macOS Tips
My 2¢ while using a MacBook Pro: Safari, the keyboard, an Ada toolchain on Apple silicon, and a ChatGPT app that wouldn't load.
Ada toolchain
Work in Progress, but the important tips are already included!
Build Toolchains
You'll need basic build tools such as gprbuild and gnatmake, and those come with this GCC build.
On Apple silicon you should use simonjwright/distributing-gcc.
Download the newest release whose name ends in aarch64 (not the -cross ones): a .pkg installer.
Follow the wiki. But basically you'll do
- Right Click the downloaded file to install
- Add the path
/opt/gcc-VERSION-aarch64/binto$PATH
Then, you can check the installation with $ which gprbuild. Output should show where it was installed (default: /opt/gcc-VERSION-aarch64/bin).
Test
You can type below under the filename of 'hello.adb'
with Ada.Text_IO;
use Ada.Text_IO;
procedure Hello is
begin
Put_Line ("Hello WORLD!");
end Hello;
And now you can
$ gcc -c hello.adb
$ gnatbind hello
$ gnatlink hello
$ ./hello
> Hello WORLD!
OR, simply
$ gnatmake hello.adb
$ ./hello
> Hello WORLD!
Build Alire, If you want
Easier today: Alire publishes a ready-made alr for Apple silicon (alr-VERSION-bin-aarch64-macos.zip on its releases page), and alr toolchain --select can also install a GNAT for Apple silicon itself. Building it yourself still works:
You can clone alire-project/alire.
(You may want to switch to a release branch, e.g. $ git switch release/2.1)
And then, since zsh is macOS's default shell (bash is there too, but an old 3.2), run $ ALIRE_OS=macos gprbuild -j0 -p -P alr_env.
Lastly, your alr binary is generated under the git repository's bin directory. So you have to edit the path.
(In the future, when a new release comes out, you can simply switch to the release branch and run alr build)
MOST IMPORTANTLY, You should run $ alr toolchain --select and select gnat version gnat_external=VERSION [Detected at /opt/gcc-VERSION-aarch64/bin/gnat] one, which we installed above. Also, for the gprbuild, gprbuild=VERSION [Detected at /opt/gcc-14.1.0-aarch64/bin/gprbuild].
Test
Move to the desired dev directory, and run
$ alr get hello
$ cd hello[tab]
$ alr run
> Hello, World!
ChatGPT app stuck on the logo
If your ChatGPT desktop app on macOS opens to nothing but the logo in the center—and never loads—you’re not alone. I ran into this issue myself, and the usual fixes didn’t work at first.
Here’s what I tried, what failed, and the solution that finally fixed it.
The Problem
After launching the ChatGPT macOS app:
- The window opens normally
- Only the ChatGPT logo appears
- The app never loads the interface
- No error messages are shown
What Didn’t Work
Before finding the solution, I tried the obvious fixes:
- Deleting and reinstalling the app
- Clearing cache files
Neither of these solved the issue.
The Fix (That Actually Worked)
The problem turned out to be leftover app data that wasn’t removed during a standard uninstall.
Step-by-step solution:
- Quit the ChatGPT app completely
- Open Finder → press
Cmd + Shift + G - Enter the following paths one by one and delete the folders/files if they exist:
~/Library/Application Support/OpenAI
~/Library/Logs/OpenAI
~/Library/Preferences/com.openai.chatgpt.plist
- (In my case, simply re-starting ChatGPT app worked)
- (Optional) Restart your Mac
- (Optional) Reinstall the ChatGPT app
Why This Works
A normal reinstall doesn’t fully reset the app. Some configuration or corrupted data remains in your user Library folder, which can prevent the app from loading properly.
By removing these files, you’re forcing the app to start fresh.
If It Still Doesn’t Work
Here are a few additional things to check:
- Disable VPNs or network filtering apps
- Try a different network (e.g., mobile hotspot)
- Check macOS privacy/security settings
- Test with a different macOS user account
Final Thoughts
This issue is frustrating because the app gives no error message—it just sits there. But the fix is actually simple once you know where to look.
Hopefully this saves you some time.