Bare Metal (The Emacs Essay)

I've got another one for macOS. It's not as significant as what you listed, but it was extremely annoying for me.

When using homebrew, the default package for emacs is version 30. However, that version on macOS was compiled with an outdated tree-sitter version (v14). If one tries to install a tree-sitter parser for a particular language it will not work with emacs 29 or 30 on macOS (or at least not that I could figure out). I tried the D12Frosted version and some of the alternatives for macOS (I forget the names).

The current tree-sitter project is on v15+, so in order for it to work, one has to go through the commit histories for each language and find when the tree sitter's parser version was last compatible with v14. To my knowledge, this was not clearly listed in all the git commits in a nice clean manner.

One alleged workaround was to install everything via RedHat OS and move the installed parsers over to macOS since their version of emacs and the tree-sitter parsers are still compatible. However, I was not interested in doing this.

Interestingly enough, Neovim does not have this issue on macOS to my knowledge, but I could be wrong.

Now, this is not entirely an emacs issue, but it does introduce an interesting issue. If emacs has to be compiled with a particular tree-sitter version, then it makes things quite difficult to maintain. The maintainers apparently are discussing various options on how to handle this going forward, but there isn't a clear way to work around this.

Now, I do believe if one builds emacs from source, he or she can work around this issue to some degree, but I was also not interesting in doing this because one loses out on some of the nice features that the macOS focused emacs versions come with. I also did not want to install all the dependencies required to build from source since those dependencies are not something I have any other use for other than building emacs.

So, I just gave up on the tree-sitter. Though, I think emacs 31 has this fixed until it happens again, but I also encountered a lot of other bugs when trying other use it since it's a prerelease version anyway.

This was all months ago, so maybe things have changed. I haven't kept up. I can live without tree-sitter for the time being.

Now you mention it, come to think of it, I'm actually unemployed at the moment. We can't both be wrong!

My thoughts too. I've never had any real problems running it on Windows myself, but it doesn't feel like Mac users are any worse off.

I've been using Mac Emacs since about 2010, so perhaps I dodged some of the worst stuff, and it's always felt - for good or for ill - very much the same as using it on Windows or Linux/X-Windows, both of which I've been using since 2005 or so.

I've always used the standard GNU version rather than any specific Mac port. Over the years I'll have done some mix of downloading prebuilt binaries, building it from source, or getting the MacPorts version.

Perhaps I have missed out on some Fancy Extra Mac Stuff, the sort of thing that would come as standard with any project that took the platform seriously, but it's never really felt like a problem. And indeed, I figure if I'm happy enough with it running on X-Windows and on Windows, why wouldn't I be just as happy with exactly the same thing on macOS too? A consistent (or near enough) cross-platform experience is part of why I use Emacs anyway.

Is he bigger on the inside? Does he represent a bet that didn’t pay off, on a future that will never, ever be allowed to come? Does he embody the charming but mistaken belief that creating tools to make people freer will begin a movement toward freedom?

2) I have eshell bound to a keybind, but I've never use it. I prefer shell-mode and shell-command. They make it easy to use cli utilities. TUI is something that I find myself no longer needing. And I've become so accustomed with the cli that the only two I'm using in a terminal is `less` and `top`.

3) I think the best way is to find some vanilla base config that will smooth out the rough parts, then, once you understand the internal concepts, tweak them to your liking. It's certainly a long term plan, but the pro is not having to wait on "features" from another company or group.