encoding used by those #some.txt# backup?

encoding used by those #some.txt# backup?

[Xah Lee]

it seems emacs's crash backup files saved my utf-8 files in a
different encoding.

for example, if i'm editing x.txt, i have a line:

∑ something

then emacs will create a file #x.txt# while x.txt buffer has not been
saved.
If i have a crash, then opening the #x.txt# file the content is:

\234\364\370\261 something

since my files have lots of unicode, this renders those auto-save not
usable. (diff becomes useless because about every line is shown as
different due to the encoding)

is this a bug or am i missing some file encoding for this type of
backups?
Those normal backup x.txt~ files are fine.

Thanks.

  Xah
∑ http://xahlee.org/

☄

Re: encoding used by those #some.txt# backup?

[Eli Zaretskii]

> From: Xah Lee <xahlee@gmail.com>
> Date: Mon, 19 Jan 2009 10:30:34 -0800 (PST)
> 
> it seems emacs's crash backup files saved my utf-8 files in a
> different encoding.

It's the internal representation of characters used by Emacs.  Since
auto-save files must be always encoded correctly, the only safe way of
doing that is to dump the internal representation.

> If i have a crash, then opening the #x.txt# file the content is:
> 
> \234\364\370\261 something
> 
> since my files have lots of unicode, this renders those auto-save not
> usable.

M-x recover-file RET should DTRT.

> is this a bug or am i missing some file encoding for this type of
> backups?

M-x recover-file RET should decode them correctly.

Re: encoding used by those #some.txt# backup?

[fnn]

You should set a apropriate Language Environment