Related
I'm aware of the use of lisp funcitons window-configuration-to-register and jump-to-register, and also packages like desktop.el and winner-mode.
These can all be used to save a layout during a session and recover it sometime later. I can get these to work nicely.
However if I'm running in daemon mode and am attached via emacsclient - if I try save a layout to a register, for example - I do not seem to be able to recover this layout in a separate emacsclient on the same daemon using any of the above methods - even though they share exactly the same windows.
Can anyone suggest a lisp function or any other methodology to allow to persisting of a client window layout after closing the original emacsclient session?
I should add I have no option but to use use emacs in console.
Update:
To half-answer my own question -
Saving the frame configuration to a register seems to work:
C-x r f – frame-configuration-to-register
C-x r j – jump-to-register
It's a bit shakey - when you jump back to the register you seem to have to exit the frame using 'C-x 5 0'.... after that you're still left with an underlying client frame that you can exist using 'C-x C-c' without killing the daemon.
If you try to directly exit the register stored frame using other methods it will either not work at all, or you'll kill the daemon entirely!
So it looks to me like jumping to the register seems to kinda grab ownership of the daemon process.
So follow-up questions:
1) Is there a way to make this play a bit nicer with the daemon?
2) Is there a way to automatically save your frame to a register (which will persist on the daemon) when emacs client exits or dies?
2nd Update
Actually having used it for a day - the above doesn't really work at all - the behaviour is random when restoring windows - sometimes it works, and it is often impossible to quit the emacsclient using any command!
I often find myself running kill on the emacsclient - which in-turn kills the daemon process as well as the client!
To answer my own question and having tried many ways to achieve the above, I've only found one method that at least so far has proved stable/reliable even when using the daemon - this is to use the workgroups2 pacakge that is available on Melpa.
I also noted that the original workgroups package (of which workgroups2 is a more recent fork) also works, but has more limited funcitonality. This is also available from Melpa.
The key default commands on workgroups2 that will interest you are:
C-c z c - create
C-c z v - view
I refer you to the manuals for more details - I've included reference links below to the 2 packages
https://github.com/tlh/workgroups.el
http://workgroups2.readthedocs.io/en/latest/
I use Emacs v. 22 (the console version, either remotely with PuTTY or locally with Konsole) as my primary text editor on Linux. It takes a while to load up each time I start it though, probably almost a second, although I never timed it. I tend to open and close Emacs a lot, because I'm more comfortable using the Bash command-line for file/directory manipulation and compiling.
How can I speed up the start-up time?
Others have covered using gnuserve and emacsclient, and I'd suggest compiling within emacs (being able to jump to compilation errors is a win).
But, specifically speeding up the .emacs can be done by:
Byte compiling the .emacs file, which you can do automatically by using this snippet of code
Replacing as many of the (require 'package) statements with autoloaded functionality. This will delay loading of lisp until it's actually required. Using this technique allowed me to speed up my startup from >6 seconds to <1. This takes a little bit of work because not all libraries come properly marked autoload.
Removing code/functionality you no longer use.
Try running emacs with the option --no-site-file to avoid loading unnecessary packages in the site installation site-start.el.
If you are really serious, you can roll your own emacs with your favorite functionality already loaded. This, of course, means it's more involved to make changes to what you have in your .emacs because it's a part of the binary. Follow the link for information on how to use dump-emacs.
Buy a faster computer and/or faster disk.
How to determine what your .emacs loads
Now, how do you find out what your .emacs loads? With the goal to remove the functionality, or to delay it? Check your *Messages* buffer, which contains lines like:
Loading /home/tjackson/.emacs.tjackson.el (source)...
Loading /home/tjackson/installed/emacs/lisp/loaddefs.el (source)...done
Loading /user/tjackson/.elisp/source/loaddefs.el (source)...done
Loading autorevert...done
Loading /home/tjackson/.emacs.tjackson.el (source)...done
If you'll notice, the Loading statements can nest: the first .emacs.tjackson.el ends with ... and the last line shows the .emacs.tjackson.el load is ...done. All those other files are loaded from inside my .emacs.tjackson.el file. All the other loads are atomic.
Note: If you have a large .emacs, it's possible that the *Messages* buffer will lose some of the messages because it only keeps a fixed amount of information. You can add this setting early on to your .emacs to keep all the messages around:
(setq message-log-max t)
Note: It the 'load command will suppress the messages if its fourth argument nomessage is non-nil, so remove any such invocations (or, advise 'load and force the fourth argument to be nil).
In addition to Adam Rosenfield's solution, I recommend to use Emacs in server mode. You may add (server-start) to your dotemacs, and run emacsclient instead of emacs whenever you want to open file in Emacs. That way you have to pay the loading cost of Emacs only once, after then clients pop up immediately.
Edit
You're right, v22 does not create a new frame. Create a shell script that do the trick:
#!/bin/bash
# Argument: filename to open in new Emacs frame
/usr/bin/emacsclient -e '(let ((default-directory "`pwd`/")) (select-frame (make-frame)) (find-file "'$1'"))'
Edit 2
In v24+, you can do emacsclient -c to create a new frame.
Don't close Emacs every time you want to use the shell. Use Ctrl-Z to move Emacs to the background and the fg command in Bash to move it back to the foreground.
A couple of tips:
Use autoloads
Using autoload saves you from loading libraries until you use them.
For example:
(if (locate-library "ediff-trees")
(autoload 'ediff-trees "ediff-trees" "Start an tree ediff" t))
Compile your .emacs
Gives you a slight speed increase although there are pitfalls if you
work with version control and your .emacs is newer than .emacs.elc.
One common trick is:
(defun autocompile nil
"compile itself if ~/.emacs"
(interactive)
(require 'bytecomp)
(let ((dotemacs (file-truename user-init-file)))
(if (string= (buffer-file-name) (file-chase-links dotemacs))
(byte-compile-file dotemacs))))
(add-hook 'after-save-hook 'autocompile)
Learn to love emacs server.
Running emacs as a server means never having to close it down. However
I note your still using emacs22. emacs23 supports multi-tty which makes
it a lot easier to run emacs in one screen session and then bring up
new windows in another terminal. I use emacs to edit mail for my mail
client (mutt) and emacsclient is fantastic for these sort of quick edits.
One of
M-x shell
M-x eshell
M-x term
M-x ansi-term
should meet your command-line needs from within Emacs.
You can also use M-! (aka M-x shell-command) to execute a one-liner without dropping to the shell.
Check your .emacs file to see if you're loading unnecessary packages. Loading packages can take a significant amount of time. For example, you might only want to load the php-mode package if you're editing a PHP file. You can do that by installing a hook procedure, although I'm not certain of the details.
Also make sure that any packages you're loading are compiled (.elc files). You can compile an elisp file by running
emacs -batch -f batch-byte-compile thefile.el
Compiled packages load much faster than uncompiled packages.
"I tend to open and close emacs a lot, because I'm more comfortable using the bash command line for file/directory manipulation and compiling."
You're describing the way an editor like vim is used like. Shoot in&out. Emacs is usually kept open, and mostly all is done from "within it". hiena already answered what would be the correct approach here.
The fastest way is to profile your .emacs. I cut down my load time from >3s to 1s in 5 minutes after I found that 4 particular lines in my .emacs were taking up more than 80% of the load time.
One thing that helped me reduce the load time of my .emacs, in addition to autoload (as others have suggested), is eval-after-load. In the following example, delaying the call to sql-set-product saves you from having to load sql in your .emacs, making the exisiting sql autoloads more effective.
(eval-after-load "sql"
'(progn
(sql-set-product 'mysql)
(setq sql-mysql-options '("-C" "-t" "-f" "-n"))
(setq sql-sqlite-program "sqlite3")
))
Of course, for some packages there will be a hook available that you can do the same thing, but sometimes there isn't, or else this way just proves easier to think about.
Emacs is designed to run "all the time" (or at least for long periods of time), thus starting and stopping Emacs several times during a day is not recommended.
I would suggest using screen. Screen is a terminal multiplexer, giving you an unlimited virtual terminals in one terminal.
After installing simply write "screen emacs" in your terminal. Emacs will start as usual, but pressing "c-a c" (that is press ctrl-a and then c) will open a new virtual terminal. You can get back to emacs by pressing "c-a c-a" (that's two times ctrl-a).
You can even detach from the running screen session, the key sequence is "c-a d".
Re-attach to the session by issuing "screen -R" and you will be back where you left. This enables you to start an emacs session at work, detach, go home, and re-attach from home.
I've been running Emacs like this for months in a row.
Here's the official web site: http://www.gnu.org/software/screen/ but try googling for screen tutorials and howtos
You can use benchmark-init to profile your Emacs startup. It will keep track of what modules are being loaded and how much time is spent on each. The results can be presented either in a tabulated form or as a tree. The tree makes it easier to track who loads what, which can be helpful when you load a package with a lot of dependencies, and the tabulated form helps you quickly find where most of the time is being spent.
Once you have these results try to figure out if all of the modules have to be loaded all the time or if you can perhaps load some of them on-demand. For instance, in my configuration I only load Emacs extensions that are specific to certain modes when that mode is actually activated since most of the time I only use a small subset of them in a session. eval-after-load and mode hooks will be your friends here.
By applying this method my Emacs starts in 3-4 seconds and I have close to 200 extensions installed. Most of the time is spent loading Helm, which I always load since it replaces find-file and other core functions that are always needed, and CEDET, since I use the latest version and it has to be loaded before Emacs tries to load the older built-in version.
Try using the https://github.com/jwiegley/use-package macro to define your package loads and customizations. It handles deferred loading of packages for you, making it relatively easy to get good startup times even in the presence of large numbers of configured packages. I have almost 100 packages referenced in my .emacs, but my startup time is under 2 seconds on Linux, and 2.2s on the Mac.
One thing that others haven't mentioned is to include the elisp libraries you use as part of the dumped Emacs to move the library loading time from Emacs startup to Emacs build. It is not for the faint-hearted, but if you load several libraries in .emacs it could win you a few seconds of startup time.
I had around 120sec start time. I was able to find the fix installing this:
https://github.com/dholm/benchmark-init-el
put on top of your init.el
(let ((benchmark-init.el "~/.emacs.d/el-get/benchmark-init/benchmark-init.el"))
(when (file-exists-p benchmark-init.el)
(load benchmark-init.el)))
then once your emacs started, run:
M-x benchmark-init/show-durations-tree
On my side the problem was 127 secs in tramp-loaddefs
I fixed it by adding
127.0.0.1 host.does.not.exist
to /etc/hosts and that made my startup fast
see more here: https://github.com/emacs-helm/helm/issues/1045
another thing that maybe helpful to you: https://www.emacswiki.org/emacs/ProfileDotEmacs
This doesn't answer the question, but is kind of relevant
I don't know how to make it start faster, but there are a few things I could suggest:
for most things you do on the command line, you can do them in emacs:
compile: M-x compile, then type the command you use
my experience is only with C++, but with g++ you can press C-x ` to jump to lines that the compiler complains about
run shell commands: M-!, dumps output into a buffer
interactive shell: M-x shell
alternatively, you could run emacs like this:
emacs file.ext &
which opens emacs in the background so you can still use the shell ( this works best with putty and X forwarding with something like Xming)
I was trying to solve the same problem, when I came across this question here. I just wanted to add that the problem for me was not because of the load time of emacs lisp packages, but the fact that the host did not have a fully resolved hostname
To check your package load time do
M-x emacs-init-time
For me it was 0.3 seconds, and yet the load time was extremely high.
After changing my hostname correctly, it fixed the problem.
To configure your fully resolved hostname edit /etc/hostname, and /etc/hostsfile with:
127.0.0.1 localhost localhost.localdomain
192.168.0.2 hostname hostname.domain
I would have to check my customization, but there is a package called gnuserve or emacsclient. It migrates a lot so you will have to google for it.
It runs one emacs session in the background. Any further sessions of emacs are essentially just new frames of that session. One advatage is quick startup times for your later sessions.
I've been fooling around with comint-mode lately and I'm noticing some weird behaviors. Its very poorly documented, so I'm wondering if anyone has any insight on this.
In some modes, comint-send-string causes whatever is sent to be inserted into the comint buffer and then sent to the associated process, whereas in others, the input is send directly to the process without being placed into the buffer. For example, do run-python with the new (24.3) python.el and then do (comint-send-string "*Python*" "x=3\n"), the string x=3 is inserted into the buffer and then executed. If you do M-x shell, however, and then (comint-send-string "*shell*" "x=3\n"), no text is inserted into the buffer, the input is simply sent to the shell process directly to be executed.
Does anyone know why this difference in behavior exists or how I can change it?
I observe identical behavior on linux (emacs-version == "24.3.50.7", both GUI and emacs -Q -nw): neither
(comint-send-string "*Python*" "x=3\n")
nor
(comint-send-string "*shell*" "x=3\n")
insert anything in the comint buffer (i.e., the next prompt appears
right after the previous prompt - without even a newline between them).
I eventually figured it out. For some reason the system python on OSX causes this behavior, installing python from homebrew fixed it.
I've found that terminal emacs does not render the correct colors unless I explicitly set TERM=xterm-256color. I use gnome-terminal, and from what I understand, TERM should be set to gnome-256color. Similarly, I tend to use tmux a lot, which advises against any TERM setting other than screen-256color. Unfortunately, both of those settings (within their respective context - gnome-terminal or tmux) result in emacs having wrong colors, whereas vim displays colors correctly. However, if I export TERM=xterm-256color, the colors work just fine in emacs.
Can anyone explain what's going on, or offer a solution?
Update
Here's what I'm dealing with:
I can get the colors to look correct in the terminal by adding the following to my init.el:
(defun terminal-init-gnome ()
"Terminal initialization function for gnome-terminal."
;; This is a dirty hack that I accidentally stumbled across:
;; initializing "rxvt" first and _then_ "xterm" seems
;; to make the colors work... although I have no idea why.
(tty-run-terminal-initialization (selected-frame) "rxvt")
(tty-run-terminal-initialization (selected-frame) "xterm"))
This feels really, really wrong though. There has to be a logical explanation for this...
P.S.
I have very little knowledge of terminfo and the precise role that $TERM plays in the process of color terminal behavior. If it's safe to always use xterm-256color (even when $TERM "should" be gnome-256color or screen-256color), I'll go with that.
Maybe I'm not understanding something, buy why don't you run emacs like this:
TERM=xterm-256color emacs -nw
This way Emacs has its own TERM setting that you know works. You can also make an alias or wrap this in shell-script.
Terminals are a special type of device. When a process sends special byte sequences (called control sequences) to the terminal, it performs some action (like cursor positioning, change colors, etc).
You can read the ANSI terminal codes to find more detail about control sequences.
But terminals come from 70s, when hardware was limited in its capabilities, and a terminal cannot provide info about its capabilities (ie. which sequences it supports).
$TERM was used to resolve this issue - it allows programs to know what to send to the terminal to get the job done. termcap and terminfo are databases that store info about terminal capabilities for many $TERM names. If your $TERM is not in the db, you must ask an administrator to add it.
All terminal emulators inherit these limitations from old hardware terminals. So they need a properly set $TERM, and the terminfo/termcap DB MUST have data for this terminal. When a virtual terminal starts it sets the $TERM variable for you (and inside programs like bash). If $TERM is not in the terminfo/termcap you can quickly define an alias from $TERM to xterm-256color (you can find examples in the termcap file on how to do that).
This behavior has to do with the logic EMACS uses to determine whether the terminal background is dark or light. Run M-x list-colors-display with TERM set to either xterm-256color or screen-256color and you'll see that the exact same colors are listed. As you pointed out in the comments, the difference in color schemes that you've observed is due to the frame background mode. To see this, with your TERM set to screen-256color, compare the colors in
emacs -Q -nw --eval "(setq frame-background-mode 'light)"
and
emacs -Q -nw --eval "(setq frame-background-mode 'dark)"
The function frame-set-background-mode (in frame.el) checks to see whether the terminal type matches "^\\(xterm\\|\\rxvt\\|dtterm\\|eterm\\)" if it can't deduce the background color otherwise.
Within a running session, you can change the color scheme to 'light by evaluating
(let ((frame-background-mode 'light)) (frame-set-background-mode nil))
I am not that familiar with how emacs handles different terminals exactly. But looking at lisp/term directory in emacs sources, I found out that the existence of a function terminal-init-xxx allows you to add support for different terminals. For example, I've got:
(defun terminal-init-screen ()
"Terminal initialization function for screen."
;; Use the xterm color initialization code.
(xterm-register-default-colors)
(tty-set-up-initial-frame-faces))
in my .emacs, which adds support for screen-256color. You may try defining a similar function for gnome by renaming the above function to terminal-init-gnome.
NOTE: If you are interested, you can try to track down the calls from tty-run-terminal-initialization code. It first gets the terminal type using tty-type function, then looks at certain locations to load a relevant terminal file, then tries to locate the matching terminal-init-xxx function, and finally calls it. It may help you figure out the correct name for gnome-terminal.
It looks like unless your TERM indicates that your terminal has 256 colors, emacs will only use 8. Changing TERM to gnome-256color allowed the color registration functions to work.
There is a way to cheat, after all. When I run gnome-terminal, my terminal is set to xterm by default. Instead of changing TERM variable, it is possible to redirect xterm to another terminal, say, gnome-256color. Simply create the directory $(HOME)/.terminfo/x, then run ln -s /usr/share/terminfo/g/gnome-256color ~/.terminfo/x/xterm. I think this is better than setting TERM manually in .bashrc, because it only redirects a particular terminal to something else. A console login would still leave TERM as linux, and not xterm-256color.
Add this to your ~/.emacs:
(add-to-list 'term-file-aliases
'("st-256color" . "xterm-256color"))
It tells emacs that if it sees TERM=st-256color then it should initialize the terminal as if it had seen TERM=xterm-256color.
Longer answer:
Emacs is showing strange colors because it thinks your terminal can only support 8 colors. In Emacs, run M-x list-colors-display to see the colors it thinks are available. The correct number of colors is detected during terminal-specific initialization. It says, in part:
Each terminal type can have its own Lisp library that Emacs loads when run on that type of terminal.
On my machine, the terminal-specific initialization files are in /usr/local/share/emacs/25.*/lisp/term. It has files for xterm, rxvt, screen, etc. but nothing for st. We need to help Emacs find the right initialization file. The documentation further says:
If there is an entry matching TERM in the term-file-aliases association list, Emacs uses the associated value in place of TERM
So that association list is a recommended way to handle unknown terminals. It works without you having to manually override the TERM environment variable.
On ubuntu 10.04 I too had noticed that running emacs -nw inside byobu/tmux/screen was using different colours from emacs -nw in the regular gnome-terminal.
I found that this is because byobu was setting TERM to screen-bce. Then setting TERM to xterm (for me, in the normal gnome-terminal TERM=xterm) gave me the same syntax highlighting when not running through byobu/screen.
So still not sure what the proper solution is.
See also this post:
Emacs Python-mode syntax highlighting
When I start Emacs, its windows and its minibuffer start up floating on the main frame, leaving a lot of empty space (see: here and here). Any idea of what is going on and how to fix it?
I'm using emacs version 23.1.1 on Xmonad
Thanks!
I get the same thing occasionally when doing a TAB expand in shell mode. I don't end up with an empty mini-buf, it usually has all the contents from the shell expand. The only solution I've found so far is to kill the current shell buffer
You might try C-l (recenter-top-bottom). I've used this over the years as a pretty reliable "repaint" operation that will clear up various detritus that occurs on the frame. It is even more helpful when you do not have a window system.
It looks like an odd interaction between Emacs and your window manager. It's as if Emacs is configuring itself to have a certain number of lines and columns in its active window, but the frame the window manager has allocated it is larger.
Can you try running X without Xmonad (ie. just run startX) and then run Emacs? You won't get frame decorations, but you can still use xterm to start it and see what 'natural' size it gets allocated. With the same Emacs configuration, that would narrow it down to some issue with Xmonad.
It's not obvious from the screenshots, as one is definitely in GUI mode, but the other (green on black) looks like it is in text mode. If you force it to run in text mode (emacs -nw), does it take up the same size? Within the shell, do normal commands take up the entire size of the allocated window?
Are the shell variables COLUMNS or LINES set correctly? Do they match the window dimensions? Assuming you are running bash, is shopt checkwinsize turned on? Do regular commands such as ls wrap at the expected margins and scroll to use the entire window?
It is worth ensuring that text mode is being configured correctly before getting GUI mode to work. I have a hunch the way that Xmonad is managing its window sizing may be the culprit, if it is 'forcing' windows to tile in certain arrangements, and the correct X hints aren't being sent. But - that's all just supposition for now...
This partially works (taken from here):
(defun fullscreen ()
(interactive)
(set-frame-parameter nil 'fullscreen
(if (frame-parameter nil 'fullscreen) nil 'fullboth)))
(add-hook 'emacs-startup-hook 'fullscreen)
The windows get corrected, although the mini-buffer is still larger than the usual.
Sounds like a bug, to me. Consider reporting it: M-x report-emacs-bug. Try to provide a recipe starting from emacs -Q.