bug#8435: misuse of error ("...%d...", ...) on 64-bit hosts

[Paul Eggert <eggert@cs.ucla.edu>]

On 04/09/2011 11:32 AM, Eli Zaretskii wrote:
> I don't think line count is a compelling argument in favor of the
> change.

It's certainly good evidence that the change will simplify
Emacs maintainance overall.  It's not often that we can fix
multiple bugs like this by removing over 250 lines of source code.

> we now have to solve those problems in users of that function

Yes, that's a disadvantage, but it's a minor one; only a
17 lines of code are affected.  And there is a real advantage
to sticking with a stable, widely-used standard interface like
vsnprintf, as this makes it easier on code readers who are not
expert in Emacs internals.

As Emacs evolves, if we find that more lines of code are affected,
then the tradeoffs will change.  If that happens, it shouldn't
be hard to come up with a doprnt replacement that uses
vsnprintf internally and that also handles multibyte character
truncation and non-8-bit codepoints.  But given Emacs's current
use of doprnt, this would be overkill and would add complexity:
it would save 17 lines in doprnt's callers but require considerably
more than 17 lines to implement.