indentation problem

[c02y]

This fixes the problem of indentation from https://github.com/wwwjfy/emacs-fish, so please upload this to melpa so I can install/update using package.el

[c02y]

It seems I made a wrong conclusion, it didn't fix my problem of indentation.

Put this in config.fish:

set -gx LESS_TERMCAP_me \e'[0m'     # turn off all appearance modes (mb, md, so, us)
set -gx LESS_TERMCAP_se \e'[0m'     # leave standout mode
set -gx LESS_TERMCAP_ue \e'[0m'     # leave underline mode
set -gx LESS_TERMCAP_so \e'[01;44m' # begin standout-mode – info
set -gx LESS_TERMCAP_mb \e'[01;31m' # enter blinking mode
set -gx LESS_TERMCAP_md \e'[01;38;5;75m' # enter double-bright mode
set -gx LESS_TERMCAP_us \e'[04;38;5;200m' # enter underline mode

You'll notice the problem of indentation after the second line of the snippet.

[d12frosted]

Sorry for late response - was in travel mode. My changes are part of emacs-fish. So it's expected that we have the same problems :smile:

But good thing - I am able to reproduce your issue and will update package soon. In short - the problem comes from the fact that begin is not ignored if it's inside of comment.

Thanks for the finding!

[c02y]

Thanks, you can close this.

[c02y]

I found that once delete one space in one comment line, the indention problem will be gone, the original code snippet:

set -gx LESS_TERMCAP_me \e'[0m'     # turn off all appearance modes (mb, md, so, us)
set -gx LESS_TERMCAP_se \e'[0m'     # leave standout mode
set -gx LESS_TERMCAP_ue \e'[0m'     # leave underline mode
set -gx LESS_TERMCAP_so \e'[01;44m' # begin standout-mode – info
set -gx LESS_TERMCAP_mb \e'[01;31m' # enter blinking mode
set -gx LESS_TERMCAP_md \e'[01;38;5;75m' # enter double-bright mode
set -gx LESS_TERMCAP_us \e'[04;38;5;200m' # enter underline mode

the indention of all the lines after the 4th line will be wrong. if I change the 4th line to:

set -gx LESS_TERMCAP_so \e'[01;44m' #begin standout-mode – info

Note the deleted space before begin, then the indentation will be normal.

[d12frosted]

Note the deleted space before begin, then the indentation will be normal.

Yeah, it's because of how special keywords are parsed. Fish-mode doesn't treat #begin the same way as begin.

P. S. Sorry for solving this issue for so long.

[c02y]

The indentation problem is still there.

Note that I didn't use

(add-hook 'fish-mode-hook (lambda ()
                        (add-hook 'before-save-hook 'fish_indent-before-save)))

in my init.el, I got my own indent-buffer(basciall I use (indent-region (point-min) (point-max) for this function) before-save-hook, it will not work perfectly (I'll explain later) even I put your two lines, the point or the window will jump around several lines after saving the buffer, the problem is the fish_indent function of your code. I can put several lines of code inside my before-save-hook or totally comment my part out and use yours for fish-mode file to make indentation right, but it is not perfect thought (I'll explain later), and I feel ridiculous to add several lines of code or your yours code for just fish-mode file .

---

Another problem is this single line in my conf.fish:

alias tt 'tmux switch-client -t'

will make the indentation of all the lines after this line wrong, exactly like the indentation I mentioned about the # begin earlier. I cannot figure out what is the problem of that line, if I comment it out, the problem will go wrong.

As for the # begin problem, even thought I make #begin, the updated fish-mode source code will not fix it any more, it seems the two bugs were made after I finally found the solution (make it #begin), right now I just deleted the begin word and put the alias line at the end of my conf.fish file to make it right, maybe because I use my own indent-buffer before-save-hook instead of the two lines of code you recommended.

Any way, you should not define another function to indent the fish-mode file, just let my (indent-region (point-min) (point-max) do the work.

[d12frosted]

Thanks for posting more information about problems with indentation. Actually, your cases show that current implementation is easily broken and far from ideal.

Sorry for not fixing those problems. I am not sure when, but I'll try to fix them soon.