha wine binaries also. alpha binaries would not start the emulator, 
but just link against the lib and i386 win bins would start the
emulator.
Isn't it that way its already on intel?

> work would need to be done...mostly by Compaq since only they have the
> source to em86 from what I recall).  That would be the ideal setup
> allowing us to run some things native if possible.
>
> > Since we have the source fo wine and its lib, what we now need is an
> > emulator that maps the i386 win api calls to a native alpha wine
> > lib. Thats the point where an emulator should become useable.
> 
> So we would have to incorporate em86 somehow into the backend of Wine so
> that specific i386 calls are emulated by it, but API calls are handled by
> the lib....correct?  I think I'm understanding this idea better now :-)

Yep, thats it. It might be better to use bochs or freemware instead of 
em86, because of its license.

> Another thing we would have to ensure is, if we (or someone) decide to
> implement this, that the API calls are properly implemented based on the
> arch's headers.  I'm not sure if i386 Windows NT and Alpha NT have the
> same header files (meaning that the functions are called with arguments of
> the same types across platforms), but if they do, then running Alpha NT
> binaries would be ALOT easier and not a whole lot of work would be
> required on Wine to accomplish this (well, at least compared to getting
> i386 stuff to work).

Otherwise we need an wine-i386 and wine-alpha for alpha. But I don't
think thats a big problem. How many alpha-NT programs are out there
that one wants to use?

> > Yes, such a switch is pretty handy, but those are quite expensive and
> > the cheaper once have poor quality. Its easier to use remote displays
> > with X and telnets.
> 
> Agreed.  Too bad I haven't found a really good X server for most of my Win
> equipment yet :-(

May the Source be with you.
			Goswin


Reply to: