bug#64253: 26.3; Bug in revert-buffer-with-coding-system

bug#64253: 26.3; Bug in revert-buffer-with-coding-system

[meta1729]

When a buffer is terminated with an un-encodable character,
revert-buffer-with-coding-system modifies the buffer in unexpected
ways when coding system of the buffer is changed from default
iso-latin-1-unix to utf-8-unix, when the visited file contains
characters that cannot be encoded in either iso-latin-1-unix or
utf-8-unix.

Procedure to reproduce the bug.

[Create a file which has un-encodable characters and some text between them.]
$ echo 3b3b20410a3b3b20420aa80a3b3b20430a3b3b20440aa9 | xxd -r -p > sample

Open in Emacs.
$ emacs -q -nw --no-site-file -nsl sample

Enter: C-h C [describe-coding-system]
Output:
Coding system for saving this buffer:
  1 -- iso-latin-1-unix (alias: iso-8859-1-unix latin-1-unix)

Enter: M-x revert-buffer-with-coding-system RET utf-8 RET yes

[The buffer has been incorrectly modified.  You can see escaped
character code at the beginning of buffer, which previously existed at
the end of the buffer.]

Enter : C-h C [describe-coding-system]
Output:
Coding system for saving this buffer:
  U -- utf-8-unix (alias: mule-utf-8-unix cp65001-unix)

[Save the buffer to demonstrate that the file has really been
incorrectly modified.]

Enter: C-x C-w [write-file]
Filename: sample2
Select coding system: default raw-text 

Exit Emacs. C-x C-c

$ xxd -p sample2
a93b3b20410a3b3b20420aa80a3b3b20430a3b3b20440a

[Compare sample and sample2.
Last character (hex code a9) has been moved to beginning of file.]

Retry above procedure with different choice for
revert-buffer-with-coding-system.
Does occur when utf-8 or utf-16 or utf-7 is chosen.
Does not occur when iso-2022-7bit is chosen.

In GNU Emacs 26.3 (build 2, x86_64-pc-linux-gnu, GTK+ Version 3.24.14)
 of 2020-03-26, modified by Debian built on lcy01-amd64-020
System Description:	Ubuntu 20.04.6 LTS

bug#64253: 26.3; Bug in revert-buffer-with-coding-system

[Eli Zaretskii]

> Cc: meta1729 <meta1729@gmail.com>
> Date: Fri, 23 Jun 2023 12:43:19 +0530 (IST)
> From: meta1729 <meta1729@gmail.com>
> 
> 
> When a buffer is terminated with an un-encodable character,
> revert-buffer-with-coding-system modifies the buffer in unexpected
> ways when coding system of the buffer is changed from default
> iso-latin-1-unix to utf-8-unix, when the visited file contains
> characters that cannot be encoded in either iso-latin-1-unix or
> utf-8-unix.
> 
> Procedure to reproduce the bug.
> 
> [Create a file which has un-encodable characters and some text between them.]
> $ echo 3b3b20410a3b3b20420aa80a3b3b20430a3b3b20440aa9 | xxd -r -p > sample
> 
> Open in Emacs.
> $ emacs -q -nw --no-site-file -nsl sample
> 
> Enter: C-h C [describe-coding-system]
> Output:
> Coding system for saving this buffer:
>   1 -- iso-latin-1-unix (alias: iso-8859-1-unix latin-1-unix)
> 
> Enter: M-x revert-buffer-with-coding-system RET utf-8 RET yes
> 
> [The buffer has been incorrectly modified.  You can see escaped
> character code at the beginning of buffer, which previously existed at
> the end of the buffer.]

Thanks.  This 21-year old bug (a simple omission of a single code
line) should now be fixed on the emacs-29 branch.

bug#64253: 26.3; Bug in revert-buffer-with-coding-system

[Eli Zaretskii]

> Cc: 64253@debbugs.gnu.org
> Date: Sat, 24 Jun 2023 11:51:39 +0300
> From: Eli Zaretskii <eliz@gnu.org>
> 
> > $ emacs -q -nw --no-site-file -nsl sample
> > 
> > Enter: C-h C [describe-coding-system]
> > Output:
> > Coding system for saving this buffer:
> >   1 -- iso-latin-1-unix (alias: iso-8859-1-unix latin-1-unix)
> > 
> > Enter: M-x revert-buffer-with-coding-system RET utf-8 RET yes
> > 
> > [The buffer has been incorrectly modified.  You can see escaped
> > character code at the beginning of buffer, which previously existed at
> > the end of the buffer.]
> 
> Thanks.  This 21-year old bug (a simple omission of a single code
> line) should now be fixed on the emacs-29 branch.

No further comments, so I assume the bug is indeed fixed, and I'm
closing it.