bug#40407: [PATCH] slow ENCODE_FILE and DECODE_FILE

[Eli Zaretskii]

> From: Mattias Engdegård <mattiase@acm.org>
> Date: Sat, 4 Apr 2020 20:01:12 +0200
> Cc: 40407@debbugs.gnu.org
> 
> 4 apr. 2020 kl. 19.04 skrev Eli Zaretskii <eliz@gnu.org>:
> 
> > How so?
> 
> Because then NOCOPY suddenly matters for almost all coding systems, not just 
> nil. After all, an all-ASCII input string and ASCII-compatible coding is not 
> an unusual combination. This forces us to be careful when correcting the 
> NOCOPY sense, and may expose latent bugs.

That's true.  But as a matter of fact, I don't see any calls to
code_convert_string with NOCOPY non-zero, they all pass zero or false
to it.  So none of the existing direct calls from C wants or expects
to get the same string.

> But you are right: we should trust calls to {en,de}code-coding-system with 
> NOCOPY=t, and the rest can also remain as they are until someone cares enough 
> to change them.

Agreed.

Btw, this bug was introduced in commit 4031e2b, 18 years ago.