On Thu, Apr 02, 2015 at 12:41:53AM +0200, Thomas Graf wrote: > On 04/01/15 at 08:06am, Ben Pfaff wrote: > > On Wed, Apr 01, 2015 at 11:34:39AM +0200, Thomas Graf wrote: > > > On 04/01/15 at 07:45am, Kavanagh, Mark B wrote: > > > > Agreed; given the precedent set by OVS_ENABLE_WERROR, I figured that > > > > this was the convention. > > > > > > The only reason this still exists is for backwards compatibility > > > to not break existing build scripts. Specifying CFLAGS=-Werror has > > > the same effect and f.e. travis builds are doing exactly this. > > > > On the contrary, specifying CFLAGS=-Werror on the "configure" command > > line can break configuration because it makes any warnings reported by > > the compiler during feature tests look like test failures. > > Good point. Autoconf has been very good in fixing any tests > which fail to work with CFLAGS=-Werror though because a lot > of folks want to pass in -Werror like that but it's definitely > a risk for faulty tests.
I'm probably extra-sensitive to this because I maintain the Autoconf packaging for Debian (since sometime in the 1990s) and remember seeing several "bug reports" that were really users enabling -Werror in situations where it broke things. > Let me know if you think we should fix the travis scripts to > pass in -Werror differently. As long as it works, it works, and there's no need to change anything ;-) _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
