Robert> So if I read the patch right, doing (require 'ibuffer) Robert> would be enough to get the keybinding changed? Urk. Please think carefully about this. Changing keybindings is always one of the most controversial aspects of patching. In particular, we've made a point in many cases of trying to ensure that user-visible changes are never activated by a require (the user normally does not know that they are require'ing a library). (Cf the "howcum JPEGs don't enter image-mode in 21.4" thread, etc.) Are people going to be shocked if something require's ibuffer behind their backs? I'm sure you think the shock will be pleasant; the point is some people will not agree. [This is not opposition to this particular patch, I haven't looked at it yet.] (Unofficial) Guide to Keybinding Changes 1. If you change any keybindings, say so in the ChangeLog. 2. Explain in the text why this change is a good thing ("ibuffer has been blessed..." is probably enough here, what's good for the GNUs is not always good for the Gnader, but usually). 3. Explain if the principle of no surprise has been violated, and why this patch complies with the principle of least surprise. 4. Think twice if it seems likely that more than about two carrots and a rutabaga will be tempted to reach for C-h k to figure out what's going on. Anything with a spinal cord should get it immediately. :-) -- Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Don't ask how you can "do" free software business; ask what your business can "do for" free software.

Robert> PS Why does being in the To: field make me feel like Robert> you're talking to me, when I *know* you're talking to Robert> John? Because you're the one who reduced the issue to two very quotable lines? In fact I wasn't really addressing John, because I haven't actually read his patch. Probably I should have phrased it "let's" and "we should". And trimmed the headers! ;-) -- Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Don't ask how you can "do" free software business; ask what your business can "do for" free software.

John Paul Wallington writes: Me no like. Switching bindings around just because a particular package file exists is not something I'd want (although it's fairly harmless here). Would it not be better to just move ibuffer.el into core lisp and remove the other implementation? I dislike having multiple implementations of the same thing. Robert --

jpw> Moving ibuffer.el out of packages and into core lisp is jpw> better for a future XEmacs, but then deprives existing jpw> XEmacsen. Hmm. If ibuffer.el doesn't use any non-public APIs, move it to xemacs-base, and the existing file in core can go away in future versions. (And fsf-compat et al should come out of that package, but that's another matter.) Then dump xemacs-base at build-time. This may not yet be practical, but it's the right thing to do in the long run. We should try to get anything that uses only public APIs out of the core. -- Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Don't ask how you can "do" free software business; ask what your business can "do for" free software.

John Paul Wallington <jpw(a)shootybangbang.com&gt; wrote: Idiot. Because of the :set keyword it ain't setq-able. So the answer would instead be: "umm. You have to customize it. No, setq will not do." Which would not placate them. So changing core is probably better. I just did this: --- lisp/menubar-items.el.orig Thu Jan 31 16:26:26 2002 +++ lisp/menubar-items.el Thu Jan 31 16:30:44 2002 ＠＠ -1442,7 +1442,7 ＠＠ ["Go To %_Previous Buffer" switch-to-other-buffer] ["Go To %_Buffer..." switch-to-buffer] "----" - ["%_List All Buffers" list-buffers] + ["%_List All Buffers" (if (fboundp 'ibuffer) (ibuffer) (list-buffers))] ["%_Delete Buffer" kill-this-buffer :suffix (if put-buffer-names-in-file-menu (buffer-name) "")] "----" And of course it works fine but C-h k (click on the menu item) prints: misc-user runs (if (fboundp (quote ibuffer)) (ibuffer) (list-buffers)) `if' is a built-in function... Ha-ha. Is it reasonable to just replace list-buffers in the menu with ibuffer? Should there be a list-all-buffers wrapper function with (if (fboundp 'ibuffer) (ibuffer) (list-buffers)) in it? -- John Paul Wallington