On Sat, 16 Aug 2014, Jan-Benedict Glaw wrote:
On Sat, 2014-08-16 17:10:12 +0200, Manuel López-Ibáñez <lopeziba...@gmail.com>
wrote:
On 16 August 2014 16:56, Jan-Benedict Glaw <jbg...@lug-owl.de> wrote:
There's --enable-werror-always, though that didn't trigger before (et
least not for the printf'ish checks.) So what's specifically different
with this new function?
Sorry, I don't find that option here: https://gcc.gnu.org/install/configure.html
so I'm not sure what it does.
The .ac fragment is:
,---------------------------------------------
| # Enable -Werror in bootstrap stage2 and later.
| AC_ARG_ENABLE(werror,
| [AS_HELP_STRING([--enable-werror],
| [enable -Werror in bootstrap stage2 and later])], [],
| [if test -d ${srcdir}/gcc && test x"`cat $srcdir/gcc/DEV-PHASE`" =
xexperimental; then
| enable_werror=yes
| else
| enable_werror=no
| fi])
| case ${enable_werror} in
| yes) stage2_werror_flag="--enable-werror-always" ;;
| *) stage2_werror_flag="" ;;
| esac
| AC_SUBST(stage2_werror_flag)
`---------------------------------------------
This was added on Wed Mar 24 22:08:39 2004 +0000. The `config-list.mk'
file uses it since it was added to GCC on Thu May 5 01:47:47 2011 +0000.
But if the compiler using -Werror is an earlier revision as the one
that modified c-format.c, it should not use -Werror because it will
produce Wformat errors. Is this the case here?
Not sure:
gcc20: g++ (Debian 4.4.5-8) 4.4.5
gcc76: g++ (Debian 4.4.5-8) 4.4.5
pluto: g++ (Debian 4.9.1-1) 4.9.1
However, the "pluto" box also fails, and it's GCC is about up-to-date.
For your tests, I thought the sequence was:
checkout some revision of gcc
build a native gcc at this revision
use it to build a bunch of cross-compilers at the same revision
Otherwise, I don't see how -Werror can work.
--
Marc Glisse
