[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: buffer-modified-tick and text properties
From: |
Eli Zaretskii |
Subject: |
Re: buffer-modified-tick and text properties |
Date: |
Sat, 20 Apr 2024 17:09:10 +0300 |
> From: JD Smith <jdtsmith@gmail.com>
> Date: Sat, 20 Apr 2024 08:47:42 -0400
>
> I had hoped to compare a saved and current buffer-modified-tick in a
> post-command-hook, to tell me whether
> the command that lead to it had modified the buffer text, but this is not a
> reliable way to do so when text
> properties are being changed as a result of non-edit commands (like motion).
If you want to know only about changes to characters, perhaps
buffer-chars-modified-tick is a better function to use?
Or how about if you explain why the fact that buffer-modified-tick
changes even under with-silent-modifications gets in your way, and
what would you like to do that need that tick to be sensitive to what
with-silent-modifications does?
> Plenty of modes compare the buffer tick against a saved value to infer
> whether the buffer text has been
> modified since the last update. That logic will be defeated if other modes
> make text property changes. Is
> there a way to inhibit changes in the buffer-tick when you are just changing
> simple text properties?
You cannot do that without breaking redisplay, because the display
engine uses the tick value to know when it might need to redisplay
portions of a window. But I'm still uncertain why this presents a
problem, given that we have buffer-chars-modified-tick. What am I
missing?