E-Maculation

Paul Says:

http://minivmac.sourceforge.net/

"Today's Development source snapshot is the first in the 3.2.x series. This snapshot includes a first draft of merging in Ross Martin's FPU emulation code. There's still a lot of work to do, but it already allows much more software for the Mac II to run without crashing.

One thing I changed is that it was originally using the native floating pointing abilities of the computer it ran on, which I'm not enthusiastic about, because this could cause the emulation to give different results on different computers, and because of the even worse problem of getting it compile on various compilers that provide different ways of accessing obscure floating point features. Instead, I've made it use SoftFloat (by John Hauser), which I found being used in the Bochs emulator, plus some extensions to SoftFloat in Bochs by Stanislav Shwartsman. Though it's slower, it should give consistent results with minimum trouble. (The native floating point option is still available in the code for anyone who wants to fiddle with it, but I don't intend to support it.) SoftFloat does require the compiler to support 64 bit integers, and I haven't updated the build system yet to adjust this for different compilers (It is currently hard wired to use "long long").

One fun thing the improved FPU emulation makes possible is that the final MPW development tools from Apple now run well enough within Mini vMac to successfully compile the 68000 version of Mini vMac. (Mini vMac doesn't actually run inside the Mac II emulation unless it is compiled without sound. The incomplete sound emulation now seems to be the main reason programs don't work in the Mac II emulation.) "

VC6 won't build the SoftFloat without errors. So I must change my build environment to VC8 or Dev-C++ to do further test.

yksoft1 wrote:VC6 won't build the SoftFloat without errors. So I must change my build environment to VC8 or Dev-C++ to do further test.

Dev-C++ did compile the new FPU-enabled Mini vMac IIx.




The new Mini vMac IIx now runs Stuffit 5.1.5 and Stuffit Expander 5.5, so sit5 Stuffit file could now be unstuffed inside Mini vMac IIx.



running Photoshop 2.5.1, QuickTime Player 4.0.3(VERY SLOW), Nisus 3.5.




Running Chinese and Japanese System 6.0.7 in color
(I can't enable color in US English 6.0.8 without crashing Mini vMac).



Download address for my new Windows builds:
http://www.mediafire.com/?lxtntyzj2mz

Hi,

i'm trying to build vMac on my own, but i see from your post that you are using some compilation options which are not present in vMac build guide ...
Do you have a doc explaining all the new switches and the new available machine types ( the -m switch).

thanks !
bye

enryfox wrote:Hi,

i'm trying to build vMac on my own, but i see from your post that you are using some compilation options which are not present in vMac build guide ...
Do you have a doc explaining all the new switches and the new available machine types ( the -m switch).

Have a look at the Mini vMac Variations page:
http://www.gryphel.com/c/minivmac/var/download.html

That's really cool yksoft. Thanks for posting the links and (especially) the screenshots.

Just change the following three lines in FPMATHEM.h to make the new Mini vMac II compiles in MSVC6 or VC8:
Line 99, 107 and 119
#if 0
to
#ifdef _MSC_VER

I also compiled Mini vMac II 091014 FPU in my Leopard and Xcode 2.5.


I hope Paul will support 16M or more memory and 32-bit clean mode in the final 3.2.0, as Stuffit Expander needs 6M+ memory to unstuff a Sit5 file.

Download
http://www.mediafire.com/?jhonjnwzdnn

Hi,

I compiled too mini vMac using Snow leopard and Xcode 3.2: it actually works, but I have to manually change the target SDK from 10.4 to 10.5 (maybe I forgot to install tiger SDK ... i do not remember exactly...). Mac plus is quite stable, but mac II is having some problems (but I think they are all known):

- sound does not work, better to compile without sound support to avoid system crash
- in system 6.08 changing from grey to colours crashes the emulation.
- overall the mac II simulation seems a bit slower than the mac plus ... i do not know if it is the real time clock mimicking the real speed of a mac II or the more complex simulation of 68020 + FPU.

I also downloaded system 7.01 installation from Apple websites and it is contained in a disk image which can only be mounted with system 7.01 ... kind of loophole !

thanks

enryfox wrote:Hi,

I compiled too mini vMac using Snow leopard and Xcode 3.2: it actually works, but I have to manually change the target SDK from 10.4 to 10.5 (maybe I forgot to install tiger SDK ... i do not remember exactly...). Mac plus is quite stable, but mac II is having some problems (but I think they are all known):

- sound does not work, better to compile without sound support to avoid system crash
- in system 6.08 changing from grey to colours crashes the emulation.
- overall the mac II simulation seems a bit slower than the mac plus ... i do not know if it is the real time clock mimicking the real speed of a mac II or the more complex simulation of 68020 + FPU.

I also downloaded system 7.01 installation from Apple websites and it is contained in a disk image which can only be mounted with system 7.01 ... kind of loophole !

thanks

Does XCode 3.x generate Panther-compatible binaries?

yksoft1 wrote: Does XCode 3.x generate Panther-compatible binaries?

Actually I do not know, i have an intel iMac and compiled vMac as an intel-only binary. I do not remember also if SDK for 10.4 and 1.3 are included in the Xcode installation: in my customised installation minimum SDK is 10.5.

I might try and install other SDK's, but i do not know if an intel mac can compile a PPC only binary.

bye

Snow leopard Xcode can compile binaries only from 10.4 up; 10.3 is not supported.

Given the incomplete sound emulation for Mac II, wouldn't it be better to compile vmac Mac II without sound support ? if sound support is enabled, every time an applications trys to reproduce a sound, vMac hangs. If sound is disabled, the menu bar will flash but emulation will continue unaltered.

thanks

enryfox wrote: Given the incomplete sound emulation for Mac II, wouldn't it be better to compile vmac Mac II without sound support ? if sound support is enabled, every time an applications trys to reproduce a sound, vMac hangs. If sound is disabled, the menu bar will flash but emulation will continue unaltered.

thanks

Some applications just use Gestalt Manager's 'snd ' selector to get the systems' sound hardware type without detecting it directly. You could try hacking the Gestalt Manager so that Mac OS would think the emulated system does not have Sound Manager installed.
In the beginning of MinivMac 3.1.x development, I thought Paul should implement emulation of an LC instead of a II or IIx, because LCs don't have FPU by default and have simpler sound subsystem. However Paul did not implement LC emulation even at 3.2.x.

Jsut wanted to say THANKS for the OSX compiles, YKSOFT1!! :D