Bug#853421: glib2.0: ftbfs with GCC-7

[Emilio Pozuelo Monfort]

On 31/01/17 11:31, Simon McVittie wrote:
> On Tue, 31 Jan 2017 at 09:31:40 +0000, Matthias Klose wrote:
>> The full build log can be found at:
>> http://people.debian.org/~doko/logs/gcc7-20170126/glib2.0_2.50.2-2_unstable_gcc7.log
>> The last lines of the build log are at the end of this report.
> 
> ... and are completely useless, because this is a heavily parallel build.
> 
> Here is the actually useful part:
> 
> /<<PKGBUILDDIR>>/./glib/gscanner.c: In function 'g_scanner_unexp_token':
> /<<PKGBUILDDIR>>/./glib/gscanner.c:1469:61: error: '%.3f' directive output may be truncated writing between 5 and 314 bytes into a region of size 48 [-Werror=format-truncation=]
>        _g_snprintf (token_string, token_string_len, "number '%.3f'", scanner->value.v_float);
>                                                              ^~~~
> In file included from /usr/include/stdio.h:938:0,
>                  from /<<PKGBUILDDIR>>/./glib/gscanner.c:38:
> /usr/include/x86_64-linux-gnu/bits/stdio2.h:64:10: note: format output between 15 and 324 bytes into a destination of size 56
>    return __builtin___snprintf_chk (__s, __n, __USE_FORTIFY_LEVEL - 1,
>           ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>         __bos (__s), __fmt, __va_arg_pack ());
>         ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
> I don't know why "%.3f" % some_float can be 324 bytes long.

%.3f limits the fractional part to 3 digits, but not the integer part. Thus a
large integer will be truncated.

Cheers,
Emilio