On Sun, 31 May 2020 at 20:03, NightStrike via Gcc-help
<gcc-help@xxxxxxxxxxx> wrote:
>
> I'm curious if this is a gcc bug or not. The warning I get is trying to
> highlight a real problem, but it's referring to a string literal as a
> directive, which I thought was just for the %XX printf commands. Given the
> following:
>
> #include <stdio.h>
>
> void f() {
> char x[4];
> char y[5];
> sprintf(x, "%s_%s", y, y);
> }
>
> $ gcc a.c -c -Wall
> a.c: In function 'f':
> a.c:6:16: warning: '_' directive writing 1 byte into a region of size
> between 0 and 4 [-Wformat-overflow=]
> 6 | sprintf(x, "%s_%s", y, y);
> | ^
> a.c:6:2: note: 'sprintf' output between 2 and 10 bytes into a destination
> of size 4
> 6 | sprintf(x, "%s_%s", y, y);
> | ^~~~~~~~~~~~~~~~~~~~~~~~~
>
>
> I hope the fixed width display shows this correctly. The point is that the
> ^ points to the underscore, which is right, but the message calls the
> underscore a printf directive.
>
> Maybe I'm splitting hairs, but this confused me for a good half hour before
> I realized what was wrong.
That is the correct term. The %s parts are conversion specifications,
the "_" between them is a directive.
See https://pubs.opengroup.org/onlinepubs/9699919799/functions/printf.html
"The format string is composed of zero or more directives: ordinary
characters (not %), which are copied unchanged to the output stream;
and conversion specifications, each of which results in fetching zero
or more subsequent arguments."
