I’ve been providing nightly builds of Emacs CVS HEAD for Mac OS X for a few months now. I’m going to keep building them, since the NS port is in bad shape and CVS should surpass 23.1 soon.
Moving forward, I hope to transition the nightlies to S3 and build them from Git instead of CVS.
For people who prefer the stable release, I built Cocoa Emacs 23.1. This is a vanilla build of the official release.
Lots. It’s significantly slower, less stable, and lacks features which were in the 22.x Carbon port, like fullscreen. It’s really not ready for prime time, and it sucks that people will get such a bad experience on OS X. What sucks even more is that new users will get a bad first impression. This could have been a chance to turn people on to Emacs and GNU software, but it’s more likely to alienate them instead.
The Carbon port was discontinued in the hopes that a single NextStep port could work on OS X and GNUStep. Unfortunately, relatively few people cared to work on it, and the result is that the stable Carbon port is dead and it’s replacement just isn’t as good. I also hear that the GNUstep builds are completely broken, though I haven’t personally verified such.
The politics of GNU bear some of the blame, too. Non-free platforms like Macintosh are second class citizens in their eyes, and there is very little desire to improve the experience of GNU software on them. The idea being that if they want things to work well, they should use GNU/Linux instead.
This is backwards, painfully stupid and wrong. If a user tries some of your software and it sucks, they’re going to stop using all of your software. Nobody is going to use crap software and think it’s a great idea to switch to a whole platform of it.
Discussion