On Tue, Jan 23, 2018 at 08:00:08AM +0100, Florian Weimer wrote: > On 01/22/2018 10:15 PM, Adam Jackson wrote: > > On Mon, 2018-01-22 at 19:19 +0100, Florian Weimer wrote: > > > On 01/22/2018 06:26 PM, Adam Jackson wrote: > > > > > > > > I'm trying to prepare xserver for this change, and it seems to provoke > > > > an awkward warning when building on F27: > > > > > > > > In file included from ../os/rpcauth.c:47:0: > > > > /usr/include/tirpc/rpc/rpc.h:83:12: warning: redundant redeclaration of ‘bindresvport’ [-Wredundant-decls] > > > > extern int bindresvport(int, struct sockaddr_in *); > > > > ^~~~~~~~~~~~ > > > > In file included from /usr/include/tirpc/rpc/rpc.h:40:0, > > > > from ../os/rpcauth.c:47: > > > > /usr/include/netinet/in.h:502:12: note: previous declaration of ‘bindresvport’ was here > > > > extern int bindresvport (int __sockfd, struct sockaddr_in *__sock_in) __THROW; > > > > ^~~~~~~~~~~~ > > > > > > > > Is there any reasonable fix for this? > > > > > > Redeclarations in system headers are expected. Do you compile with > > > -Wsystem-headers? Or do you something else which is unusual, such as > > > running the preprocessor separately? > > > > Not that I'm aware of. The generated cc line is: > > > > ninja: Entering directory `build' > > [1/17] ccache cc -Ios/libxserver_os@sta -Ios -I../os -Ixfixes -I../xfixes > > ccache runs the preprocessor separately. 8-) > > > Is -Wredundant-decls just not a thing one should try to use? > > Certainly with ccache, yes. The command uses -I /usr/include/tirpc/ rather than -isystem /usr/include/tirpc, so while /usr/include/netinet/in.h is treated like a system header, /usr/include/tirpc/rpc/rpc.h is not. Jakub _______________________________________________ devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx
