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.
Hi everybody!

View attachment 848578

Acceleration achieved on DP4 with two NOPs to SkyLightOriginal: I saw loginwindow crashes, found "assertion failed" messages in DiagnosticReports, and... uh... ummm... patched out the assertions... Hey, it works!

As usual, I'll edit in details in a moment once I get over my excitement.

Edit: Alright, here you go:
There are no new symbols in either SkyLight or CoreDisplay. However, the function SLSReenableUpdateTokenWithSeed was bombing with "Assertion failed: (wid != 0)" and "Assertion failed: (sid != 0)". So I just NOP'd the places where it jumps to those messages:
Code:
0x787c6
0x787cf

This is, of course, incredibly dumb, but one thing that lends some credence to my solution is that the new function looks mostly unchanged (from a quick skim), but it does not have those assertions anymore. So, it is possible that a value of zero for those IDs is no longer indicative of a problem. It's also quite possible that I'm just masking the real issue...

If anybody who actually understands assembler wants to take a quick poke around (looking at you, @pkouame and @testheit) and confirm or refute my suspicion (and maybe find a better solution), that would be great :)

Latest wrapped frameworks are attached as always. (The only change since my Dock/login fixes is the SkyLightOriginal modification, but I'll put the full frameworks here for convenience.)

As for changes in DP4, a quick list of things I've seen so far:
- The IOHIDFamily patches don't seem to be necessary anymore. I'll edit to confirm this after a couple more timeout-free reboots, though. Edit: seems stable. @dosdude1 and @0403979, you may want to update your patchers.
- There are now notifications confirming whether you want to allow an app to notify you (I think this is new?)
View attachment 848583
Edit: There are also new toggle switches for this in System Preferences:
View attachment 848589
- Apple seemingly followed my lead (as with Xcode!) and fixed the Siri crash by disabling the "flames view" on non-Metal systems:
View attachment 848584
(If you're reading this, thank you Apple engineers!)
- There's an empty "Legacy Desktops" folder in System Preferences. No idea what it does. Maybe it's for those who installed over a previous OS?
View attachment 848585
- Edit: One more change I've noticed -- last edit, I swear... Our old friend Safari is back in Launchpad! (I'm very happy about this one.)
View attachment 848588

My apologies for the long (as usual) post. Please report back as always! And have a good night :)

Ah then in this beta4 not the wrapper but the SkyLightOriginal needed a new binary patch, amazing quick work again!

I confirm too that IOHIDFamily beta4 doesn't produce more boot delays or kext_stall .

About Siri.app , it's weird, as happened for Xcode 11 beta2, working "Metal-apps" on non-metal machines, I guess apple macos developers made that to allow using them when booting Catalina in safe mode (-x) where the Metal driver is disabled by default, otherwise I don't see other reasons, or most probably some are really watching this thread and waiting for your Photos.app non-metal fix.

edit:
also Notes.app has evolved.
 

Attachments

  • notes beta4.png
    notes beta4.png
    94.5 KB · Views: 191
Last edited:
Hi everybody!

View attachment 848578

Acceleration achieved on DP4 with two NOPs to SkyLightOriginal: I saw loginwindow crashes, found "assertion failed" messages in DiagnosticReports, and... uh... ummm... patched out the assertions... Hey, it works!

As usual, I'll edit in details in a moment once I get over my excitement.

Edit: Alright, here you go:
There are no new symbols in either SkyLight or CoreDisplay. However, the function SLSReenableUpdateTokenWithSeed was bombing with "Assertion failed: (wid != 0)" and "Assertion failed: (sid != 0)". So I just NOP'd the places where it jumps to those messages:
Code:
0x787c6
0x787cf

This is, of course, incredibly dumb, but one thing that lends some credence to my solution is that the new function looks mostly unchanged (from a quick skim), but it does not have those assertions anymore. So, it is possible that a value of zero for those IDs is no longer indicative of a problem. It's also quite possible that I'm just masking the real issue...

If anybody who actually understands assembler wants to take a quick poke around (looking at you, @pkouame and @testheit) and confirm or refute my suspicion (and maybe find a better solution), that would be great :)

Latest wrapped frameworks are attached as always. (The only change since my Dock/login fixes is the SkyLightOriginal modification, but I'll put the full frameworks here for convenience.)

As for changes in DP4, a quick list of things I've seen so far:
- The IOHIDFamily patches don't seem to be necessary anymore. I'll edit to confirm this after a couple more timeout-free reboots, though. Edit: seems stable. @dosdude1 and @0403979, you may want to update your patchers.
- There are now notifications confirming whether you want to allow an app to notify you (I think this is new?)
View attachment 848583
Edit: There are also new toggle switches for this in System Preferences:
View attachment 848589
- Apple seemingly followed my lead (as with Xcode!) and fixed the Siri crash by disabling the "flames view" on non-Metal systems:
View attachment 848584
(If you're reading this, thank you Apple engineers!)
- There's an empty "Legacy Desktops" folder in System Preferences. No idea what it does. Maybe it's for those who installed over a previous OS?
View attachment 848585
- Edit: One more change I've noticed -- last edit, I swear... Our old friend Safari is back in Launchpad! (I'm very happy about this one.)
View attachment 848588

My apologies for the long (as usual) post. Please report back as always! And have a good night :)
Can I have source code please?
 
successfully installed macOS Catalina Dev beta 4 thank you Drift does not work on my iMac 9,1 Photo's crashes got about to 91% restoring photo library but then it crashed all in all running smoothly did get a few like Office update asking for permission. I think the legacy desktop is if you downloaded pictures or got a hold of a Catalina Heic :)
 

Attachments

  • Screen Shot 2019-07-17 at 2.33.29 AM.png
    Screen Shot 2019-07-17 at 2.33.29 AM.png
    2.3 MB · Views: 213
  • Screen Shot 2019-07-17 at 2.31.58 AM.png
    Screen Shot 2019-07-17 at 2.31.58 AM.png
    2.3 MB · Views: 217
  • Screen Shot 2019-07-17 at 2.46.42 AM.png
    Screen Shot 2019-07-17 at 2.46.42 AM.png
    2.5 MB · Views: 190
@testheit your SkyLight hybrid "light mode" offset of course is unchanged (and works) because the SkyLight mojave base is essentially the same, only thing when you will hex-patch it again, if you can fork the "FAT 32bit coding" from OpenGL, SkyLight and CoreDisplay, since for those who use the verbose booting, the many apfs_spec_read_wrapper instances (even if not severe) are a bit annoying and don't allow to read/debug correctly the verbose lines.

Code:
file /System/Library/PrivateFrameworks/SkyLight.framework/Versions/A/SkyLight
/System/Library/PrivateFrameworks/SkyLight.framework/Versions/A/SkyLight: Mach-O universal binary with 2 architectures: [x86_64:Mach-O 64-bit dynamically linked shared library x86_64] [i386]
/System/Library/PrivateFrameworks/SkyLight.framework/Versions/A/SkyLight (for architecture x86_64):    Mach-O 64-bit dynamically linked shared library x86_64
/System/Library/PrivateFrameworks/SkyLight.framework/Versions/A/SkyLight (for architecture i386):    Mach-O dynamically linked shared library i386

file /System/Library/Frameworks/CoreDisplay.framework/Versions/A/CoreDisplay
/System/Library/Frameworks/CoreDisplay.framework/Versions/A/CoreDisplay: Mach-O universal binary with 2 architectures: [x86_64:Mach-O 64-bit dynamically linked shared library x86_64] [i386]
/System/Library/Frameworks/CoreDisplay.framework/Versions/A/CoreDisplay (for architecture x86_64):    Mach-O 64-bit dynamically linked shared library x86_64
/System/Library/Frameworks/CoreDisplay.framework/Versions/A/CoreDisplay (for architecture i386):    Mach-O dynamically linked shared library i386

file /System/Library/Frameworks/OpenGL.framework/Versions/A/OpenGL
/System/Library/Frameworks/OpenGL.framework/Versions/A/OpenGL: Mach-O universal binary with 2 architectures: [x86_64:Mach-O 64-bit dynamically linked shared library x86_64] [i386:Mach-O dynamically linked shared library i386]
/System/Library/Frameworks/OpenGL.framework/Versions/A/OpenGL (for architecture x86_64):    Mach-O 64-bit dynamically linked shared library x86_64
/System/Library/Frameworks/OpenGL.framework/Versions/A/OpenGL (for architecture i386):    Mach-O dynamically linked shared library i386
 

Attachments

  • Screenshot 2019-07-17 at 14.36.37.png
    Screenshot 2019-07-17 at 14.36.37.png
    1 MB · Views: 201
Last edited:
  • Like
Reactions: TimothyR734
He just written, the new non-metal beta4 fix it's an assembly code NOP on 10.14.5 Skylight unix exec between that offset interval to bind a new catalina function (in the mojave skylight) that crashed, while all his wrappers source code are unchanged.
The CoreDisplay wrapped framework also seems to be different. I ran diff on it and it showed changes.
 
The CoreDisplay wrapped framework also seems to be different. I ran diff on it and it showed changes.

Since the macOS frameworks need to be codesigned to load and work I guess each one who will re-sign the frameworks will give them a slightly binary-modification inside their unix exec, that's why "diff" produces differences versus the previous CoreDisplay.
 
  • Like
Reactions: TimothyR734
1.0b11 is out now and working.
Can confirm that for cMP 4,1/5,1. Installed DP4 over DP3, no hickups, finished in ~30min.

cMP & Cat DP4.png

There's a cosmetic problem here with fonts displayed system-wide as in pic above, quite pale and frayed.

Edit, found the culprit. An option in TinkerTool was still active »...force font smoothing the old way...« – incompatible.
 
Last edited:
Catalina PB3 runs great so far on my 4,1(5,1), no issues so far. Installed it via MBP 2015 on temporarily removed ssd from the cMP. Just the no compat check nvram added (actually added back when DP1 came out).

A shame Apple blocks this machine from legitimately getting Catalina:confused:
 

Attachments

  • Cat PB3.png
    Cat PB3.png
    3.5 MB · Views: 195
Here is the latest LightMode fix (except top menubar)

@dosdude1 maybe you could strip the binaries in your patcher down to x86_64 only. It would shrink the size of your patcher and would remove all these annoying log messages in verbose boot :)
I’m gonna avoid that for mine since it supports other versions apart from Catalina. But, I could run lipo thin on all of them for the 10.15 release zip only.
 
I’m gonna avoid that for mine since it supports other versions apart from Catalina. But, I could run lipo thin on all of them for the 10.15 release zip only.

"lipo -thin x86_64" (or "lipo -remove i386") just these unix exec for Catalina, all the others are already "64 bit only" after a Catalina installation: OpenGL (from Mojave 10.14.3), SkyLightOriginal and CoreDisplayOriginal (from @ASentientBot patches respectively 10.14.5 and 10.14.4 ).
 
Last edited:
Tried to upgrade an existing install from the USB stick. It went to the second part rebooting, but then could not complete install due to an error unmounting the disk. This is on a Mac Pro 5,1.
Has anybody else seen this?
Have any of you been successful with an upgrade?

In any case, thanks Dosdude!
 
i have yet 19A501i is this public beta or develioper beta?? my mac doesnt found an update :(
i have apfs drive..
You have to use the Catalina Patcher to download the latest Cat beta then create a usb patcher then reboot and install :)
[doublepost=1563377226][/doublepost]
Tried to upgrade an existing install from the USB stick. It went to the second part rebooting, but then could not complete install due to an error unmounting the disk. This is on a Mac Pro 5,1.
Has anybody else seen this?
Have any of you been successful with an upgrade?

In any case, thanks Dosdude!
make sure you are formatting your usb extended journaled before you make a new Cat patcher :)
 
You have to use the Catalina Patcher to download the latest Cat beta then create a usb patcher then reboot and install :)
[doublepost=1563377226][/doublepost]
make sure you are formatting your usb extended journaled before you make a new Cat patcher :)
Not the issue. That part is working fine. It does the install and goes to the second phase booting from the installed part to finish the installation - that's when it cannot dismount the disk.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.