Hide implementation buffers from buffer switching about polymode

[doublep]

To make it even worse, this changes the stack order of buffers, so that C-x b RET no longer switches to the "previous" buffer. At this point polymode is unusable for me.

[srnnkls]

If you, like me, do use ivy-switch-buffer and friends you can go bruteforce with this regex:

(add-to-list 'ivy-ignore-buffers "\\[.+\\]")

EDIT (Ido version):

(setq ido-ignore-buffers '("\\` " "\\[.+\\]"))

[doublep]

Actually, just found this in Elisp manual:

Buffers that are ephemeral and generally uninteresting to the user have names starting with a space, so that the ‘list-buffers’ and ‘buffer-menu’ commands don’t mention them (but if such a buffer visits a file, it is mentioned). A name starting with space also initially disables recording undo information; see *note Undo::.

[vspinu]

Hiding those buffers is a non-trivial change. I have in my agenda to re-start working on polymode this week after a long break. This issue is a priority.

this changes the stack order of buffers, so that C-x b RET no longer switches to the "previous" buffer.

What change do you mean more concretely? I don't understand what you mean.

Polymode shifts only its own "non-interesting" buffers at the end of the buffer stack. I am using ido and the most relevant buffer is always in front. I don't understand how a couple of more buffers in ido menu make polymode unusable.

[doublep]

Well, maybe it tries to shift, but fails. Tried both polymode from melpa and git head. I can just:

• C-x f file1.jsp
• C-x f file2.jsp

Now C-x b shows 'file1.jsp[js]' as the buffer it will switch to by default. This seems to be OK, since switching to that buffer effectively switches to 'file1.jsp', but in the course of my normal work I had worse cases where it would switch to something different entirely.

I also tried to change variable 'pm--ib-prefix' in the source code from empty string to one space, but that broke the mode entirely.

[ashiklom]

Loosely related, but I am currently using a function like this to kill all polymode buffers for the current buffer (useful when switching to plain markdown-mode). Perhaps a similar approach can be used as advice (?) for the buffer list.

(defun ans/kill-polymode-buffers ()
  "List all polymode implementation buffers."
  (interactive)
  (let* ((b (buffer-name))
         (pattern (concat b "\\[.+\\]"))
         (matched-buffers (-filter (lambda (buf) (s-matches? pattern (buffer-name buf))) (buffer-list)))
         (n-matched (length matched-buffers)))
    (-map #'kill-buffer matched-buffers)
    (message "Killed %d polymode buffers." n-matched)))

[mavit]

We can make that regular expression a little bit tighter, since the part in square brackets is always at the end of the buffer name.

(add-to-list 'ivy-ignore-buffers "\\[.+\\]\\'")

[srnnkls]

Good one. Just out of curiosity:

Is it more appropriate to use \\' than simply $ here?

[mavit]

Marginally. The following is a valid (if foolish) buffer name, since buffer names can contain newlines:

foo[bar]
baz