Mastering EmacsThe Challenge
Jon over at Irreal’s Emacs blog posted an interesting solution to a challenge raised by Xah Lee:
How do you convert a string like this
37°26′36.42″N 06°15′14.28″Winto a decimal answer like this37.44345 -6.25396.
First off, I’m not sure Xah’s example answer is entirely correct; my understanding of latitude and longitude is limited to what I can google, and if I type the original degrees, minutes and seconds into this tool by the U.S. FCC it returns 37.44345 6.25396.
Anyway, on with the challenge. Jon’s solution is very interesting, but it got me thinking: surely Emacs has the facility in place to do this already? It’s Emacs, right? Right.
Fun with Emacs Calc
Emacs is equipped with a really, really awesome and spiffy RPN calculator (HP calculator fans, rejoice) capable of manifold things like algebraic and symbolic manipulation; matrix manipulation; conversion routines and much, much more. It’s truly wonderful but really complex, but it does come with a really nice Info manual (type C-h i g (calc) and check it out.) It’s a shame so few people know about its potential, as it’s basically a much simpler version of Mathematica, or even Wolfram Alpha (arguably you’ll have as much trouble telling Calc what you want as you would Wolfram Alpha…)
Anyway, I figured the Emacs calculator would have a facility in place for converting Deg-Min-Sec to decimal form, and sure enough, it does.
To try it out, type C-x * * and the calculator will open. Two windows will appear: the calculator mode and the trail containing a trail — a history — of commands and actions. The first thing we need to do is switch the calculator to “HMS” mode so we can try it out. To do this, type m h in the mode (the left) window and the modeline will change and say something like Calc: 12 Hms. The 12 is the floating point precision.
Next, type in the expression, replacing the unicode symbols above with @ for degrees; ' for minutes; and " for seconds. If you typed it in correctly, it will appear in the calculator window.
All we have to do now is convert it. Calc can convert between a wide range of units and systems, but we only care about decimals. Type c d and Calc will convert it to a decimal number. If you entered 37@26'36.42" you should see 37.44345 appear in its place.
OK, so we know it can do it, but how do we weaponize it? It so happens that Calc comes with a neat, little (though underdocumented) command called calc-eval.
Entering IELM, M-x ielm, we can query the calculator in real time:
ELISP> (calc-eval "1+2") "3"
Yep. It does work. The hardest part about using it is mapping the “algebraic” notation used above with the indirect, keybinding-based input you use in the RPN calculator. Thankfully, the manual and (often) the trail will tell you the name of the function you are calling.
Let’s digress a little so I can show you how neat this calculator actually is: solving the elementary equation 2x+5=10. In the calculator, type m a to go to algebraic mode; next, type (2x + 5 = 10) — don’t forget the brackets — and it should appear as an equation. Finally, type a S to “solve for” a variable — and when prompted, answer x. The answer will appear in your calculator window. How awesome is that?
Back to the challenge. Calling the “convert to degrees” function is what we need to do, and the answer is hidden in the trail — it’s called deg.
Putting it all together, and we get:
ELISP> (calc-eval "deg(37@ 26' 36.42\")") "37.44345"
That looks right. But the original challenge said that we had to take a string, like the one given above, and map that. So here’s my solution:
(defun hms-to-dec (hms-str) (let ((hms (split-string hms-str "[°′″NW ]" t))) (flet ((to-deg () (string-to-number (calc-eval (format "deg(%s@ %s' %s\")" (pop hms) (pop hms) (pop hms)))))) (list (to-deg) (to-deg)))))
Calling it from IELM yields the following answer:
ELISP> (hms-to-dec "37°26′36.42″N 06°15′14.28″W") (37.44345 6.25396666667)
Looks good to me. Job done, and a fun challenge.
In my What’s New In Emacs 24 series (part one, part two) I briefly mentioned that pcomplete, the programmable completion library featured prominently in Eshell, now supports M-x shell out of the box. That’s great news for shell mode fans as the completion mechanism adds a lot of nifty functionality to a mode that lacks the native completion provided by underlying the shell itself.
The most amazing thing about the completion mechanism is that it has been in Emacs for ages but never made much of a public appearance and has gone virtually unnoticed due to its limited use in Emacs. In fact, I think it’s only used in EShell, ERC, Org Mode and now, finally, Shell Mode.
Programmable, Context-Sensitive Completion
To use pcomplete you won’t have to do anything, because as of Emacs 24 it is now supported automatically when you launch a new shell session. Emacs ships with a handful of pcomplete functions that enhance the otherwise drab filename completion with context-sensitive completion similar to what you can do with bash/zsh completion. Of particular note is the scp, ssh, mount, umount and make support.
The following table lists the commands supported by shell mode (or indeed any mode that supports pcomplete, including Eshell.)
| Command | Notes |
|---|---|
| bzip2 | Completes arguments and lists only bzipped files. |
| cd | Completes directories. |
| chgrp | Completes list of known groups on the system. |
| chown | Completes user and group perms, but only if you use user.group. |
| cvs | Completes commands and parameter options and cvs entries and modules. |
| gdb | Completes only directories or files with eXecute permission. |
| gzip | Completes arguments and lists only gzipped files. |
| kill | Lists signals if completed with just a -, otherwise it completes all system PIDs. |
| make | Completes arguments and valid makefiles in the directory; if a valid makefile is completed with make -f FILE a list of rule names from the file itself are completed. |
| mount | Completes arguments and valid filesystem types if completed with mount -t TYPE. |
| pushd | Identical to cd. |
| rm | Completes arguments and filenames and directories. |
| rmdir | Completes directories. |
| rpm | Very sophisticated completion mechanism for most of rpm, the Redhat Package Manager. Context-sensitive completion for almost all commands, including package lookup. |
| scp | Completes arguments, SSH known hosts and remote file lookup (using TRAMP) if the format is scp host:/. |
| ssh | Completes arguments and SSH known hosts. |
| tar | Completes arguments, including context-sensitive completion for POSIX arguments, and file name completion. |
| time | Completes directories and files with eXecutable permission. |
| umount | Completes arguments, mounted directories and filesystem types (like mount) |
| which | Supposed to provide simple filename completion of all known binaries (wouldn’t be useful otherwise!) but appears to not work right. |
| xargs | Completes directories and files with eXecutable permission. |
Custom Completion
It goes without saying that a completion library called programmable completion is, well, programmable.
Adding simple parameter completion is an easy job but anything more than that and it gets hairy as, not surprisingly, this library is virtually undocumented (though an optimist would say the source is all the documentation you need…)
I’ll demonstrate how to add rudimentary support for git.
The first thing we need to do is establish the order in which parameters must be given; for git, it’s somewhat consistent: git [options] <command> [<args>]
For now I’ll stick to the commands as that’s what people use the most anyway. The commands, in list form, are:
(defconst pcmpl-git-commands '("add" "bisect" "branch" "checkout" "clone" "commit" "diff" "fetch" "grep" "init" "log" "merge" "mv" "pull" "push" "rebase" "reset" "rm" "show" "status" "tag" ) "List of `git' commands")
The syntax for pcomplete is rather clever: it will use dynamic dispatch to resolve the elisp function provided it is named a certain way. All commands are named pcomplete/COMMAND or pcomplete/MAJOR-MODE/COMMAND. Provided you follow that naming scheme your command will automagically work.
Next, we need to present a list of valid commands — in this case the ones in pcmpl-git-commands, but it could be any form — to the command pcomplete-here.
(defun pcomplete/git () "Completion for `git'" (pcomplete-here* pcmpl-git-commands))
Now when you try to tab-complete the first argument to git it will list our commands. Sweet.
Let’s extend it further by adding support for the add and rm commands. I want the aforementioned commands to provide the standard filename/filepath completion if, and only if, the command is add or rm.
This is surprisingly easy to do using pcomplete-match, a function that asserts a certain regexp matches a particular function argument index. Note that the call to pcomplete-here is in a while loop; this is so you can complete as many files as you like, one after another. One advantage of pcomplete-here is that it won’t display files you have already completed earlier in the argument trail — that’s very useful for a command like add.
(defun pcomplete/git () "Completion for `git'" ;; Completion for the command argument. (pcomplete-here* pcmpl-git-commands) ;; complete files/dirs forever if the command is `add' or `rm'. (if (pcomplete-match (regexp-opt '("add" "rm")) 1) (while (pcomplete-here (pcomplete-entries)))))
Ok, that was easy. Now let’s make it a bit more dynamic by extending our code to support the git checkout command so it will complete the list of branches available to us locally.
To do this we need a helper function that takes the output of a call to shell-command and maps it to an internal elisp list. This is easily done with some quick hackery.
The variable pcmpl-git-ref-list-cmd holds the shell command we want Emacs to run for us. It gets every ref there is and we then filter by sub-type (heads, tags, etc.) later. The function pcmpl-git-get-refs takes one argument, type, which is the ref type to filter by.
(defvar pcmpl-git-ref-list-cmd "git for-each-ref refs/ --format='%(refname)'" "The `git' command to run to get a list of refs") (defun pcmpl-git-get-refs (type) "Return a list of `git' refs filtered by TYPE" (with-temp-buffer (insert (shell-command-to-string pcmpl-git-ref-list-cmd)) (goto-char (point-min)) (let ((ref-list)) (while (re-search-forward (concat "^refs/" type "/\\(.+\\)$") nil t) (add-to-list 'ref-list (match-string 1))) ref-list)))
And finally, we put it all together. To keep the code clean I’ve switched to using a cond form for readability.
(defconst pcmpl-git-commands '("add" "bisect" "branch" "checkout" "clone" "commit" "diff" "fetch" "grep" "init" "log" "merge" "mv" "pull" "push" "rebase" "reset" "rm" "show" "status" "tag" ) "List of `git' commands") (defvar pcmpl-git-ref-list-cmd "git for-each-ref refs/ --format='%(refname)'" "The `git' command to run to get a list of refs") (defun pcmpl-git-get-refs (type) "Return a list of `git' refs filtered by TYPE" (with-temp-buffer (insert (shell-command-to-string pcmpl-git-ref-list-cmd)) (goto-char (point-min)) (let ((ref-list)) (while (re-search-forward (concat "^refs/" type "/\\(.+\\)$") nil t) (add-to-list 'ref-list (match-string 1))) ref-list))) (defun pcomplete/git () "Completion for `git'" ;; Completion for the command argument. (pcomplete-here* pcmpl-git-commands) ;; complete files/dirs forever if the command is `add' or `rm' (cond ((pcomplete-match (regexp-opt '("add" "rm")) 1) (while (pcomplete-here (pcomplete-entries)))) ;; provide branch completion for the command `checkout'. ((pcomplete-match "checkout" 1) (pcomplete-here* (pcmpl-git-get-refs "heads")))))
And that’s that. A simple completion mechanism for git. Put this in your .emacs or init file and you’re done.
As I mentioned in my What’s New In Emacs 24 series (part one, part two) Emacs 24 can now sort the colors in M-x list-colors-display based on the RGB or HSV distance to another color. This is perhaps of most interest to web designers or color theme creators, but it’s still a fun little addition.
Unfortunately, this functionality is only available through the customize interface, and that means most people will never bother with it at all. Hence, I’ve written two helper commands below to make this neat addition to Emacs 24 easier to use.
If you’re looking up a named color (like red) then it must exist in M-x list-colors-display; if it’s an HTML color then as long as it is valid it will work.
To use it, you must invoke M-x find-nearest-color and enter the name of a color. The other command, find-nearest-color-at-point does just that: it looks at what your point is on and tries to look it up.
(defun find-nearest-color (color &optional use-hsv) "Finds the nearest color by RGB distance to `color'. If called with a universal argument (or if `use-hsv' is set) use HSV instead of RGB. Runs \\[list-colors-display] after setting `list-colors-sort'" (interactive "sColor: \nP") (let ((list-colors-sort `(,(if (or use-hsv current-prefix-arg) 'hsv-dist 'rgb-dist) . ,color))) (if (color-defined-p color) (list-colors-display) (error "The color \"%s\" does not exist." color)))) (defun find-nearest-color-at-point (pt) "Finds the nearest color at point `pt'. If called interactively, `pt' is the value immediately under `point'." (interactive "d") (find-nearest-color (with-syntax-table (copy-syntax-table (syntax-table)) ;; turn `#' into a word constituent to help ;; `thing-at-point' find HTML color codes. (modify-syntax-entry ?# "w") (thing-at-point 'word))))
If you hack a lot of elisp you will notice that I’ve chosen to hack the syntax table (temporarily) to make thing-at-point play nice with the HTML color syntax (#ABCDEF) as the character — and this will depend on your mode and thus your syntax table, of course — # is rarely a “word” constituent. I think this is a much cleaner way of “getting” things at point than messing with regular expressions as you will inevitably end up reinventing most of thing-at-point doing so.
This is part two of my What’s New in Emacs 24 series. Part one.
Trash changes
`delete-by-moving-to-trash’ now only affects commands that specify
trashing. This avoids inadvertently trashing temporary files.
An important change as some features like flymake create
new temporary files every time it runs an external make or linting
tool, resulting in a gazillion temp files ending up in your trash can.
Calling `delete-file’ or `delete-directory’ with a prefix argument
now forces true deletion, regardless of `delete-by-moving-to-trash’.
Very handy addition that mimicks Windows and most major window managers.
New option `list-colors-sort’ defines the color sort order
for `list-colors-display’.
Wow, this is an… interesting.. addition to Emacs. You can govern the sort order of colors now such as HSV or RGB distance from a particular color… very neat.
An Emacs Lisp package manager is now included.
This is a convenient way to download and install additional packages,
from a package repository at http://elpa.gnu.org.
Package Manager
`M-x list-packages’ shows a list of packages, which can be
selected for installation.
New command `describe-package’, bound to `C-h P’.
By default, all installed packages are loaded and activated
automatically when Emacs starts up. To disable this, set
`package-enable-at-startup’ to nil. To change which packages are
loaded, customize `package-load-list’.
A much sought-after addition to Emacs. Either Emacs has joined the “App Store” bandwagon or it has finally caught up to XEmacs circa 2001. Your call. Of course, the official stance is that all packages must have its copyright assigned to the FSF, so it’s unlikely that it’ll ever be the definitive package source.
I suggest you alter the package manager archive list and add the repository, Marmalade:
(require 'package) (add-to-list 'package-archives '("marmalade" . "http://marmalade-repo.org/packages/"))
Custom Themes
`M-x customize-themes’ lists Custom themes which can be enabled.
New option `custom-theme-load-path’ is the load path for themes.
Emacs no longer looks for custom themes in `load-path’. The default
is to search in `custom-theme-directory’, followed by a built-in theme
directory named “themes/” in `data-directory’.
New option `custom-safe-themes’ records known-safe theme files.
If a theme is not in this list, Emacs queries before loading it, and
offers to save the theme to `custom-safe-themes’ automatically. By
default, all themes included in Emacs are treated as safe.
Color themes, a very popular package and one of the first a lot of people install, is now superseded by a new version bundled with Emacs.
TRAMP
The user option `remote-file-name-inhibit-cache’ controls whether
the remote file-name cache is used for read access.
This caching mechanism will stop TRAMP from resolving remote file attributes over and over again in a short timespan; you can change numeric value of the variable which govern the cache expiry (in seconds) or enable or disable it permanently with nil or t respectively.
File- and directory-local variable changes
You can stop directory local vars from applying to subdirectories.
Add an element (subdirs . nil) to the alist portion of any variables
settings to indicate that the section should not apply to
subdirectories.
A welcome change to people who use directory-local variables a lot.
Directory local variables can apply to some file-less buffers.
Affected modes include dired, vc-dir, and log-edit. For example,
adding “(diff-mode . ((mode . whitespace)))” to .dir-locals.el will
turn on `whitespace-mode’ for *vc-diff* buffers. Modes should call
`hack-dir-local-variables-non-file-buffer’ to support this.
Another step in the direction towards “project”-like functionality.
Using “mode: MINOR-MODE” to enable a minor mode is deprecated.
Instead, use “eval: (minor-mode 1)”.
Pretty self-explanatory.
Misc
New primitive `secure-hash’ that supports many secure hash algorithms
including md5, sha-1 and sha-2 (sha-224, sha-256, sha-384 and sha-512).
The elisp implementation sha1.el is removed. Feature sha1 is provided
by default.
Excellent addition, though I cannot find any commands that let you use it on a string or buffer directly.
Window changes
The behavior of `display-buffer’ is now customizable in detail.
New option `display-buffer-base-action’ specifies a list of
user-determined display “actions” (functions and optional arguments
for choosing the displaying window).This takes precedence over the default display action, which is
specified by `display-buffer-fallback-action’.New option `display-buffer-alist’ maps buffer name regexps to
display actions, taking precedence over `display-buffer-base-action’.
Interesting change that should make it easier for people to control which frame or window a buffer should appear in. Having used the defaults for so long I would find it hard to switch, but I’m sure there are lots of people — especially tiling WM fans — who would want Emacs to use more frames.
New option `window-combination-limit’.
The new option `window-combination-limit’ allows to return the space
obtained for resizing or creating a window more reliably to the window
from which such space was obtained.
This command does a lot more, seemingly, than what the description here says it does; reading the docstring for the variable clouds the issue as it is pretty obtuse. From what I can tell, when you C-x 0 a window while this option is enabled you will remain in its parent, and that parent is the one in which the new window was spawned from (with C-x 2 or C-x 3). That it claims to return the space more effciently may be a side effect of this change in behavior — or maybe it’s the other way around? Either way, enable it and see if you like how it works.
New option `window-combination-resize’.
The new option `window-combination-resize’ allows to split a window that
otherwise cannot be split because it’s too small by stealing space from
other windows in the same combination. Subsequent resizing or deletion
of the window will resize all windows in the same combination as well.
Now this is a useful switch (unlike the one above) as now Emacs will assign an equal amount of space to each window when you split or delete windows.
New commands `maximize-window’ and `minimize-window’.
These maximize and minimize the size of a window within its frame.
New commands `switch-to-prev-buffer’ and `switch-to-next-buffer’.
These functions allow to navigate through the live buffers that have
been shown in a specific window.
Sorely missed window and buffer commands finally make an (official) appearance in Emacs. I suggest you bind these commmands to something right away! Of course there’s no restore-window command.
New functions `window-state-get’ and `window-state-put’.
These functions allow to save and restore the state of an arbitrary
frame or window as an Elisp object.
Useful for Elisp hackers looking for a way to save and restore the window states.
The inactive minibuffer has its own major mode `minibuffer-inactive-mode’.
This is handy for minibuffer-only frames, and is also used for the “mouse-1
pops up *Messages*” feature, which can now easily be changed.
Even more streamlining of Emacs.
Editing Changes in Emacs 24.1
Search changes
C-y in Isearch is now bound to isearch-yank-kill, instead of
isearch-yank-line.
M-y in Isearch is now bound to isearch-yank-pop, instead of
isearch-yank-kill.
M-s C-e in Isearch is now bound to isearch-yank-line.
So this change is a controversial one, as you had to press M-y to yank in isearch, whereas now it is consistent with the rest of Emacs: C-y yanks and M-y cycles the kill ring. I like the change as it keeps Emacs consistent and easier to use for new users. Not to mention that adding the rest of the line to an isearch query is an uncommon action.
New commands `count-words-region’ and `count-words’.
`count-lines-region’ is now an alias for `count-words-region’,
bound to M-=, which shows the number of lines, words, and characters.
Another sorely missed feature but do be careful using it. The notion of a “word” is governed entirely by the syntax table, and so what counts as a single word in one mode may well be more than one in another. For example, the command “count-words-region” — is that one or three words? Again, it depends on the mode.
The default value of `backup-by-copying-when-mismatch’ is now t.
This variable is only of interest to people who use backup-by-copying and then only if you care about how Emacs preserves the owner and group of the file.
The command `just-one-space’ (M-SPC), if given a negative argument,
also deletes newlines around point.
The just-one-space command is already a very useful tool in programming, and with the new negative argument toggle, even more so.
Deletion changes
New option `delete-active-region’.
If non-nil, [delete] and DEL delete the region if it is active and no
prefix argument is given. If set to `kill’, these commands kill
instead.New command `delete-forward-char’, bound to [delete].
This is meant for interactive use, and obeys `delete-active-region’.
The command `delete-char’ does not obey `delete-active-region’.
Another standardization change aimed at helping new users. Confusingly, however, it does not replace the behavior afforded by M-x delete-selection-mode. That mode will only delete the selected text if, and only if, you type a character. So alldelete-active-region does is make it so if you press a key bound to delete-[forward/backward]-char it will delete the active region. The two of them, combined, make Emacs behave like other applications.
`delete-backward-char’ is now a Lisp function.
Apart from obeying `delete-active-region’, its behavior is unchanged.
However, the byte compiler now warns if it is called from Lisp; you
should use delete-char with a negative argument instead.
If you’re an elisp author and you use delete-backward-char (or indeed, its forward counterpart) while you have the region active, expect it to nuke the entire selection if delete-active-region is t. Better to switch to delete-char and avoid this problem altogether.
The option `mouse-region-delete-keys’ has been deleted.
Because selections are now deleted by default (if delete-active-region is t) with any key bound to delete-[forward/backward]-char, this option is redundant.
Selection changes.
The default handling of clipboard and primary selections was changed
to conform with modern X applications. In short, most commands for
killing and yanking text now use the clipboard, while mouse commands
use the primary selection.In the following, we provide a list of these changes, followed by a
list of steps to get the old behavior back if you prefer that.
`select-active-regions’ now defaults to t.
Merely selecting text (e.g. with drag-mouse-1) no longer puts it in
the kill ring. The selected text is put in the primary selection, if
the system possesses a separate primary selection facility (e.g. X).
`select-active-regions’ also accepts a new value, `only’.
This means to only set the primary selection for temporarily active
regions (usually made by mouse-dragging or shift-selection);
“ordinary” active regions, such as those made with C-SPC followed by
point motion, do not alter the primary selection.
`mouse-drag-copy-region’ now defaults to nil.
mouse-2 is now bound to `mouse-yank-primary’.
This pastes from the primary selection, ignoring the kill-ring.
Previously, mouse-2 was bound to `mouse-yank-at-click’.
`x-select-enable-clipboard’ now defaults to t on all platforms.
`x-select-enable-primary’ now defaults to nil.
Thus, commands that kill text or copy it to the kill-ring (such as
M-w, C-w, and C-k) also use the clipboard—not the primary selection.
The “Copy”, “Cut”, and “Paste” items in the “Edit” menu are now
exactly equivalent to, respectively M-w, C-w, and C-y.
Note that on MS-Windows, `x-select-enable-clipboard’ was already
non-nil by default, as Windows does not support the primary selection
between applications.
Yet again, more standardization. That can only be a good thing. This alone should cut the number of mailinglist and IRC channel questions in half.
If you’re a curmudgeonly Emacs hacker, follow this advice to get the old behavior back:
To return to the previous behavior, do the following:
Change `select-active-regions’ to nil.
Change `mouse-drag-copy-region’ to t.
Change `x-select-enable-primary’ to t (on X only).
Change `x-select-enable-clipboard’ to nil.
Bind `mouse-yank-at-click’ to mouse-2.
Support for X cut buffers has been removed.
X clipboard managers are now supported.
To inhibit this, change `x-select-enable-clipboard-manager’ to nil.
Rectangle Commands
New command `rectangle-number-lines’, bound to `C-x r N’, numbers
the lines in the current rectangle. With a prefix argument, this
prompts for a number to count from and for a format string.
Sorely needed numbering rectangle command makes an appearance at last. Saves you the hassle of using a macro or regular expression.
The default value of redisplay-dont-pause is now t
This makes Emacs feel more responsive to editing commands that arrive
at high rate, e.g. if you lean on some key, because stopping redisplay
in the middle (when this variable is nil) forces more expensive
updates later on, and Emacs appears to be unable to keep up.
I blogged about a similar command earlier, and this, too, is one of those commands you will have to use for a while to find out if it benefits you or not.
The behavior of <TAB> for active regions in Text mode has changed.
In Text and related modes, typing <TAB> (`indent-for-tab-command’)
when the region is active causes Emacs to indent all the lines in the
region, aligning them with the line previous to the first line in the
region (or with the left margin if there is no previous line).
About. Time. Ever hit TAB in a mode with no indentation engine only to see your lines indented like this:
foo
bar
baz
Well, your woes are over. It will now apply a fixed indentation to the entire region instead of doing it recursively to each one.
Changes in Specialized Modes and Packages in Emacs 24.1
Archive Mode has basic support for browsing and updating 7z archives.
As 7zip is becoming more and more popular, this change can only be a good thing.
browse-url has a new variable `browse-url-mailto-function’
specifies how mailto: URLs are handled. The default is `browse-url-mail’.
This is a great addition, especially to people who use Emacs for email. Make sure you (require 'browse-url) before assigning to the variable.
BibTeX mode
BibTeX mode now supports biblatex.
Use the variable bibtex-dialect to select support for different BibTeX
dialects. bibtex-entry-field-alist is now an obsolete alias for
bibtex-BibTeX-entry-alist.
New command `bibtex-search-entries’ bound to C-c C-a.
New `bibtex-entry-format’ option `sort-fields’, disabled by default.
New variable `bibtex-search-entry-globally’.
Calendar, Diary, and Appt
Diary entries can contain non-printing `comments’.
See the variable `diary-comment-start’.
Appointments can specify their individual warning times.
See the variable `appt-warning-time-regexp’.
The function specified by `appt-disp-window-function’ may be passed
lists of arguments if multiple appointments are due at similar times.
If you are using a custom function for this, you should update it.
New function `diary-hebrew-birthday’.
Elements of `calendar-day-abbrev-array’ and `calendar-month-abbrev-array’
may no longer be nil, but must all be strings.
The obsolete (since Emacs 22.1) method of enabling the appt package
by adding appt-make-list to diary-hook has been removed. Use appt-activate.
Some appt variables (obsolete since Emacs 22.1) have been removed:
appt-issue-message (use the function appt-activate)
appt-visible/appt-msg-window (use the variable appt-display-format)
Some diary function aliases (obsolete since Emacs 22.1) have been removed:
view-diary-entries, list-diary-entries, show-all-diary-entries
The diary and calendar functionality in Emacs is woefully underused. The calendar, in particular, is useful; and like all Emacs features it is extremely advanced and capable of a great many things.
CC Mode (C, C++, etc.)
New feature to “guess” the style in an existing buffer.
Perhaps one of the more significant features to make an appearance in Emacs 24. This feature should work with any cc-mode-derived mode, and hopefully save you the hassle of having to configure the 50 billion different indentation settings Emacs has.
Comint
comint and modes derived from it use the generic completion code.
All this change really means is that comint-derived modes will, by default, use completion-at-point instead of a now deprecated, comint-specific version.
Compilation mode
Compilation mode can be used without font-lock-mode.
`compilation-parse-errors-function’ is now obsolete.
`compilation-filter-start’ is let-bound to the start of the text
inserted by the compilation filter function, when calling
compilation-filter-hook.
`compilation-error-screen-columns’ is obeyed in the editing buffer.
So programming language modes can set it, whereas previously only the value
in the *compilation* buffer was used.
Customize
Customize buffers now contain a search field.
The search is performed using `customize-apropos’.
To turn off the search field, set custom-search-field to nil.
Yay! Very useful and a welcome change. If you’re not comfortable rummaging around Emacs’s internals to track down settings, you now have the option of using the Search interface in customize.
Custom options now start out hidden if at their default values.
Use the arrow to the left of the option name to toggle visibility.
custom-buffer-sort-alphabetically now defaults to t.
The color widget now has a “Choose” button, which allows you to
choose a color via list-colors-display.
Definitely a welcome addition.
D-Bus
It is possible now, to access alternative buses than the default
system or session bus.
dbus-register-{service,method,property}
The -method and -property functions do not automatically register
names anymore.The new function dbus-register-service registers a service known name
on a D-Bus without simultaneously registering a property or a method.
D-Bus functionality is a nice-to-have (even though its Windows support is lacking) but I have yet to see any Emacs packages make great use of it. Hopefully external shells like iPython will add support for it to eliminate the reliance on parsing the stdout of a comint process.
Dired-x
dired-jump and dired-jump-other-window called with a prefix argument
read a file name from the minibuffer instead of using buffer-file-name.
The `dired local variables’ feature provided by Dired-x is obsolete.
The standard directory local variables feature replaces it.
ERC changes
New vars `erc-autojoin-timing’ and `erc-autojoin-delay’.
If the value of `erc-autojoin-timing’ is ‘ident, ERC autojoins after a
successful NickServ identification, or after `erc-autojoin-delay’
seconds. The default value, ‘ident, means to autojoin immediately
after connecting.
Very handy if you use ERC for your IRC needs, especially on networks like Freenode where you cannot join certain channels without identifying to nickserv beforehand.
New variable `erc-coding-system-precedence’: If we use `undecided’
as the server coding system, this variable will then be consulted.
The default is to decode strings that can be decoded as utf-8 as
utf-8, and do the normal `undecided’ decoding for the rest.
Eshell changes
The default value of eshell-directory-name is a directory named
“eshell” in `user-emacs-directory’. If the old “~/.eshell/” directory
exists, that is used instead.
This change was probably prompted by a need for better cross-platform support as some older versions of Windows don’t handle directories and files beginning with ‘.’ very well.
GDB-MI
GDB User Interface migrated to GDB Machine Interface and now
supports multithread non-stop debugging and debugging of several
threads simultaneously.
IDO Mode
In ido-mode, C-v is no longer bound to ido-toggle-vc.
The reason is that this interferes with cua-mode.
Sensible change for people who use CUA.
Image mode
RET (`image-toggle-animation’) toggles animation, if the displayed
image can be animated.
Option `image-animate-loop’, if non-nil, loops the animation.
If nil, `image-toggle-animation’ plays the animation once.
Info
New command `info-display-manual’ displays an Info manual
specified by its name. If that manual is already visited in some Info
buffer within the current session, the command will display that
buffer. Otherwise, it will load the manual and display it. This is
handy if you have many manuals in many Info buffers, and don’t
remember the name of the buffer visiting the manual you want to
consult.
This command is essentially a wrapper around M-: (info "(emacs)Top").
Misc Changes
The Landmark game is now invoked with `landmark’, not `lm’.
MH-E has been upgraded to MH-E version 8.3.1.
See MH-E-NEWS for details.
Modula-2 mode provides auto-indentation.
mpc.el: Can use pseudo tags of the form tag1|tag2 as a union of two tags.
Prolog mode has been completely revamped, with lots of additional
functionality such as more intelligent indentation, electricity, support for
more variants, including Mercury, and a lot more.
Neat. I’ve not done Prolog in many years, but last time I did I found Prolog support in Emacs lacking.
Rmail
The command `rmail-epa-decrypt’ decrypts OpenPGP data
in the Rmail incoming message.
The variable `rmail-message-filter’ no longer has any effect.
This change was made in Emacs 23.1 but was not advertised at the time.
Try using `rmail-show-message-hook’ instead.
Shell mode
This entry is a duplicate of the completion change mentioned in part 1
Shell mode uses pcomplete rules, with the standard completion UI.
The `shell’ command prompts for the shell path name if the default
directory is a remote file name and neither the environment variable
$ESHELL nor the variable `explicit-shell-file-name’ is set.
New variable `shell-dir-cookie-re’.
If set to an appropriate regexp, Shell mode can track your cwd by
reading it from your prompt.
This will no doubt please a lot of people tired of shell losing track of the PWD. As an alternative you can use dirtrack which is what I have historically used.
SQL Mode enhancements.
`sql-dialect’ is an alias for `sql-product’.
New variable `sql-port’ specifies the port number for connecting
to a MySQL or Postgres server.
With the new connection string functionality this new option is probably moot for most people who interact with databases frequently.
The command `sql-product-interactive’ now takes a prefix argument,
which causes it to prompt for an SQL product instead of the current
value of `sql-product’.
Product-specific SQL interactive commands now take prefix args.
These commands (`sql-sqlite’, `sql-postgres’, `sql-mysql’, etc.),
given a prefix argument, prompt for a name for the SQL interactive
buffer. This reduces the need for calling `sql-rename-buffer’.
Handy, but it would be better if the buffer named itself more sensibly to begin with – such as using the database and username in the buffer name.
SQL interactive modes suppress command continuation prompts, and
replace tabs with spaces. The first change impacts multiple line SQL
statements entered with C-j between each line, statements yanked into
the buffer and statements sent with `sql-send-*’ functions. The
second change prevents the MySQL and Postgres interpreters from
listing object name completions when sent text via `sql-send-*’
functions.
This is a good change. If it works properly in practice it will eliminate all the annoying continuation prompts and other text flotsam in the inferior sql shell. The only other issue has always been the sending of large buffers to the inferior sql buffer. Earlier Emacsen would hang.
New custom variables control prompting for login parameters.
Each supported product has a custom variable `sql-*-login-params’,
which is a list of the parameters to be prompted for before a
connection is established.
New variable `sql-connection-alist’ for login parameter values.
This can be used to store different username, database and server
values. Connections defined in this variable appear in the submenu
SQL->Start… for making new SQLi sessions.
New command `sql-connect’ starts a predefined SQLi session,
using the login parameters from `sql-connection-alist’.
New “Save Connection” menu item in SQLi buffers.
This gathers the login params specified for the SQLi session, if it
was not started by a connection, and saves them as a new connection.
These changes all govern the much-needed and much-welcome connection list system now in place. As an added cherry on top the dynamic way databases are loaded makes it (relatively) easy — in theory anyway — to add new databases and have most of the features work out of the box.
Commands for listing database objects and details.
an SQLi session, you can get a list of objects in the database.
The contents of these lists are product specific.
`C-c C-l a’ or the “SQL->List all objects” menu item
lists all the objects in the database. With a prefix argument, it
displays additional details or extend the listing to include other
schemas objects.
`C-c C-l t’ or the “SQL->List Table details” menu item
prompts for the name of a database table or view and displays the list
of columns in the relation. With a prefix argument, it displays
additional details about each column.
Now *this* is neat. This completion functionality is a serious aid if you are using a database like Oracle where every object is defined as metadata.
New options `sql-send-terminator’ and `sql-oracle-scan-on’.
The first option is useful if, like me, you often forget to terminate your queries; the second is mandatory if you use the & placeholders in Oracle as now Emacs will query for the values (like SQLPlus does.)
An API for manipulating SQL product definitions has been added.
There are detailed instructions on how to do this in the commentary section of the library, sql.el.
TeX modes
latex-electric-env-pair-mode keeps \begin..\end matched on the fly.
Tramp
There exists a new inline access method “ksu” (kerberized su).
I’ve never used kerberos with su before, but that it is now supported will surely come in handy in the future.
The following access methods are discontinued: “ssh1_old”,
“ssh2_old”, “scp1_old”, “scp2_old”, “imap”, “imaps” and “fish”.
The option `ange-ftp-binary-file-name-regexp’ has changed its
default value to “”.
`url-queue-retrieve’ downloads web pages asynchronously, but allow
controlling the degree of parallelism.
VC and related modes
Support for pulling on distributed version control systems.
The vc-pull command runs a “pull” operation, if it is supported.
This updates the current branch from upstream. A prefix argument
means to prompt the user for specifics, e.g. a pull location.`vc-update’ is now an alias for `vc-pull’.
Currently supported by Bzr, Git, and Mercurial.
This is perhaps one of the bigger changes a lot of people have waited for. With Emacs using a DVCS now as its official repository, the addition was inevitable.
Support for merging on distributed version control systems.
The vc-merge command now runs a “merge” operation, if it is supported.
This merges another branch into the current one. This command prompts
the user for specifics, e.g. a merge source.Currently supported for Bzr, Git, and Mercurial.
Another very welcome feature.
New option `vc-revert-show-diff’ controls whether `vc-revert’
shows a diff while querying the user. It defaults to t.
Good to know that this behavior can be toggled, but I’ll leave it on; you never know when you might revert something you want to keep.
Log entries in some Log View buffers can be toggled to display a
longer description by typing RET (log-view-toggle-entry-display).
In the Log View buffers made by `C-x v L’ (vc-print-root-log), you can
use this to display the full log entry for the revision at point.Currently supported for Bzr, Git, and Mercurial.
Packages using Log View mode can enable this functionality by
binding `log-view-expanded-log-entry-function’ to a suitable function.
This is useful if you want to customize how Emacs displays the root log view (C-x v L)
New command `vc-ediff’ allows visual comparison of two revisions
of a file similar to `vc-diff’, but using ediff backend.
I’m not sure I’ll switch, as much as I love ediff. Ediff is the 800 lbs gorilla of diffing and merging and, well, when I want to diff something the simple diff view mode works fine.
`vc-toggle-read-only’ is an obsolete alias for `toggle-read-only’.
Since Emacs 23, it has done the same thing as `toggle-read-only’, but
this was not advertised at the time.
The option `vc-initial-comment’ was removed in Emacs 23.2, but
this was not advertised at the time.
Obsolete modes
partial-completion-mode is obsolete.
You can get a comparable behavior with:
(setq completion-styles ‘(partial-completion initials))
(setq completion-pcm-complete-word-inserts-delimiters t)
Never used this mode, but I suppose the more sophisticated completion mechanisms available now have made it redundant.
pc-mode.el is obsolete.
With Emacs slowly standardizing on the same keyboard shortcuts as modern apps, this package has outlived its usefulness. The fact that it aims to emulate a PC gives you an idea of how old it is (16 years.)
sregex.el is obsolete, since rx.el is a strict superset.
rx is a wonderful s-expression based regex engine, and there is no reason why Emacs should have other packages lying around that do essentially the same thing.
s-region.el and pc-select are obsolete.
They are superseded by shift-select-mode enabled by default in 23.1.
More standardization and more obsolescence.
Miscellaneous
f90.el has some support for Fortran 2008 syntax.
Fortran users rejoice.
`copyright-fix-years’ can optionally convert consecutive years to ranges.
I never actually knew this library existed. Cool.
New command `nato-region’ converts text to NATO phonetic alphabet.
That’s not terribly useful, except when you have to spell the word “Zyzygy” to somebody over the phone. It’s right up there with the [un]morse-region commands.
New Modes and Packages in Emacs 24.1
Occur Edit mode applies edits made in *Occur* buffers to the
original buffers. It is bound to “e” in Occur mode.
Probably one of the best changes in Emacs 24 yet. Forget color themes and package managers. This is the sort of thing that gives Emacs the edge, and now we no longer need a 3rd party package to do this.
New global minor modes electric-pair-mode, electric-indent-mode,
and electric-layout-mode.
These modes are designed to replace the umpteen different electric functions each major mode invariably reinvent. Still more standardization!
tabulated-list.el provides a generic major mode for tabulated data,
from which other modes can be derived.
It seems there are no actual modes that inherit from this generic framework yet, but looking code it seems like it’ll be a doozy to support a variety of tabulated formats with only about 10-15 lines of code.
pcase.el provides the ML-style pattern matching macro `pcase’.
secrets.el is an implementation of the Secret Service API, an
interface to password managers like GNOME Keyring or KDE Wallet. The
Secret Service API requires D-Bus for communication. The command
`secrets-show-secrets’ offers a buffer with a visualization of the
secrets.
Seems like a useful addition to Emacs. I’ll have to study this in greater detail.
notifications.el provides an implementation of the Desktop
Notifications API. It requires D-Bus for communication.
Useful if you combine it with ERC to notify you if you’ve received an IRC message when Emacs is not in focus.
soap-client.el supports access to SOAP web services from Emacs.
soap-inspect.el is an interactive inspector for SOAP WSDL structures.
Very handy, but I wonder if it blocks the thread while it is in operation.
xmodmap-generic-mode for xmodmap files.
New emacs-lock.el package.
(The pre-existing one has been renamed to old-emacs-lock.el and moved
to obsolete/.) Now, Emacs Lock is a proper minor mode
`emacs-lock-mode’. Protection against exiting Emacs and killing the
buffer can be set separately. The mechanism for auto turning off
protection for buffers with inferior processes has been generalized.
If you’re prone to twitch killing this could be the mode for you.
Incompatible Lisp Changes in Emacs 24.1
`char-direction-table’ and the associated function `char-direction’
were deleted. They were buggy and inferior to the new support of
bidirectional editing introduced in Emacs 24. If you need the
bidirectional properties of a character, use `get-char-code-property’
with the last argument `bidi-class’.
`copy-directory’ now copies the source directory as a subdirectory
of the target directory, if the latter is an existing directory. The
new optional arg COPY-CONTENTS, if non-nil, makes the function copy
the contents directly into a pre-existing target directory.
`compose-mail’ now accepts an optional 8th arg, RETURN-ACTION, and
passes it to the mail user agent function. This argument specifies an
action for returning to the caller after finishing with the mail.
This is currently used by Rmail to delete a mail window.
For mouse click input events in the text area, the Y pixel
coordinate in the POSITION list now counts from the top of the text
area, excluding any header line. Previously, it counted from the top
of the header line.
Removed obsolete name `e’ (use `float-e’ instead).
A backquote not followed by a space is now always treated as new-style.
Test for special mode-class was moved from view-file to view-buffer.
FIXME: This only says what was changed, but not what are the
programmer-visible consequences.
Passing a nil argument to a minor mode function now turns the mode
ON unconditionally.
During startup, Emacs no longer adds entries for `menu-bar-lines’
and `tool-bar-lines’ to `default-frame-alist’ and `initial-frame-alist’.
With these alist entries omitted, `make-frame’ checks the value of the
variable `menu-bar-mode’/`tool-bar-mode’ to determine whether to create
a menu-bar or tool-bar, respectively. If the alist entries are added,
they override the value of `menu-bar-mode’/`tool-bar-mode’.
Regions created by mouse dragging are now normal active regions,
similar to the ones created by shift-selection. In previous Emacs
versions, these regions were delineated by `mouse-drag-overlay’, which
has now been removed.
cl.el no longer provides `cl-19′.
The menu bar bindings’s caches are not used any more.
Use (where-is-internalnil t) instead.
The following obsolete functions and aliases were removed:
comint-kill-output, decompose-composite-char, outline-visible,
internal-find-face, internal-get-face, frame-update-faces,
frame-update-face-colors, x-frob-font-weight, x-frob-font-slant,
x-make-font-bold, x-make-font-demibold, x-make-font-unbold
x-make-font-italic, x-make-font-oblique, x-make-font-unitalic
x-make-font-bold-italic, mldrag-drag-mode-line, mldrag-drag-vertical-line,
iswitchb-default-keybindings, char-bytes, isearch-return-char,
make-local-hook
The following obsolete variables and varaliases were removed:
checkdoc-minor-keymap, vc-header-alist, directory-sep-char, and
font-lock-defaults-alist.
The following obsolete files were removed:
sc.el, x-menu.el, rnews.el, rnewspost.el
Lisp changes in Emacs 24.1
Code can now use lexical scoping by default instead of dynamic scoping.
The `lexical-binding’ variable lets code use lexical scoping for local
variables. It is typically set via file-local variables, in which case it
applies to all the code in that file.
`eval’ takes a new optional argument `lexical’ to choose the new lexical
binding instead of the old dynamic binding mode.
Lexically scoped interpreted functions are represented with a new form
of function value which looks like (closure ENV ARGS &rest BODY).
New macro `letrec’ to define recursive local functions.
New function `special-variable-p’ to check whether a variable is
declared as dynamically bound.
The first of many changes to Emacs’s elisp engine to make it more performant, feature rich and in line with what people expect of a modern Lisp. A large amount of the existing Emacs packages have already been switched to lexical scoping, but it is apparently difficult to get right as Emacs’s codebase is heavily reliant on dynamic scoping to monkeypatch or extend other parts of the code.
An Emacs Lisp testing tool is now included.
Emacs Lisp developers can use this tool to write automated tests for
their code. See the ERT info manual for details.
Emacs gets its first built-in unit testing framework. Time will tell if it is powerful and easy enough to use for package maintainers to adopt it.
Changes for bidirectional display and editing
New function `current-bidi-paragraph-direction’.
This returns the actual value of base direction of the paragraph at
point.
New function `bidi-string-mark-left-to-right’.
Given a string containing characters from right-to-left (RTL) scripts,
this function returns another string which can be safely inserted into
a buffer, such that any following text will be always displayed to the
right of that string. (This works by appending the Unicode
“LEFT-TO-RIGHT MARK” character when the argument string might need that.)
is useful when the buffer has overall left-to-right (LTR)
paragraph direction and you need to insert a string whose contents and
directionality are not known in advance, without disrupting the layout
of the line.
Window changes
Window tree functions are accessible in Elisp.
Functions are provided to return the parent, siblings or child windows
of any window including internal windows (windows not associated with a
buffer) in the window tree.
New function `window-valid-p’ gives non-nil for live and internal
windows.
Window manipulation can deal with internal windows.
Many window handling functions like `split-window’, `delete-window’, or
`delete-other-windows’ as well as the window resizing functions can now
act on any window including internal ones.
window-total-height/-width vs window-body-height/-width.
The function `window-height’ has been renamed to `window-total-height’
and `window-width’ has been renamed to `window-body-width’. The old
names are provided as aliases. Two new functions `window-total-width’
and `window-body-height’ are provided.
Window parameters specific to window handling functions.
For each window you can specify a parameter to override the default
behavior of a number of functions like `split-window’, `delete-window’
and `delete-other-windows’. The variable `ignore-window-parameters’
allows to ignore processing such parameters.
New semantics of third argument of `split-window’.
The third argument of `split-window’ has been renamed to SIDE and can be
set to any of the values ‘below, ‘right, ‘above, or ‘left to make the
new window appear on the corresponding side of the window that shall be
split. Any other value of SIDE will cause `split-window’ to split the
window into two side-by-side windows as before.
Window resizing functions.
A new standard function for resizing windows called `window-resize’ has
been introduced. This and all other functions for resizing windows no
longer delete any windows when they become too small.
Deleting the selected window now selects the most recently selected
live window on that frame instead.
`adjust-window-trailing-edge’ adjustments.
`adjust-window-trailing-edge’ can now deal with fixed-size windows and
is able to resize other windows if a window adjacent to the trailing
edge cannot be shrunk any more. This makes its behavior more similar to
that of Emacs 21 without compromising, however, its inability to delete
windows which was introduced in Emacs 22.
Window-local buffer lists.
Windows now have local buffer lists. This means that removing a buffer
from display in a window will preferably show the buffer previously
shown in that window with its previous window-start and window-point
positions. This also means that the same buffer may be automatically
shown twice even if it already appears in another window.
This is perhaps the only change to Emacs’s windowing functionality that most people care about; namely that windows had a habit of forgetting the order in which buffers were shown if you had multiple windows.
`switch-to-buffer’ has a new optional argument FORCE-SAME-WINDOW,
which if non-nil requires the buffer to be displayed in the currently
selected window, signaling an error otherwise. If nil, another window
can be used, e.g. if the selected one is strongly dedicated.
`split-window-vertically’ and `split-window-horizontally’ renamed
to `split-window-below’ and `split-window-right’ respectively.
The old names are kept as aliases.
A sensible name change, especially in light of the above that let you split windows from a certain “side” of the active window.
Display actions
The second arg to `display-buffer’ and `pop-to-buffer’ is now
named ACTION, and takes a display action of the same form as
`display-buffer-base-action’ (see Changes, above). A non-nil,
non-list value is treated specially, as the old meaning.
New variable `display-buffer-overriding-action’.
The procedure of `display-buffer’ etc. to choose a window is
determined by combining `display-buffer-overriding-action’,
`display-buffer-alist’, the ACTION arg, `display-buffer-base-action’,
and `display-buffer-fallback-action’. The second and fourth of these
are user-customizable variables.See the docstring of `display-buffer’ for details.
This harkens back to the display buffer change I talked about earlier.
New behavior of `quit-window’.
The behavior of `quit-window’ has been changed in order to restore the
state before the last buffer display operation in that window.
The new option `frame-auto-hide-function’ lets you choose between
iconifying or deleting a frame when burying a buffer shown in a dedicated
frame or quitting a window showing a buffer in a frame of its own.
This is probably only of interest to people who rely heavily on frames instead of windows.
Completion
New variable completion-extra-properties used to specify extra properties
of the current completion:
- :annotate-function, same as the old completion-annotate-function.
- :exit-function, function to call after completion took place.
Functions on completion-at-point-functions can return any of the properties
valid for completion-extra-properties.
completion-annotate-function is obsolete.
New `metadata’ method for completion tables. The metadata thus returned
can specify various details of the data returned by `all-completions’:
- `category’ is the kind of objects returned (e.g., `buffer’, `file’, …),
used to select a style in completion-category-overrides.
- `annotation-function’ to add annotations in *Completions*.
- `display-sort-function’ to specify how to sort entries in *Completions*.
- `cycle-sort-function’ to specify how to sort entries when cycling.
I’m curious to see what package authors will do with this new functionality. The greater support for annotation in particular should prove interesting.
minibuffer-local-filename-must-match-map is not used any more.
Instead, the bindings in minibuffer-local-filename-completion-map are
combined with minibuffer-local-must-match-map.
New variable `completing-read-function’ allows overriding the
behavior of `completing-read’.
This change is probably prompted by a desire to make it easier for packages like IDO mode to replace the default completing-read function without resorting to monkey patching.
`glyphless-char-display’ can now distinguish between graphical and
text terminal display, via a char-table entry that is a cons cell.
`open-network-stream’ can now be used to open an encrypted stream.
It now accepts an optional `:type’ parameter for initiating a TLS
connection, directly or via STARTTLS. To do STARTTLS, additional
parameters (`:end-of-command’, `:success’, `:capabilities-command’)
must also be supplied.
I hope this change means we’ll see native support for TLS in mail programs like GNUS.
pre/post-command-hook are not reset to nil upon error.
Instead, the offending function is removed.
New hook types
New function `run-hook-wrapped’ for running an abnormal hook by
passing the hook functions as arguments to a “wrapping” function.
New macro `with-wrapper-hook’ for running an abnormal hook as a
set of “wrapping” filters, similar to around advice.
`server-eval-at’ is provided to allow evaluating forms on different
Emacs server instances.
Hmm. There’s a lot of potential locked away in this one command. Large-scale mapreduce clusters are, I suppose, now possible with Emacs thanks to Elisp and the functions map and reduce 
`call-process’ allows a `(:file “file”)’ spec to redirect STDOUT to
a file.
Variable `stack-trace-on-error’ removed.
Also the debugger can now “continue” from an error, which means it will jump
to the error handler as if the debugger had not been invoked instead of
jumping all the way to the top-level.
The ability to continue after encountering an error is a pretty big deal if you use edebug a lot!
The function format-time-string now supports the %N directive, for
higher-resolution time stamps.
The exact nature of how the high-resolution time stamps are sourced is left undefined. This functionality is likely to differ slightly depending on the platform.
New function `read-char-choice’ reads a restricted set of characters,
discarding any inputs not inside the set.
This is pretty handy if you’re making your own minibuffer prompts.
`image-library-alist’ is renamed to `dynamic-library-alist’.
The variable is now used to load all kind of supported dynamic libraries,
not just image libraries. The previous name is still available as an
obsolete alias.
New variable `syntax-propertize-function’.
This replaces `font-lock-syntactic-keywords’ which is now obsolete.
This allows syntax-table properties to be set independently from font-lock:
just call syntax-propertize to make sure the text is propertized.
Together with this new variable come a new hook
syntax-propertize-extend-region-functions, as well as two helper functions:
syntax-propertize-via-font-lock to reuse old font-lock-syntactic-keywords
as-is; and syntax-propertize-rules which provides a new way to specify
syntactic rules.
If I understand the implication of this change, it should eliminate the explicit font lock requirement and thus propertizing text with syntax table properties can now be done even if font locking is disabled. I think.
New hook post-self-insert-hook run at the end of self-insert-command.
You can now hook each keypress post facto.
Syntax tables support a new “comment style c” additionally to style b.
Useful if your programming mode has a very particular, or complex, comment syntax.
frame-local variables cannot be let-bound any more.
Major and minor mode changes
`prog-mode’ is a new major mode from which programming modes
should be derived.
`prog-mode-hook’ can be used to enable features for programming
modes, e.g. (add-hook ‘prog-mode-hook ‘flyspell-prog-mode) to enable
on-the-fly spell checking for comments and strings.
Hopefully this switch will mean the end of programming mode-specific hooks for things that you want to set once, and once only (such as binding newline-and-indent.)
New hook `change-major-mode-after-body-hook’, run by
`run-mode-hooks’ just before any other mode hooks.
Enabled globalized minor modes can be disabled in specific modes,
by running (FOO-mode-hook 0) via a mode hook.
This is a useful change if global minor modes are thrust upon you.
`define-minor-mode’ accepts a new keyword :variable.
This let’s you define the variable name in which to store the state of the minor mode. Historically this name was chosen for you automatically.
`delete-file’ and `delete-directory’ now accept optional arg TRASH.
Trashing is performed if TRASH and `delete-by-moving-to-trash’ are
both non-nil. Interactively, TRASH defaults to t, unless a prefix
argument is supplied (see Trash changes, above).
This change echos the other trash change made earlier.
`facemenu-read-color’ is now an alias for `read-color’.
The command `read-color’ now requires a match for a color name or RGB
triplet, instead of signaling an error if the user provides a invalid
input.
Tool-bars can display separators.
Tool-bar separators are handled like menu separators in menu-bar maps,
i.e. via menu entries of the form `(menu-item “–”)’.
I don’t use the tool bar (who does?) but it’s nice to see functionality like that added to Emacs.
Image API
Animated images support (currently animated gifs only).
Time to dust off all those animated GIFs from the 90s.
`image-animated-p’ returns non-nil if an image can be animated.
`image-animate’ animates a supplied image spec.
`image-animate-timer’ returns the timer object for an image that
is being animated.
`image-extension-data’ is renamed to `image-metadata’.
If Emacs is compiled with ImageMagick support (see Startup
Changes), the function `imagemagick-types’ returns a list of image
file extensions that your installation of ImageMagick supports. The
function `imagemagick-register-types’ enables ImageMagick support for
these image types, minus those listed in `imagemagick-types-inhibit’.See the Emacs Lisp Reference Manual for more information.
XML and HTML parsing
If Emacs is compiled with libxml2 support (which is the default),
two new Emacs Lisp-level functions are defined:
`libxml-parse-html-region’ (which will parse “real world” HTML)
and `libxml-parse-xml-region’ (which parses XML). Both return an
Emacs Lisp parse tree.
The snarky remarks about XML as a poor man’s verbose s-expression have finally come full circle now, with the ability to import xml and “real world” HTML to a lisp parse tree. This is actually a surprisingly nifty addition and as I work with xml frequently, a rather welcome one as well. All I need now is proper XPath support.
GnuTLS
Emacs can be compiled with libgnutls support
This is the default. You will then be able to use the functionality
in gnutls.el, namely the `open-gnutls-stream’ and `gnutls-negotiate’
functions. It’s easiest to use these functions through
`open-network-stream’ because it can upgrade connections through
STARTTLS opportunistically or use plain SSL, depending on your needs.Only versions 2.8.x and higher or GnuTLS have been tested.
[FIXME: this statement needs clarifying, given that GnuTLS >= 2.6.6
is the test used by configure.]
gnutls-log-level
Set `gnutls-log-level’ higher than 0 to get debug output. 1 is for
important messages, 2 is for debug data, and higher numbers are as per
the GnuTLS logging conventions. The output is in *Messages*.
As mentioned a few times earlier, Emacs can now be compiled with native TLS support.
Isearch
New hook `isearch-update-post-hook’ that runs in `isearch-update’.
More Misc Changes
Progress reporters can now “spin”.
The MIN-VALUE and MAX-VALUE arguments of `make-progress-reporter’ can
now be nil, or omitted. This makes a “non-numeric” reporter. Each
time you call `progress-reporter-update’ on that progress reporter,
with a nil or omitted VALUE argument, the reporter message is
displayed with a “spinning bar”.
This is pretty cool. I never knew Emacs had a progress bar feature at all!
New variable `revert-buffer-in-progress-p’ is true while a buffer is
being reverted, even if the buffer has a local `revert-buffer-function’.
New variables `delayed-warnings-list’ and `delayed-warnings-hook’ allow
deferring warnings until the main command loop is executed.
`set-auto-mode’ now respects mode: local variables at the end of files,
as well as those in the -*- line.
rx.el has a new `group-n’ construct for explicitly numbered groups.
keymaps can inherit from multiple parents.
This is a subtle but interesting change: you can now build multiple keymaps and pull the keybindings from both of them. How conflicts are resolved is not explained, though.
`debug-on-event’ lets you debug Emacs when stuck because of inhibit-quit.
New reader macro ## which stands for the empty symbol.
This means that the empty symbol can now be read back. Also, #: by itself
(when not immediately followed by a possible symbol character) stands for
an empty uninterned symbol.
Obsolete functions and variables
buffer-substring-filters is obsolete.
Use `filter-buffer-substring-functions’ instead.
`byte-compile-disable-print-circle’ is obsolete.
`deferred-action-list’ and `deferred-action-function’ are obsolete.
`font-lock-maximum-size’ is obsolete.
Changes in Emacs 24.1 on non-free operating systems
New configure.bat option –enable-checking builds Emacs with extra
runtime checks.
New configure.bat option –distfiles to specify files to be
included in binary distribution.
New configure.bat option –without-gnutls to disable automatic
GnuTLS detection.
New configure.bat option –lib for general library linkage, works
with the USER_LIBS build variable.
New make target `dist’ to create binary distribution for MS Windows.
Function `w32-default-color-map’ is now obsolete.
On Nextstep/OSX, the menu bar can be hidden by customizing
ns-auto-hide-menu-bar.
The End
Phew. That’s a lot of stuff! It’s difficult to tell for sure what impact the new additions and changes will have, so we’ll just have to wait and see what people come up with in time.
I think the Emacs maintainers have done a phenomenal job improving the quality of Emacs for us old-timers, and by making tough decisions and aligning Emacs more with modern editors to rope in, and not scare away, new users. Yes, it may well upset a lot of old-timers that this and that changed; good thing, then, that Emacs makes it easy to change things (if you know where to look.)
With Emacs 24 looming around the corner I figured it was time I took a close look at all the new features and changes. As it’s not officially out yet I will settle for the pretest build (get it here.) With that said, there’s little difference (in my experience anyway) between the pretest versions of Emacs and the real mccoy. They’re usually rock-solid and won’t crash on you, but the featureset is, of course, still subject to change.
I’ve annotated most changes and given my view of how useful each change is to most people. Obviously I do not speak for all of us, so don’t neglect to read the NEWS file in its entirety yourself (C-h n.)
Installation Changes in Emacs 24.1
** By default, the installed Info and man pages are compressed.
You can disable this by configuring –without-compress-info.
Useful configure flag for people who value time over space; of course, most of us use a package manager and thus we have no say in how our software is built.
Emacs can be compiled with ImageMagick support.
Emacs links to ImageMagick if version 6.2.8 or newer of the library is
present at build time. To inhibit ImageMagick, use the configure
option `–without-imagemagick’ .
Neat. Imagemagick is a set of tools for manipulating, viewing and converting images and other media files like PDF and PS. With any luck we might get “native” PDF/PS support in Emacs. DocView (the Emacs document viewer) converts each PDF/PS page to PNG and displays that on your screen. Hardly ideal.
Startup Changes in Emacs 24.1
On Windows, Emacs now warns when the obsolete _emacs init file is used,
and also when HOME is set to C:\ by default.
Warning when the HOME directory is unset (or set to C:\)
will aid new Emacs users greatly as it is a rather obscure requirement
that you have to set the HOME environment variable to
control where Emacs will read the init.el or .emacs file from.
Changes in Emacs 24.1
auto-mode-case-fold is now enabled by default.
The auto-mode-alist variable is an association list (hence the -alist bit) of key-value pairs that define what major mode Emacs should set when you open a certain file type. By setting the case fold flag to t Emacs will search the list twice: first time around it will use case sensitive matching and, if no matches were found, it will try again but with insensitive matching. This switch has no effect on Windows.
Completion
shell-mode uses pcomplete rules, with the standard completion UI.
Yowzah! There’s a lot of cool, new functionality hidden away in this gem of a change. The author of pcomplete is John Wiegley — the same guy who wrote EShell — and in Eshell you can complete commandline arguments for things like kill, rmdir, etc. in the same way zsh (or newer versions of bash) does it: by writing your own “programmable” (that’s the p in pcomplete) completion engines for individual commands. Nifty. Of course, this concept is not limited to just shells. You could write completion libraries for programing modes and what have you as well.
This feature is a Big Deal to people (like me) who use M-x shell all the time for day-to-day work. If you’re using Emacs 24 right now try this on for size (Linux, et al only): chgrp TAB and you should see a completion window appear listing all the various groups on your system. Cool! Type M-x apropos pcomplete/ to see a list of all the completion engines available.
I’m definitely going to cover this in greater detail later.
Many packages have been changed to use `completion-at-point’
rather than their own completion code.
Another lovely change made in the name of standardization. The command completion-at-point is a generic command that, by fiddling with the variable completion-at-point-functions, you can control how Emacs does completion in a particular buffer. Lots of modes and packages have their own little completion engines, thus diminishing the value of having such a generic facility in place.
`completion-at-point’ now handles tags and semantic completion.
CEDET’s Semantic library is now usable from the Emacs completion framework. The easiest way to test it is to open up a C source code file and typing M-TAB or C-M-i and watch as Semantic automagically completes for you. For languages where semantic is not enabled by default, you can add it the following line to your major mode hook of choice:
(add-to-list 'completion-at-point-functions 'semantic-completion-at-point-function)
Completion in a non-minibuffer now tries to detect the end of completion
and pops down the *Completions* buffer accordingly.
This behaviour has always been a major source of frustration for me and, I’m sure, others as well. Historically Emacs would leave completion windows hanging around well after you were done using them; not so any more. Emacs will try to hide them automatically “when you’re done with them.” This works really well for me in M-x shell as I’m always TAB-completing paths and filenames.
Completion can cycle, depending on completion-cycle-threshold.
Another productivity enhancer! If you set completion-cycle-threshold to t then Emacs will always cycle through the completion choices if there is more than one left. And because this feature affects completion-at-point it should work in most modes. Set it to an integer and it will only cycle if there are fewer than N items. Set it to nil (the default) and it is disabled entirely.
I like this feature, so I have set it to cycle if there are fewer than 5 items:
(setq completion-cycle-threshold 5)
New completion style `substring’.
The new substring style gives you “wildcard”-style partial completion where Emacs will cast as wide a net as it can to complete your text.
Historically, Emacs has gone through quite a few completion “styles”; that is, different ways in which Emacs tries to complete your query when you press (usally) TAB. There’s a host of them, and you can (surprise!) add your own as well if you really want to. The association list completion-styles-alist lists all the current styles and the variable completion-styles contains an ordered list of styles to cycle through. To use the new substring style you can add it to the completion styles list. Keep in mind that some modes will override the defaults. See the next item.
(add-to-list 'completion-styles 'substring)
Completion style can be set per-category `completion-category-overrides’.
As I mentioned in the above entry you can tell Emacs to only use certain styles in certain parts of Emacs.
Completion of buffers now uses substring completion by default.
This is no doubt a welcome change to the, oh, five or so people out there who choose not to use iswitchb or IDO Mode. Sarcasm aside, this is useful if you screw up your init file and you need to switch to the *Messages* buffer.
Mail changes
default of `send-mail-function’ is now `sendmail-query-once’,
which asks the user (once) whether to use the smtpmail package to send
email, or to use the old defaults that rely on external mail
facilities (`sendmail-send-it’ on GNU/Linux and other Unix-like
systems, and `mailclient-send-it’ on Windows).
The first time you use C-x m (compose-mail) and try to send an e-mail Emacs will now ask you how it should send the e-mail. Simple stuff, really.
smtpmail changes
smtpmail now uses encrypted connections (via STARTTLS) if the
mail server supports them. It also uses the auth-source framework for
getting credentials.
Good news for people who can’t use unencrypted mail servers.
The variable `smtpmail-auth-credentials’ has been removed.
That variable used to have the default value “~/.authinfo”, in which
case you won’t see any difference. But if you changed it to be a list
of user names and passwords, that setting is now ignored; you will be
prompted for the user name and the password, which will then be saved
to ~/.authinfo.You can also manually copy the credentials to your ~/.authinfo file.
For example, if you had(setq smtpmail-auth-credentials
‘((“mail.example.org” 25 “jim” “s!cret”)))then the equivalent line in ~/.authinfo would be
machine mail.example.org port 25 login jim password s!cret
The variable `smtpmail-starttls-credentials’ has been removed.
If you had that set, then you need to put
machine smtp.whatever.foo port 25 key “~/.my_smtp_tls.key” cert “~/.my_smtp_tls.cert”
in your ~/.authinfo file instead.
Useful knowledge if you’re composing mail in Emacs.
sendmail changes
can now add MIME attachments to outgoing messages with the new
command `mail-add-attachment’.
command `mail-attach-file’ was renamed to `mail-insert-file’; the
old name is now an obsolete alias to the new name.
Emacs server and client changes
New option `server-port’ specifies the port on which the Emacs
server should listen.
New emacsclient argument -q/–quiet suppresses some status messages.
New emacsclient argument –frame-parameters can be used to set the
frame parameters of a newly-created graphical frame.
If emacsclient shuts down as a result of Emacs signaling an
error, its exit status is 1.
New emacsclient argument –parent-id ID.
This opens a client frame in parent X window ID, via XEmbed, similar
to the –parent-id argument to Emacs.
More changes to Emacs’s Client-Server architecture. The port command change is probably the most interesting bit.
Internationalization changes
Emacs now supports display and editing of bidirectional text.
Text that includes characters from right-to-left (RTL) scripts, such
as Arabic, Farsi, or Hebrew, is displayed in the correct visual order
as expected by users of those scripts. This display reordering is a
“Full bidirectionality” class implementation of the Unicode
Bidirectional Algorithm. Buffers with no RTL text should look exactly
the same as before.For more information, see the node “Bidirectional Editing” in the
Emacs Manual.
New buffer-local variable `bidi-display-reordering’.
To disable display reordering in any given buffer, change this to nil.
New buffer-local variable `bidi-paragraph-direction’.
If nil (the default), Emacs determines the base direction of each
paragraph from its text, as specified by the Unicode Bidirectional
Algorithm.Setting this to `right-to-left’ or `left-to-right’ forces a particular
base direction on each paragraph in the buffer.Paragraphs whose base direction is right-to-left are displayed
starting at the right margin of the window.
Enhanced support for characters with no glyphs in available fonts.
If a character has no glyphs in any of the available fonts, Emacs
normally displays it either as a hexadecimal code in a box or as a
thin 1-pixel space. In addition to these two methods, Emacs can
display these characters as empty box, as an acronym, or not display
them at all. To change how these characters are displayed, customize
the variable `glyphless-char-display-control’.On character terminals, these methods are used for characters that
cannot be encoded by the `terminal-coding-system’.
New input methods for Farsi: farsi and farsi-translit.
Right-to-Left support in Emacs has been a long time coming and a difficult and complex change to Emacs’s display engine. If you’re a RTL user I’d love to hear how functional and useful it actually is.
`nobreak-char-display’ now also highlights Unicode hyphen chars
(U+2010 and U+2011).
If nobreak-char-display is t Emacs will highlight facsimile unicode whitespace and hyphen characters so you can tell them apart from their ASCII counterpart.
Improved GTK integration
GTK scroll-bars are now placed on the right by default.
Use `set-scroll-bar-mode’ to change this.
GTK tool bars can have just text, just images or images and text.
Customize `tool-bar-style’ to choose style. On a Gnome desktop, the default
is taken from the desktop settings.
GTK tool bars can be placed on the left/right or top/bottom of the frame.
The frame-parameter tool-bar-position controls this. It takes the values
top, left, right or bottom. The Options => Show/Hide menu has entries
for this.
The colors for selected text (the `region’ face) are taken from
the GTK theme when Emacs is built with GTK.
Emacs uses GTK tooltips by default if built with GTK. You can turn that
off by customizing x-gtk-use-system-tooltips.
A slew of GTK changes for people who use Emacs compiled with GTK.
New basic faces `error’, `warning’, `success’ are available to
highlight strings that indicate failure, caution or successful operation.
Good news for customize fans. No longer will you have to hunt down fifty different faces that do the same thing. Simply invoke M-x customize-face RET facename, where facename is one of the aforementioned three, and change the face there. The changes you make at the root faces will now properly propagate to all faces that inherit from them.
Lucid menus and dialogs can display antialiased fonts if Emacs is built
with Xft. To change font, use the X resource font, for example:
Emacs.pane.menubar.font: Courier-12
On graphical displays, the mode-line no longer ends in dashes.
Also, the first dash (which does not indicate anything) is just
displayed as a space.
Truly the end of an era.
Basic SELinux support has been added.
This requires Emacs to be linked with libselinux at build time.
Emacs preserves the SELinux file context when backing up, and
optionally when copying files. To this end, copy-file has an extra
optional argument, and backup-buffer and friends include the SELinux
context in their return values.
The new functions file-selinux-context and set-file-selinux-context
get and set the SELinux context of a file.
Tramp offers handlers for file-selinux-context and set-file-selinux-context
for remote machines which support SELinux.
Good news for hardcore SELinux hackers.
Changes for exiting Emacs
The function kill-emacs is now run upon receipt of the signals
SIGTERM and SIGHUP, and upon SIGINT in batch mode.
A nice change, for sure. However this will not trigger the usual “Do you wish to save your buffer” query as that is done at a higher level in Emacs. So don’t think this’ll save your skin if you nuke Emacs from the commandline!
kill-emacs-hook is now also run in batch mode.
If you have code that adds something to kill-emacs-hook, you should
consider if it is still appropriate to add it in the noninteractive case.
My kill-emacs-hook is mostly filled with miscellaneous gunk like saving my recentf and bookmarks.
Scrolling changes
New scrolling commands `scroll-up-command’ and `scroll-down-command’
(bound to C-v/[next] and M-v/[prior]) do not signal errors at top/bottom
of buffer at first key-press (instead move to top/bottom of buffer)
when `scroll-error-top-bottom’ is non-nil.
New variable `scroll-error-top-bottom’ (see above).
About time! This is useful if you make macros and you don’t want Emacs to cancel out of your macro recorder (or indeed in the macro itself) if it reaches the end of the buffer. I foresee lots of custom macro recorder functions that disable the top/bottom error signal when you’re recording…
New scrolling commands `scroll-up-line’ and `scroll-down-line’
scroll a line instead of full screen.
Pretty self-explanatory!
New property `scroll-command’ should be set on a command’s symbol to
define it as a scroll command affected by `scroll-preserve-screen-position’.
This is an elisp change that only really concerns people who write elisp packages. Basically, it means you can with greater granularity control if a particular movement command is affected by the user-defined setting in scroll-preserve-screen-position; itself a command that let’s you control how Emacs scrolls on your screen.
If you customize `scroll-conservatively’ to a value greater than 100,
Emacs will never recenter point in the window when it scrolls due to
cursor motion commands or commands that move point (e.f., `M-g M-g’).
Previously, you needed to use `most-positive-fixnum’ as the value of
`scroll-conservatively’ to achieve the same effect.
Interesting change, but not something I find all that useful. I rather prefer knowing where a line (or what have you) I want to see is in roughly the same spot every time.
``Aggressive'' scrolling now honors the scroll margins.
If you customize `scroll-up-aggressively' or
`scroll-down-aggressively' and move point off the window, Emacs now
scrolls the window so as to avoid positioning point inside the scroll
margin.
This is another one of those changes that people with rather interesting views on how Emacs should scroll care about. I personally just use the defaults so I don't have too much to say about it.
To Be Continued...
I think I'm going to end my post here. Emacs has hundreds of new features and changes and -- according to my modeline -- I have only covered about 30% of the NEWS file.
A little more than a year ago I set out to write a blog about the things I knew about Emacs. I had been collecting tips, concepts and snippets in a large org file for years with the goal of one day sharing what I knew with other people. I picked up Emacs when I was in University having heard about it before, but never really used it. All I knew about it was that it was really advanced, had meta keys and came with a built-in psychotherapist (try it — M-x doctor.)
When I mentioned this to someone more experienced in the world of editor-fu at the University’s computer science society, he told me that he used vim, and it was the editor of choice bar none, and that no other editor was worth using; that the modal concept was superior in every way to Emacs, and I was a fool for considering a fate worse than death if I defied his sage advice.
I immediately started using Emacs, and I have never looked back.
But I never had the opportunity to learn from an Emacs master myself. So when I finally sat down and wrote my first article a little more than a year ago I wrote it with the assumption that simply knowing what a keystroke does is not enough if you want to master Emacs; you must understand the why and not just the how. Thankfully, although Emacs has a steep learning curve, once you “get” it — you get it. And reaching that step is not an unattainable goal if you try to grasp the why. I remember when it finally “clicked” for me and I finally got how the editor functioned and why it functioned the way it did. Unfortunately, to achieve this level of enlightenment will involve a smidgeon of Lisp — and that’s probably what puts off most people. If you are an Emacs user and you are put off by elisp — overcome it and give it a shot. Elisp is arguably the most practical application of Lisp for most professional developers. You can write it, use it right away, and enrich your Emacs experience at the same time.
As I’ve written quite a few articles I have come to the realization that finding stuff on my site is, ah, not that easy. Therefore, I have created a reading guide to help newcomers find their way around.
Here’s to another year.
Overall Vimgolf Rank
Suppose you are trying to figure out what your overall rank is at Vimgolf(the sum of all the ranks you got), and as a Vim ninja you decided to use Vim to do the job. Your solution should work for every Vimgolf profile page(The actual input file for this challenge is my Vimgolf profile page). So, direct answer insertion is considered cheating.
Here’s another fun vimgolf challenge. The task is simple: get the sum of all the ranks on your vimgolf profile page.
A pretty specific goal, but the concept will work with anything involving summing numbers in a buffer. Anyway, off we go..
Given this input:
Generate Fibonacci Numbers - Rank: 1/66, Score: 20 Use your super vim powers to generate Fibonacci Numbers. Remove Accent off the Letter - Rank: 1/49, Score: 38 Remove all the accent from extremely accented statements. Remember FizzBuzz? - Rank: 1/118, Score: 53 Output FizzBuzz to 100. Start with nothing. (snip)
Get this output:
794
Now, in the vim challenge you are supposed to end up, as I understand it, with a buffer with nothing but the result — but that’s not a very productive (nor easy) way to accomplish the task.
Option 1: Using Registers
The first thing we need to do is mark the entire buffer with C-x h. Next, we need to set an emacs register — which is a simple store and recall register that you can save things like text, numbers, the point and frame or window configurations to for later re-use — by typing M-0 C-x r n a. To break that down a bit, we use the universal argument M-0 to tell Emacs that we want the number zero stored in a numeric register a by calling number-to-register, bound to C-x r n.
Next, we will use query-replace-regexp bound to C-M-% and use a little-known feature of Emacs’s regexp engine to, in effect, use the side effect of a regexp replace on each match.
Input the following:
Query replace regexp: Rank: \([0-9]+\) Query replace regexp with: \,(increment-register (string-to-number \1) ?a)
So here we are invoking increment-register (bound to C-x r +) and passing it the contents of the regexp capture group 1 (the rank) by first converting it to an integer so increment-register can make sense of it. We are also passing ?a as we need to give it the character a and not the string.
When you replace all matches (type !) you’ll see that Emacs will do a “rolling sum” by returning the contents of the new register and thus slowly add up all the ranks, one by one. Thus, if you scroll to the very bottom the last “match” will have the correct answer.
Obviously altering the text is probably not desirable so simply C-/ to undo the regexp replace.
And that’s it. Not the snappiest solution but it shows what you can do with Emacs if you really want to.
Option 2: Using Elisp
Well, if you can write elisp this problem is trivial; the code below will do the same as the above, but it’s a little bit neater and perhaps a bit easier to recycle for other tasks as well.
If you want to know how to run it, then you should read my article on Evaluating Elisp in Emacs.
(let ((counter 0)) (while (re-search-forward "Rank: \\([0-9]+\\)" (point-max) t) (incf counter (string-to-number (match-string 1)))) (message "Total sum %d" counter))
I’ve been having a lot of fun with vimgolf, a site where vim users try to complete certain tasks (sort this, format that, scrub this data, etc.) in as few keys as possible. It’s an interesting site and the problem set is varied and a good source of inspiration for Emacs blog posts. I learned about this site from Jon’s Emacs blog and I’ve been having a fun time solving the problems in Emacs.
It’s obvious to most of us that to attempt a “fewest possible keystrokes” exercise in Emacs will invariably end in tears, as we can’t compete against vim in that regard; but with that said, we do have a lot of tricks up our sleeve.
In reading — or attempting to, more like — the “solutions” I’ve made a few observations about most of them: they are tailored to the input, and that input only; thus, it wouldn’t necessarily scale to larger problem sets or even smaller ones. I’m going to try to solve these problems so they’re technically correct — the best kind of correct — and thus work with larger or smaller input.
Alphabetize the Directory
Put the contacts and their information in alphabetical order.
So the task is simple. We need to sort by a key (the name heading), preserving the details associated with each contact.
The easiest way to solve this is to imagine the input below as an org-mode file, so that’s what we’ll do:
M-x org-mode and org mode immediately picks up on the file contents and collapses the headings.
Next, we need to sort the items, whilst we preserve the order of the subitems, and this is done by first moving to the heading level you want to sort by (in this case it’s the Smith, George heading) and invoking M-x org-sort and selecting a for alphabetic sort.
Org mode will sort the file and the output should match the expected vim golf output. And that’s it.
Input
## Directory
* Smith, George
- 1234 Avenue, New York, NY
* Bailey, Stephen
- 4545 Harbor Ln, Atlanta, GA
- Phone: (412) 291-1238
* Thomas, Brent
- 1482 Mystic Oaks, Anaheim, CA
* Grant, Jenny
- 198 Circle Dr, Houston, TX
- Phone: (213) 198-9842
- Email: jenny.grant@gmail.com
* Percy, Adam
- 221 Jaguar Pkwy, Missoula, MT
* Garfield, Misty
- 5988 Apple Tree Ln, Memphis, TN
* Parsons, Betty
- Phone: (235) 523-2378
- Email: bettyboop123@gmail.com
* Sanders, Terry
- Email: sanderst@yahoo.com
* Smith, Pete
- Phone: (294) 984-2938
* Frost, Jennifer
- 2498 Temple Terrace, Miami, FL
* Matthews, Frank
- 418 Happy Trail, Phoenix, AZ
- Phone: (985) 129-2394
- Email: frank@phoenixrealtors.net
* Allen, Taylor
- Email: allen.taylor@hotmail.com
* McCullen, Sarah
- 247 Hidden Elm, Seattle, WA
- Phone: (288) 283-4568
* Perkins, Mike
- 992 Peartree Ln, Bowling Green, KY
Expected Output
## Directory
* Allen, Taylor
- Email: allen.taylor@hotmail.com
* Bailey, Stephen
- 4545 Harbor Ln, Atlanta, GA
- Phone: (412) 291-1238
* Frost, Jennifer
- 2498 Temple Terrace, Miami, FL
* Garfield, Misty
- 5988 Apple Tree Ln, Memphis, TN
* Grant, Jenny
- 198 Circle Dr, Houston, TX
- Phone: (213) 198-9842
- Email: jenny.grant@gmail.com
* Matthews, Frank
- 418 Happy Trail, Phoenix, AZ
- Phone: (985) 129-2394
- Email: frank@phoenixrealtors.net
* McCullen, Sarah
- 247 Hidden Elm, Seattle, WA
- Phone: (288) 283-4568
* Parsons, Betty
- Phone: (235) 523-2378
- Email: bettyboop123@gmail.com
* Percy, Adam
- 221 Jaguar Pkwy, Missoula, MT
* Perkins, Mike
- 992 Peartree Ln, Bowling Green, KY
* Sanders, Terry
- Email: sanderst@yahoo.com
* Smith, George
- 1234 Avenue, New York, NY
* Smith, Pete
- Phone: (294) 984-2938
* Thomas, Brent
- 1482 Mystic Oaks, Anaheim, CA
Interacting with external commands in Emacs ranges from simple one-off commands to launching interactive sessions using comint, the Emacs command interpreter library. In earlier articles I covered shell integration and terminal emulation in Emacs, but this time I want to talk about the other use case: executing shell commands and capturing their output; how to use them in elisp code; how to make them work with unicode and different coding systems; and a few unique ways of using the commands.
read more…