In case you’re unfamiliar with this slow-motion trainwreck, Apple switched from DVI display connectors to Mini DisplayPort connectors on all their new laptops. Nobody makes monitors with DisplayPort yet, so you need an adapter to use your existing DVI monitor. Which is fine so long as you don’t have a large (30”) monitor, which requires a different (Dual-Link DVI) adapter, which Apple hasn’t been able to ship. In other words: If you had a 30” monitor and bought one of their new laptops, you’ve been staring at your reflection in the tiny glossy screen instead of using your normal monitor.

Steve really must be sick. I can’t see how he’d allow garbage like this to ship.

I haven’t dug in further, but I think it’s taking the hostname from the IP assigned via DHCP. I guess it doesn’t randomly change any more, but it bugs me to no end that it changes at all.

While I’m a bit fuzzy on the specific version, this feature was removed around v2.0, replaced with the ability to play videos back when the handset was in a vertical orientation. While I prefer the old behavior, I see no reason why you shouldn’t be able to have it both ways. If the audio simply continued playing when you locked the phone during Podcast video playback, that would be fine.

I’ve found two workarounds, neither of which I find satisfying.

1. Don’t lock the phone, but stick it in my pocket.
2. Download the audio-only version of the podcast as well as the video.

The first suffers from two severe deficiencies.

1. The proximity sensor is ignored; the screen doesn’t shut off, eating your battery life.
2. Should you (or something else in your pocket) activate the Home button, playback stops.

The second is also unworkable.

1. Not all podcasts offer an audio and video version.
2. You have to make note of your position and seek to it when switching. Playback resumes from where you left off with the old behavior.
3. Disk space and bandwidth is wasted.

I’ve enjoyed this feature since I found it on my 5th generation iPod, and I’m very disappointed to see it get axed from the iPhone.

Example: I wanted to install ack, since I’m sick of getting .svn directories in my grep results. I see it’s in MacPorts, so I happily run sudo port install p5-app-ack.

Unfortunately, this begins installing the perl5.8 port. Why is this thing compiling Perl? OS X 10.5 ships with Perl 5.8. I need another copy… why? Does this seem terrible to anyone else? Ack is all of 23kb – do I really need a whole new Perl to run it?

Fortunately, OS X also ships with CPAN, which happily installed ack for me.

Re 1.3: these are Emacs limitations; we’re working on them.

Perhaps it wasn’t such a good idea to put out a press release for a major feature if there are serious known limitations. It’s a cool feature, but it’s not release quality.

Re 1.4: Other modes do NOT use Command-{/} because Command (mapped to A- in Aquamacs) is not a traditional Emacs modifier key. If you do something unwise like mapping it manually to Meta, then of course you’re on your own…

I take issue with the characterization that using command as meta is “unwise.” I am, after all, using the mechanism you yourself put in place (M-x customize-variable RET mac-command-modifier RET), which contains no mention of it being “unwise” or otherwise problematic. But I grant that this is more of a problem with my setup. I learned early on that Meta is the key to the left of space, whatever the label on it happens to be, and I find that much more palatable then stretching for the smaller Option/Alt key.

Re 3: You need a Lisp interpreter installed if you want to use common lisp. The correct way to run Emacs Lisp stuff is M-x lisp-interaction-mode or M-x emacs-lisp-mode.

Emacs is a lisp interpreter. I don’t want to use CL, I want to write, tweak, and eval Emacs lisp in a buffer. This may be my inner curmudgeon at work; lisp-mode used to do this, but perhaps that has changed to emacs-lisp-mode.

Re 4: Have you reported the bug? It won’t get fixed if it isn’t reported, and this is a problem in GNU Emacs (and not just Aquamacs specifically).

I sent a near-identical list of issues directly to David. I’m unable to participate in the Aquamacs mailing lists (including aquamacs-bugs) because any email is bounced due to my email domain lacking a “postmaster” address. I have no control over how my email is run, so this isn’t something I can fix. David said that “bugs should be non-restrictive… I’m looking into it.” Thus far, it has not been changed.

1. 1.4 introduced tabs via modification of tabbar.el. This is something I’ve desired for some time, but the implementation is lacking.
   1. If you close the only buffer in a frame, the frame closes. This is a change from 1.3 and below, which would switch you back to the last buffer you were visiting. I end up closing frames much more than I’d like to.
   2. It’s hard to manage which tabs show up in which frames. Any time you switch to a buffer, a tab gets created in that frame. There is no simple way to remove the tab from the frame at that point. You must kill the buffer, or delete the frame.
   3. Many modes break the tab bar. ERC and psvn are two common ones.
   4. Lack of standard bindings. Aquamacs is supposed to be like OS X, but they didn’t use the standard ⌘{ and ⌘} keys. Many modes already use these, so they don’t work consistently.
   5. It breaks find-alternate-file (C-x C-v). When you call it, the file opens, but it’s buffer is hidden. You have to switch back to it, negating the speed improvement you get from that binding in the first place. It’s particularly irritating when you use to correct the load of a wrong file. E.g. if you have: Foo.php and Foo/Bar.php, try to load Foo.php, but accidentally open a dired of Foo/.

   Yeah, I’m ready to give up on tabs.

2. Paths with spaces in them break dirtrack. For example, if you change to ~/Library/Preferences/Aquamacs Emacs/, dirtrack won’t follow you there.
3. When you hit M-x lisp-mode RET Aquamacs uses slime-mode, which doesn’t work right. Maybe it needs to be set up? I don’t know, but I do know that I can’t use C-x C-e to eval expressions; it just gives me a “Not connected” error.
4. find-file is broken with Tramp paths. Normally, if you start typing a new top-level path, (e.g. anything starting with / or ~), the existing path in the minibuffer gets grayed out, and ignored when you hit RET. But if you’re looking at a Tramp path, this breaks. It’s been broken since 1.4, and doesn’t seem to be getting fixed.