Emacs: help me understand file/buffer management


I've been using emacs for all of my text editing needs for the past three years now. When I'm in a single file, working on code or whatnot, I'm fairly efficient. With two files, I can "C-x b RET" between them and I do fine. However, whenever I need to be working on more than two files at a time, I tend to get lost.

Here are some of the problems that I'd like to work on:

  1. I forget what some of my buffers are called, but I don't understand why C-x C-b splits my window into two buffers and exits the mini buffer. Sure I can switch buffers and choose a buffer to visit, but this feels unintuitive, and leaves me with two buffers open.

  2. When I visit a directory rather than a file, I have a convenient list of all of the files and directories. I usually want to do one of two things with this: 1) Open a single file and never see this buffer again OR 2) Open a bunch of files and never see this buffer again. I don't really know how to do this, as moving the point to a file and hitting return doesn't do either of these things.

  3. I know that my buffers aren't like tabs, but I have an inclination to want to scroll through them to find what I want. I don't know of any key-bindings for this, but I'd like it to be M-n / M-p or the like. Then again, this may be a horribly inefficient way to switch buffers.

  4. When I open interactive help of any kind (for example in ESS), I have a habit of switching back to the buffer I was working in and using C-x 1 to get back to a single buffer. When I do this, however, the help buffer hangs around in my buffer list, further confusing me. I know I can switch to that buffer, kill it, switch back, and then go back to a single buffer, but this seems wrong.

The way I've dealt with this so far involves using a tiling window manager and a few emacs windows in different work-spaces, rather than actually learn the best way to manage a number of files in emacs. I don't necessarily want to change emacs to better fit my needs (although I am open to that if it fits in with what I'm about to say), instead, I'd like to grok the thought process behind handling files/buffers the way that emacs does, and how I can be more efficient with it.

Any answer that would help me understand the correct way, or a more efficient way to manage my buffers or files would be greatly appreciated.

Best Solution

  1. Bind C-x C-b to ibuffer. This is a better buffer listing facility with many advanced features, and its default behaviour is to replace the current buffer with the buffer listing, and then bury the listing when you select a buffer (leaving you with the newly-selected buffer in place of the original one).

    You can simply use C-x b to enter your selection in the mini-buffer, of course; however the tab-completion (which is needed to make this a viable option, IMO) does open a new window temporarily, at which point I think you might as well familiarise yourself with something with more features.

  2. Use a instead of RET when selecting from dired. This kills the dired buffer instead of leaving it behind. C-h m in any buffer will show you the help for its major mode (followed by help for the minor modes), and you can read about all the available dired key bindings there.

  3. http://www.emacswiki.org/cgi-bin/wiki/TabBarMode ? (edit: I prefer RĂ©mi's answer for this one, but TabBarMode would give you the visual tab element if you were particularly keen on that.)

  4. q is bound to a 'quit' function in a great many major modes. Generally it buries the buffer rather than killing it, but I certainly find that fine.

To elaborate a little on #1, ibuffer has lots of nice features, and M-x customize-group ibuffer RET will give you some idea of how you can customise it to your liking.

Furthermore, you can filter the buffer list by many criteria (again, use C-h m to see its help page), and then generate a 'group' definition from the current filters, and save your custom filters and groups for future usage.

For example:

  • / f ^/var/www/ RET: filter buffer list to show only filenames starting with /var/www/.
  • / s Web filters RET: name and save active filter set to your init file.
  • / g Web development RET: create a named group from the active filters.
  • / S My groups RET: name and save group definitions to your init file.
  • / r Web filters RET: invoke the "Web filters" filters.
  • / R My groups RET: invoke the "My groups" groups.
  • RET on a group name to collapse or expand it.
  • C-k and C-y to kill and yank groups, to re-arrange them.
  • C-h m for more information...

This way you can have a single Emacs instance running, and create filters and groups for different types of task, and easily switch between them.