Since I finally had to switch to RMSmacs, I find the behavior of display-buffer both incomprehensible and intolerable. It doesn't match my decades of muscle memory. I have tried tweaking the contents of display-buffer-fallback-action and various other things and still can't tell what the fuck it thinks it is doing. I can't characterize exactly what it's doing that is wrong, but... it is and it makes my typing go off the rails on the regular.
Please, is there someone out there who has managed to revert this absolute shitfuckery to have the behavior that it had from approximately 1986 through, let's say, 2002?
Actually I don't know what year RMSmacs screwed this pooch. I just know that XEmacs's behavior was traditional through 21.5 in 2007, at least.
(And no, before you suggest it without having tried it, just loading XEmacs's window.el into RMSmacs isn't gonna work.)
Wow, it's 2021.... not 1994
What a remarkable insight.
I wish. My transition to RMSmacs was a very painful time, the traumatic effects of which I still undergo, every time display-buffer is invoked for me, and I have to look at my 134k ~/.emacs.
As someone who's job it was to hack emacs, the depths of anger and disgust I feel when trying to do something in emacs now, is probably highly disproportionate. Since using Common Lisp for decades, elisp makes me cringe. I mostly type code in my own REPL now. window.el is some 10k lines of mad ass stuff that probably should be it's own window manager. I'm not sure whether to laugh or cry at window.el/window-xemacs.el. I would offer to try to fix it, but I've forgotten exactly what the traditional way was, except it didn't make me mad. I ended up writing multiple xemacs compatibility shims, but none related to display-buffer.
I can't quite characterize what's wrong either, except that seeing the "news" icon with a newspaper, and thinking of usenet, gives me feels on multiple levels. I mostly just trained myself not to do things that display-buffer, and to keep hitting C-x 1, and mostly one window per frame.
The bright side is, once switched, multiple-cursor-mode and org-mode profoundly changed what I can't imagine living without.
Welcome to the club. After over 20 years of Lemacs/XEmacs I switched back five years ago, and it was painful. I didn't try to adapt everything to make it like XEmacs, rather to get used to GNU Emacs again, and have mostly succeeded.
But it isn't the same. XEmacs felt slick and just how I liked it, while going back involved a lot of "oh noez", like going back in time to something, well, not as refined. But XEmacs support somehow became worse, and building it from source and fixing things didn't seem as fun any more as it did decades ago. And if now even you give up on it, it is literally the end of an era. It was good (most of the time) while it lasted.
I actually made the switch back in Sep 2019, and mostly I came to terms with it without too much pain, as demonstrated by the lack of blog posts here bitching about the process. It was largely a matter of putting the keybindings back to what I expected, and turning off the angry-fruit-salad color schemes. I had to rewrite and simplify beginning-of-defun and comint-next-prompt. The display-buffer thing is the only part that still drives me batty.
Not an emacs user, but even the docs for how it works very much sound insane to me.
C.
In case anyone else has the same problem as I do with the link above [1]:
[1] link fails to jump to the correct anchor in Vivaldi and possibly other Chromium-based browsers, probably because the page is so large and this is somehow a problem in 2021
Having configuration be via a list of conses of functions or lists of functions with alists of keywords is EXTREMELY ON BRAND for RMS's ideas of how data structures should work, to the point of self-parody, and this kind of shit is 60% of the reason that Lucid Emacs and XEmacs came to exist in the first place.
+1
Whatever your feelings about Emacs and
display-bufferare, please leave RMS out of this: he had nothing to do with the recent redesign and re-implementation of this functionality, nor even to its documentation.Calling Emacs "RMSmacs" is also very far from the reality, as anyone who can "git shortlog" will see.
And the prudent thing to do if you have issues with this functionality is to file a bug report (assuming that making Emacs more convenient for you is the intent), because doing that will tremendously increase the chances that someone will actually look into the issue.
Long, useless, vaguely insulting replies about nomenclature and hyphenation that ignore the real issues are also EXTREMELY ON BRAND.
Ladies and Gentlemen: The RMS Legacy.
Check this out: https://lists.gnu.org/archive/html/bug-gnu-emacs/2021-01/msg00302.html
I have three times tried to setup display-buffer-alist and all three gave up (the last one was a few days ago).
You think you made it work the way you intended but there's always one more thing that throws you off. Eventually you decide that you THOUGHT you had it under control, but in reality have no idea what will happen when a new window pops, so you give up the illusion of control and (setq display-buffer-alist nil).
Since by Emacs standards I'm a new user, I have no idea what the behaviour was until the mid 2000s. But if you can describe what you want I can try to come up with a starting config for it :)