bug#28780: 25.3; Bad signature from GNU ELPA whe coding-system-for-read is utf-8

[Iakov Davydov <dav02.davs@myths.ru>]

[-- Attachment #1: Type: text/plain, Size: 1372 bytes --]

Hi Eli,

On Tue, Oct 10, 2017 at 6:47 PM Eli Zaretskii <eliz@gnu.org> wrote:

> > 2. Execute `(setq-default coding-system-for-read 'utf-8)` (e.g. go to
> > scratch, type this command, type C-x C-e).
> Why do you need to do step 2?  You shouldn't do that, especially when
> using commands that read binary files.  coding-system-for-read is for
> temporarily forcing specific decoding of specific text, and should
> otherwise be left alone.  And you definitely shouldn't change its
> default value.
>
> IOW, this sounds like a clear cockpit error.  Am I missing something?
>
After carefully reading the documentation for coding-system-for-read, I
think you are right. It shouldn't be used like that.

However the snipped including this command to "unicodify" emacs is
mentioned in many places including some answers on stackoverflow.

I think you the bug can be closed, thank you.

> 2. https://emacs.stackexchange.com/questions/10046
> > 7. https://www.linux.org.ru/forum/security/13718006 (In Russian)
>
> I see no clear evidence that these reports are due to setting
> coding-system-for-read.
>

At least cases #2 and #7 have a very the similar error message (and the
same package in one case).

Cheers,
Iakov
-- 
Iakov Davydov
Postdoc, Department of Computational Biology and Department of Ecology and
Evolution, University of Lausanne
Swiss Institute of Bioinformatics

[-- Attachment #2: Type: text/html, Size: 2293 bytes --]