Posts

Post not yet marked as solved
10 Replies
Done, logged as FB13353555 with a simple test acse that has the Elements linbraryu projects (and the binary, so no need to build it yourself), and a very sikmple Xcode project that uses the library and fails right away. Or perhaps the new linker should be compatible with your current code generation and the fact that this doesn’t work is a bug. It’s hard to say without in-depth knowledge about how the linker works, which I don’t have )-: We’ll know more once the linker engineers have had a chance to look at your bug. 🤞🏼 as this is a bit of a breaking/show-stopper issue for us ;) thanx for your assistance!
Post not yet marked as solved
10 Replies
Okay, i'll file a bug. take it this means there is current;ly no documentation or spec available to share for what changed with regards to how Categories work, in order to work with the current/new linker?
Post not yet marked as solved
10 Replies
Hi Quinn, What compiler is that? The compiler is Elements (https://www.remobjects.com/elements). It uses LLVM as the back-end but does not use Clang, as it is not C language-based. The path forward for ld_prime issues is clear but, as this reproduces with ld64, I can’t send you down that path. i will double-check this to be sure. Any assistance would be appreciated.
Post not yet marked as solved
10 Replies
Hi Quinn, is there any documentation on "category class properties" at all? i am getting the same warning when linking a binary compiled with an (LLVM-based) compiler other than Xcode into an Xcode 15 project (and i get the reverse, when linking a Xcode 15 l,ibrary into an project with the other compiler). Also (probably not surprising), regular category methods declared in the code compiled with the other compiler don't work when linked into the Xcode app (even when called from inide the same binary), they fail with "unknown selector" at runtime. I am assuming someting has changed for how Categories are encoded on the binary level, but i can find zero coverage of this in the documentations for the Objective-C runtime or anywhere else. I'd very much like to be able to fix this issue with the compiler to restore comaptbility, so any pointers would be appreciated. FWIW, this happens whetehr usig the "classic" or new defauklt linker in Xcode. thanx, marc
Post marked as solved
2 Replies
https://developer.apple.com/documentation/apple_silicon/about_the_rosetta_translation_environment?language=objc#3616845 Determine Whether Your App Is Running as a Translated Binary On Apple silicon, a universal binary may run either natively or as a translated binary. The system runs the native version whenever possible, but the user might opt to run the code using Rosetta to support older plug-ins.  To programmatically determine when a process is running under Rosetta translation, call the sysctlbyname function with the sysctl.proc_translated flag, as shown in the following example. The example function returns the value 0 for a native process, 1 for a translated process, and -1 when an error occurs. int processIsTranslated() { 	 int ret = 0; 	 size_t size = sizeof(ret); 	 if (sysctlbyname("sysctl.proc_translated", &ret, &size, NULL, 0) == -1) 	 { 			if (errno == ENOENT) 				 return 0; 			return -1; 	 } 	 return ret; }
Post not yet marked as solved
6 Replies
Looks like 1000 is "Unknown". yay, very helpful :( public enum AuthenticationServices.ASAuthorizationError { case unknown = 1000 case canceled = 1001 case invalidResponse = 1002 case notHandled = 1003 case failed = 1004 }
Post not yet marked as solved
6 Replies
In my app (www DOT curacaoweatherapp DOT com — Jesus, you can't post links?), I see it on all devices, and from different user accounts. :(
Post not yet marked as solved
6 Replies
ps: i'm getting two conflicting messages: Credential State request returned with error: Error Domain=AKAuthenticationError Code=-7073 this is printed to the console and credentialState 2, error Domain=com.apple.AuthenticationServices.AuthorizationError Code=1000 "(null)" this is the error message i'm getting back in the callback (different codes? why?)
Post marked as solved
1 Replies
I don't see why they would not. the Universal/Fat Binary format does not seem to have changed, so as long as you have a valid x86_64 slice in your app, any "unknown" architectures should be ignored. Heck, you could include a PowerPC slice and run on 10.0, probably ;)
Post not yet marked as solved
3 Replies
1) The one remaining piece that you might be missing is the "Full Size Content View" flag on the containing window. You have to allow content to underlap / scroll beneath the titlebar region to be eligible for a full-height sidebar. Thanx. I have tried that, but when I set this, I get no toolbar *or* titlebar anymore at all, oddly. Also, what is strange, I randomly *sometimes* have my scrolling content from sidebar or main view peek thru the titlebar (not my custom bar below it, if course), as expected. But that only happens on rare occasions, and I can see not rhyme or reason to it. 99% of the time, my titlebar stays solid when I scroll the content beneath it. 2) This sounds like a good use case for NSTitlebarAccessoryViewController. You can install an accessory view controller on a window using the addTitlebarAccessoryViewController method. A bottom bar accessory is automatically fit to the width of the content area (i.e. excluding the full-height sidebar) in Big Sur, and it fills the full width of the window on previous versions of macOS. And, as you remembered, accessories live inside the titlebar blur material, so there's no need to provide a separate background. Cool. I actually came across these in slides from 2015 that I found on Google, yesterday. But one thing I'm wondering: I have two separate controls that form the bar in the sidebar vs the one in the main content area (much like Xcode, again). How would I handle this with NSTitlebarAccessoryViewController so that one stays the width of the sidebar and one the width of the rest? Would I just add the content one as NSTitlebarAccessoryViewController for Bug Sur, and leave the sidebar one as a child view of the sidebar itself? If so, how would I get it to adopt the seamless sidebar background? A third related thing I noticed: on Catalina and earlier, i *do* get the translucent sidebar ok (also just "most" of the time; sometimes it come stop gray and only becomes translucent after I go fullscreen (and stays ok when I got back out of fullscreen). But on Big Sur, my sidebar now is consistently solid white... thanx a lot form your help!