On Sat, 17 Dec 2016, Dimitry Andric wrote:
On 17 Dec 2016, at 12:46, David Chisnall <thera...@freebsd.org> wrote:
On 16 Dec 2016, at 19:31, Baptiste Daroussin <b...@freebsd.org> wrote:
Other than that, it makes more difficult to use vanilla gcc with out userland.
and it is adding more complexity to be able to build freebsd from a non freebsd
system which some people are working on.
Why? You???ll get some spurious warnings about printf, but that???s all.
Unfortunately, we compile large parts of the tree with -Werror. Thus,
"spurious warnings" will break the build, leaving the user two options:
disabling -Wformat warnings, or disabling -Werror altogether, neither of
which are very recommendable.
As far as I know, there is no -Wno-error-on-undefined-printf-specifiers.
It would also be hard to implement, since after any undefined specifiers
have been encountered, you cannot reason about the following ones
anymore either.
IMHO, if people want to use non-standard specifiers, let them define
their own almost_printf_but_not_quite() functions, and forgo any format
checking.
That would be worse than breaking format checking for the selected set
of printf()s. It gives even more unportability. %b is a BSDism that
would be detected at compile time on systems without support for %b in
printf(). almost_printf_but_not_quite() is a FreeBSDism that would
be detected at compile time on systems without the function. Using
it breaks portability even to other BSD systems including previous
versions of FreeBSD.
People who want to use non-standard specifiers added support to check
them the compiler.
Bruce
_______________________________________________
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
