When will the 30.1 tag be picked up?

[shipmints]

It's been in the github mirror for 20 hours already!!! Just curious...

[jimeh]

Thanks for the poke, and sorry for the delay. I've fired off the release build for 30.1 here: https://github.com/jimeh/emacs-builds/actions/runs/13517574012

It should be done and released in about an hour.

I haven't yet gotten around to putting together an automated way of picking up new upstream tags which are relevant for release builds. I'll try and sort something out before the next major release.

In the past it hasn't been a major issue as I'd normally see the news and fire off a build within hours. This time I've been unusually busy with other things though and missed the it till now.

[shipmints]

Aha, I assumed that tag detection was automated. 30.2 with regression fixes is likely just around the corner. Thank you.

[shipmints]

Downloaded and testing. So far, so good.

[shipmints]

Just tested this on a friend's M3 MAX laptop and got this

./Emacs -Q
2025-02-26 16:19:24.032 Emacs[4893:90989] Failed to initialize color list unarchiver: Error Domain=NSCocoaErrorDomain Code=4864 "*** -[NSKeyedUnarchiver _initForReadingFromData:error:throwLegacyExceptions:]: non-keyed archive cannot be decoded by NSKeyedUnarchiver" UserInfo={NSDebugDescription=*** -[NSKeyedUnarchiver _initForReadingFromData:error:throwLegacyExceptions:]: non-keyed archive cannot be decoded by NSKeyedUnarchiver}

And it just hangs. Weirdly, it does start from the launcher. But then this...

⛔ Warning (initialization): An error occurred while loading ‘/Users/shipmints/.emacs.d/init.el’:

Native compiler error: (lambda (arg3 &optional) (let ((f #'yes-or-no-p)) (funcall f arg3))), Compiling /Users/shipmints/.emacs.d/eln-cache/30_1-b5cf085a/subr--trampoline-7965732d6f722d6e6f2d70_yes_or_no_p_0.eln...
ld: library 'System' not found
libgccjit.so: error: error invoking gcc driver
Internal native compiler error: "failed to compile", "/Users/shipmints/.emacs.d/eln-cache/30_1-b5cf085a/subr--trampoline-7965732d6f722d6e6f2d70_yes_or_no_p_0.eln", "error invoking gcc driver"

Error: native-ice ("failed to compile" "/Users/shipmints/.emacs.d/eln-cache/30_1-b5cf085a/subr--trampoline-7965732d6f722d6e6f2d70_yes_or_no_p_0.eln" "error invoking gcc driver")
  comp--compile-ctxt-to-file0("/Users/shipmints/.emacs.d/eln-cache/30_1-b5cf085a/subr--trampoline-7965732d6f722d6e6f2d70_yes_or_no_p_0.eln")
  comp--compile-ctxt-to-file("/Users/shipmints/.emacs.d/eln-cache/30_1-b5cf085a/subr--trampoline-7965732d6f722d6e6f2d70_yes_or_no_p_0.eln")
  comp--final1()
  load-with-code-conversion("/private/var/tmp/emacs-int-comp-subr--trampoline-7965732d6f722d6e6f2d70_yes_or_no_p_0-MPdcdG.el" "/private/var/tmp/emacs-int-comp-subr--trampoline-7965732d6f722d6e6f2d70_yes_or_no_p_0-MPdcdG.el" nil t)
  command-line-1(("-l" "/var/tmp/emacs-int-comp-subr--trampoline-7965732d6f722d6e6f2d70_yes_or_no_p_0-MPdcdG.el"))
  command-line()
  normal-top-level()


To ensure normal operation, you should investigate and remove the
cause of the error in your initialization file.  Start Emacs with
the ‘--debug-init’ option to view a complete error backtrace.
⛔ Warning (native-compiler): libgccjit.so: error: error invoking gcc driver
⛔ Warning (native-compiler): /Users/shipmints/.emacs.d/elpa/exec-path-from-shell-20220826.1535/exec-path-from-shell.el: Error Internal native compiler error: "failed to compile", "/Users/shipmints/.emacs.d/eln-cache/30_1-b5cf085a/exec-path-from-shell-d7ad226c-0ff42c62.eln", "error invoking gcc driver"

[jimeh]

Interesting, though I'm afraid I can't reproduce either of those issues on my own machine, which is also a M3 Max. I tested both on my host OS which is macOS 15, and also a fresh macOS 12 VM, and native-comp and such was working as it should.

A bit of searching around for the ld: library 'System' not found error seems to point at a possible issue with Xcode or Xcode Command Line Tools. So I would suggest checking if they're installed properly and working. Possibly also check the $SDKPATH environment variable, if it's pointing to some invalid path, that might interfere with things from what I've gathered in my search.

[shipmints]

I'll look at his set up more tomorrow. I have to "screen share" so it's a bit of an arrangement.

Any idea what would cause the inconsistency between manually running the Emacs binary and launching via command-space launcher?

[jimeh]

The only stuff I've found relating to the first error goes back to Emacs 26 and macOS Mojave, which is interestingly weird.

https://stackoverflow.com/questions/52521587/emacs-error-failed-to-initialize-color-list-unarchiver-when-i-call-it-in-the-t

Removing ~/Library/Colors/Emacs.clr is supposed to fix it. In your case though, no clue if that will actually help as the fix is from 2018.

And I can't say I know why it only happens in the terminal. My best guess is that there might be some custom shell setup with env vars that causes this weirdness.

[shipmints]

Right, I forgot about that issue--I had that years ago. Thanks for googling. I'm not surprised if it still exists because Apple.

I'll try that with him later and if that works, then boom. I have a feeling it's left over from his Emacs 28 that probably used the old NS API and left behind this cruft. But, this was claimed to have been fixed before 28.

Worth adding to a FAQ if it works.

[shipmints]

My friend deleted the color file, and reinstalled xcode and the issue launching from Spotlight disappeared. I still haven't heard about command-line launching.