Re: [hatari-devel] Handing WinUAE CPU core write_log()

[Eero Tamminen <oak@xxxxxxxxxxxxxx>]

Hi Thomas,

I pushed this.  If there's something additional that you can
do to minimize Hatari & WinUAE CPU core differences, please
do.


	- Eero

On 05/07/2018 12:15 AM, Eero Tamminen wrote:
> On 05/06/2018 11:13 PM, Thomas Huth wrote:
> > Am Sun, 6 May 2018 18:28:00 +0300
> > schrieb Eero Tamminen <oak@xxxxxxxxxxxxxx>:
> > > What do you think of the attached way of handling
> > > WinUAE CPU core write_log()?
> > >
> > > It uses C99 feature to map write_log() to Log_Printf()
> > > instead of the bWriteLog stuff you added a while ago.
> >
> > Basically a good idea, but does that really work? IIRC I once tried
> > something similar, but I ran into some problems ... isn't write_log
> > also required for build68k and gencpu? But we do not link the log
> > code to these binaries, do we? Have you tried rebuilding Hatari
> > completely this way?
>
> build68k and gencpu do get rebuilt when changing these headers,
> and the did need changing[1], so they do work with it.  I guess
> they get nowadays linked with the log code.
>
> [1] build68k for abort() macro, and gencpu for commenting out
> its version of write_log().
>
>
> > Anyway, you certainly don't need these "C99" comments in the sourcecode
> > here - we already use __VA_ARGS__ in a couple of places already, so 
> > this is
> > already a prerequisite for building Hatari.
>
> Ok, that's great to know!  I was afraid there would still be
> problems on Windows or something...
>
>
>      - Eero