Re: [hatari-devel] Hatari 1.6.1 test build for Mac OS X

[Bob Carpenter <bcripon89@xxxxxxxxx>]

@Stefan - I would agree that my build would be likely to fail if X11 were not installed. Also, my build is limited to Intel 10.6 and 10.7. If Jerome's build is completed, that sounds like the best answer. If he builds SDL from the current 1.2.15 pre-release source code, I am concerned that the same crash will occur if you make a change that forces Hatari to be reset.

@Eero - I was just surprised by how much the video capture slowed down Hatari's responsiveness. I do not have a RAID array. However, I do have a 2.7Ghz Core i5 with a 7200rpm drive. Since I do not anticipate doing video captures in the future, the performance hit is not a concern for me.

Bob C

On Mon, Jan 16, 2012 at 3:39 PM, Stefan Meier <stefan_meier99@xxxxxx> wrote:

Might be hard to find ready to use frameworks for the required libs. For libpng there is a precompiled framework available here: http://ethan.tira-thompson.com/Mac_OS_X_Ports.html (although even this one is actaully outdated). For the other libraries I didn't find up-to-date frameworks..

So I guess we would need to build the frameworks ourself (easy with SDL, perhaps not so easy with the other packages, never tried this before).

Another option would be to bind statically with the libs (if supported). I think this would require a new xcode project without cmake? (I must confess I am not familiar with cmake and it is a little bit confusing)

BTW, I tried to make a build with the 10.5 SDK, but failed, because some functions names in the readline seem to be named different (on 10.6 it is _rl_complete_function, on 10.5 it is only complete_function). Did you compile your build on Leopard?

-------- Original-Nachricht --------
> Datum: Sun, 15 Jan 2012 19:01:38 +0100
> Von: Jerome Vernet <vernet.jerome@xxxxxxxxxx>

> An: hatari-devel@xxxxxxxxxxxxxxxxxxx
> Betreff: Re: [hatari-devel] Hatari 1.6.1 test build for Mac OS X

> I've just tested on my G4 MDD my ppc build( done on the MacBook): it
> fail, with an incompatible version of libpng. really a pain in... I
> need the frameworks version of libpng, libz and libreadline, as these
> dynamic library depend of the developper package installed or not...
> With the Framework embedded in the app, it will be ok !
>
> I don't know if portaudio is really needed ? Same for readline, as we
> do not use really the text debugger (well, you can, if you launch
> Hatari thru the commande line).
>
>
>
> Dyld Error Message:
>    Library not loaded: /usr/X11/lib/libpng12.0.dylib
>    Referenced from: /Volumes/Hatari/Hatari.app/Contents/MacOS/Hatari
>    Reason: Incompatible library version: Hatari requires version
> 47.0.0 or later, but libpng12.0.dylib provides version 45.0.0
>
> JV
>
> Le 15 janv. 12 à 16:56, Bob Carpenter a écrit :

>
> > @Stefan: When I use "otool", I see that libpng is referenced.
> > However, it refers to /usr/X11/lib so I am assuming it is not
> > included in the executable like the SDL framework. I have tried the
> > mono monitor switch several times without having it fail. However, I
> > agree with you that Jerome's executables are the best choice to host
> > at tuxfamily. For one, he has the only PPC build.
> >
> >
> > Bob C
> >
> > On Sun, Jan 15, 2012 at 9:01 AM, Eero Tamminen <oak@xxxxxxxxxxxxxx>
> > wrote:
> > Hi,
> >
> > On sunnuntai 15 tammikuu 2012, Bob Carpenter wrote:
> > > I had never tried the video capture before. It did seem to slow down
> > > Hatari significantly. I was trying to play Llamatron 2112 with the
> > video
> > > capture on and it was very sluggish. I was playing Hatari in a
> > window
> > > rather than full screen. However, the video file itself was full
> > speed.
> >
> > There are two things affecting its performance:
> >
> > * Video frame compression in PNG format
> >  - Faster CPU or SIMD accelerated[1] PNG library could help
> > * Disk write speed
> >  - Faster disk (RAID array?) would help
> >
> >
> >        - Eero
> >
> > [1] PNG compression isn't something that's easily SIMD accelerated
> >    unlike e.g. JPEG which is matrix based compression method.
> >
> >
> >
>