0

Sometimes Emacs stops to respond to key bindings, for example:

  • it's impossible to C-x C-s save buffer
  • it's impossible to C y yank killed text
  • when mark activated it's invisible

and many other disadvantages.

The only way to return normal behaviour is to relaunch Emacs, though such odd behaviour can occur again very soon. Also, switching buffer to and fro could help a bit, at least it is possible to write buffer.

What is possible reason of this breakage?

Remark: Please review. I've tried to simplify question, but if this still is not good enough, let me know and I'll delete it.

Geradlus_RU
  • 1,466
  • 2
  • 20
  • 37
  • This is pretty broad for SO. You might have better luck on https://emacs.stackexchange.com/. – ChrisGPT was on strike Mar 22 '15 at 14:21
  • Thanks, @Chris, I was not aware about that subsite! – Geradlus_RU Mar 22 '15 at 16:29
  • It's fairly new, and still in beta, but already a pretty good community. I hope you find some help over there. – ChrisGPT was on strike Mar 22 '15 at 16:33
  • You need to narrow things down - find out which part(s) of your init file cause the problem. You can use `comment-region` to comment out 1/2, then 3/4, 7/8, 15/16, 31/32... of your init file, to quickly find what is causing the problem. After you have done that, if you still have a (hopefully specific) question then pose that here. This question should be closed, as you are essentially telling us to debug a giant sack of stuff that you load and use somehow, and without even a recipe to repro the problem. – Drew Mar 22 '15 at 20:04
  • @Drew, I haven't understood the humor about fractions, this issue occurs spontaneously, and I can't be certain when will it occur next time. I hope maybe someone faced similar issue, and can share some useful information. Thanks. – Geradlus_RU Mar 23 '15 at 04:24
  • No humor was intended. If the issue is sporadic then what I suggested might not help. If it is systematic, what I'm suggesting is to **recursively bisect** your init file, commenting out larger and larger pieces, until you find the culprit code. `comment-region` comments a block of code, and with `C-u` it uncomments a block. Binary search is easy and very quick. Even if the problem is sporadic, you would do well to try to narrow it down - the search space is currently too large for people to be very helpful, I'm afraid. – Drew Mar 23 '15 at 13:37
  • @Drew can't be certain, I seen this last time issue occurred in __**Messages**__ buffer: `error in process sentinel: fringe-helper-modification-func: Invalid search bound (wrong side of point)` – Geradlus_RU Mar 24 '15 at 06:21
  • Well, looks like it's right catch, got same message second time. Can someone help now? (: – Geradlus_RU Mar 24 '15 at 06:26
  • finally, I've figured out that this is conflict between `git-gutter-fringe` and `flycheck`! Disabling `git-gutter-mode` gives back normal behaviour. It is safe to enable `git-gutter-mode` back, well, I suppose at least for a while. What should I do now: delete/close question or post an answer? – Geradlus_RU Mar 24 '15 at 13:07
  • If you found an answer you can post it (and you can accept it). You can edit your question to remove irrelevant stuff and make the problem (and solution) clearer, if you like. Whatever you think will be most helpful to others. – Drew Mar 24 '15 at 13:23
  • good, I'll recheck everything in a day, refine issue description and post answer. Thanks everybody! – Geradlus_RU Mar 24 '15 at 13:33

1 Answers1

0

In my particular case the issue was introduced by conflict between git-gutter-fringe plugin and FlyCheck when both of them try to place fringe marks at the same place. To spot this kind of conflicts it is helpful to open * Messages * buffer (by default C-h e), where all kind of errors logged

To resolve problem, I simply rejected git-gutter-fringe plugin.

Hope, this will be helpful for somebody.

Geradlus_RU
  • 1,466
  • 2
  • 20
  • 37