Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Status
The first post of this thread is a WikiPost and can be edited by anyone with the appropiate permissions. Your edits will be public.
I am not sure that’s needed. The only reason Macports is ever discussed here is because that’s a working way to build soft which is relatively easier than doing everything on its own. The issue is not “how to fix Macports for SL” but how to build soft for SL in an optimal way.

By and large I agree with the need to find how to get binaries built from source in the Snow Leopard on PowerPC environment, but this isn’t core to getting Snow Leopard itself functioning optimally on PowerPC Macs the way, say, testing system-level components in the builds and testing existing, previously built software in those environments are.
 
Does anyone know if X1900 XT is supported on 10A190?

I just checked Table 3 in the wikipost. It isn’t listed as a tested GPU.

Unless it’s a PCIe-based card, then chances are pretty good it won’t work in Build 10A190 or Build 10A96. If it is a PCIe-based card, the next step is to check whether 10.5.8 supports it. If it does, then it’s possible it might. After all, there’s both the Pixel Shader v. Compute Shader support question, as well as the OpenCL question — the latter being completely unavailable to PowerPC Macs.

Also, it might be worth reviewing posts #893–899 for a discussion about the X1900 XT family of cards, if you’re on the fence for buying one to try out as an experiment with Build 10A190.
 
  • Like
Reactions: barracuda156
I just checked Table 3 in the wikipost. It isn’t listed as a tested GPU.

Unless it’s a PCIe-based card, then chances are pretty good it won’t work in Build 10A190 or Build 10A96. If it is a PCIe-based card, the next step is to check whether 10.5.8 supports it. If it does, then it’s possible it might. After all, there’s both the Pixel Shader v. Compute Shader support question, as well as the OpenCL question — the latter being completely unavailable to PowerPC Macs.

Also, it might be worth reviewing posts #893–899 for a discussion about the X1900 XT family of cards, if you’re on the fence for buying one to try out as an experiment with Build 10A190.

I have checked Wikipost prior to asking here of course. Anyway, I realized that XT was released for Mac Pro, and I wanna avoid headache of flashing the card. I still have FX4500 lying around and no PCIe WinPC to flash it.
 
Here comes the Fresh Install (tm) dmg for 10A190.
[automerge]1588492590[/automerge]
I like the cat, and the original Macintosh ´84 team would have also I´m sure. As long as it has no C19-mask nowadays ;-)
Thanks @Edgecrusherr !

This is a complete thing, without any Intel parts removed, right? I am thinking to try installing it on my MacMini 2012. I have finally found out how to force 10.6 install and boot on it:

IMG_0776.JPG
 
  • Like
Reactions: Amethyst1
I am thinking to try installing it on my MacMini 2012.
I am not overly confident the hacked Ivy Bridge 10.6.8 kernel will work with that early beta. Lots of kexts may not load, and if the versions of the kernel and System.kext don’t match you’ll have weird errors like disk images not mounting IIRC.
In any case, please report your experience :)
 
Last edited:
I am not overly confident the hacked Ivy Bridge 10.6.8 kernel will work with that early beta. Lots of kexts may not load, and if the versions of the kernel and System.kext don’t match you’ll have weird errors like disk images not mounting IIRC.
In any case, please report your experience :)

That’s pretty likely, but since I made 10.6.8 work on my MacMini, I discarded the idea to buy another one just for that purpose.
 
That’s pretty likely, but since I made 10.6.8 work on my MacMini, I discarded the idea to buy another one just for that purpose.
If you only need 10A190 for the purpose of compiling stuff or testing, what about shoving it in a VM? You can spoof a Core(2)Duo CPU old enough for 10A190 to (hopefully) not fall over. No need to bother with hacked kernels ‘n’ stuff. :)
 
Last edited:
  • Like
Reactions: B S Magnet
If you only need 10A190 for the purpose of compiling stuff or testing, what about shoving it in a VM? You can spoof a Core(2)Duo CPU old enough for 10A190 to (hopefully) not fall over. No need to bother with hacked kernels ‘n’ stuff. :)

It will be horribly slow in emulation, I guess. Also I am not sure compilation works properly in such case (but I haven’t tried).
 
It will be horribly slow in emulation, I guess.
This isn’t emulation though. It’s virtualisation at near native speed. Spoofing an older CPU in a VM doesn’t imply fully emulating that CPU — just returning values OS X likes when it checks certain CPU registers.

I can confirm Tiger runs at insane speed on my 5th-gen Core i5 using that tutorial — and “naughty” VirtualBox doesn’t care if it’s a client or server build either.

;)
 
Last edited:
This isn’t emulation though. It’s virtualisation at near native speed. Spoofing an older CPU in a VM doesn’t imply fully emulating that CPU — just returning values OS X likes when it checks certain CPU registers.

I can confirm Tiger runs at insane speed on my 5th-gen Core i5 using that tutorial — and “naughty” VirtualBox doesn’t care if it’s a client or server build either.

;)

Thanks. What would you recommend to use for virtualization, software-wise?
 
Thanks. What would you recommend to use for virtualization, software-wise?
VirtualBox is free and doesn’t care about client vs. server edition.
Parallels or VMware aren’t free (AFAIK) and only allow Mac OS X Server to be virtualised if it’s older than Lion.
I’d try VirtualBox first, then VMware. I’ve no experience with Parallels.
 
Quick update on Xcode 3.2 Pre-release v. 1544 (from 10A222):

Apparently it works normally after replacing ld in /Developer/usr/bin (provided it is installed in default location).
At least I pulled an R GUI project into this volume together with R.framework, and the build by Xcode reports as successful.

In any case you will require ld and gnumake from Xcode 10A190 to make Xcode 10A222 and its compilers usable.
These files must be replaced with respective ones from 10A190:

Code:
/usr/bin/gnumake
/usr/bin/ld
/Developer/usr/bin/gnumake
/Developer/usr/bin/ld

P. S. If anyone has any success with later versions of Xcode, please update me.

I can confirm now that gcc compiler from 10A222 is perfectly functional. It is safe to install and use: I have replaced the earlier one on my main system.

Code:
36-225:~ svacchanda$ gcc -v
Using built-in specs.
Target: powerpc-apple-darwin10
Configured with: /var/tmp/gcc/gcc-5626~6/src/configure --disable-checking --enable-werror --prefix=/usr --mandir=/usr/share/man --enable-languages=c,objc,c++,obj-c++ --program-transform-name=/^[cg][^.-]*$/s/$/-4.2/ --with-slibdir=/usr/lib --build=i686-apple-darwin10 --with-gxx-include-dir=/usr/include/c++/4.2.1 --program-prefix= --host=powerpc-apple-darwin10 --target=powerpc-apple-darwin10
Thread model: posix
gcc version 4.2.1 (Apple Inc. build 5626)

Used it to rebuild isl, gcc6 and gcc7, all is good.

(Of course ld and gnumake have to be replaced after installation.)
 
  • Like
Reactions: G4fanboy
I wonder if it could possibly run on a Power Macintosh 9500 or the like with a speedy enough G4 upgrade card. Probably not, but it seems like that would be the best combo for a time-machine computer that runs every version of OS X from Rhapsody DR1 onward (not counting point releases, of course). Maybe if I ever win the lottery I'll try it just for fun...​
 
  • Like
Reactions: Amethyst1
What prevents them to boot, if you know?

AHCI — a predecessor to NVMe — is Intel-only, insofar as Macs go.

The introduction of AHCI support in OS X was intended for Intel Macs to take advantage of — which is why, despite the AppleAHCIPort and IOAHCIFamily kexts in Snow Leopard Builds 10A96, 10A190, and onward, only an Intel Mac is able to load them (as both are written as Intel only binaries). That support needs to be present at the Open Firmware or EFI level for the Mac to be able to boot from it, but Open Firmware was sort of left in the dust here.
 
I can confirm now that gcc compiler from 10A222 is perfectly functional. It is safe to install and use: I have replaced the earlier one on my main system.

Code:
36-225:~ svacchanda$ gcc -v
Using built-in specs.
Target: powerpc-apple-darwin10
Configured with: /var/tmp/gcc/gcc-5626~6/src/configure --disable-checking --enable-werror --prefix=/usr --mandir=/usr/share/man --enable-languages=c,objc,c++,obj-c++ --program-transform-name=/^[cg][^.-]*$/s/$/-4.2/ --with-slibdir=/usr/lib --build=i686-apple-darwin10 --with-gxx-include-dir=/usr/include/c++/4.2.1 --program-prefix= --host=powerpc-apple-darwin10 --target=powerpc-apple-darwin10
Thread model: posix
gcc version 4.2.1 (Apple Inc. build 5626)

Used it to rebuild isl, gcc6 and gcc7, all is good.

(Of course ld and gnumake have to be replaced after installation.)

Since I have a clone of my 10A190 system now, I will probably try in the next round to replace Xcode and Unix tools with those from 10A261 and see if that break down everything or not.

10A222 compilers are fully functional, you may consider adding that to Wikipost (which has rather misleading info that only Xcode from 10A96 works – that is verily incorrect, I never installed anything from 10A96 at all, and I assure that my Xcode installation works).

P. S. A side note re compilation: if something fails to build on 10.6 PPC by default methods, consider building against 10.5 SDK and deployment target. That will be a better solution – I think – than building on 10.5.8 and sticking it into 10.6. In Macports that can be like that: https://trac.macports.org/ticket/54332#comment:1 (amend the file temporarily to build specific port).
This turned to be a very simply solution for 2 ports I was struggling with for days.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.