On Tue, Jun 9, 2015 at 6:58 PM, Greg Troxel <[email protected]> wrote > > Vincent JARDIN <[email protected]> writes: > > That seems fine. Although I would be tempted to add a #warning for the > case of inlined system .h, as it seems wrong. > > I agree it's wrong and I was hesitant to blindly remove the inline system .h files. Not with all the gyrations I've had to do lately in order to get the various systems up and running.
> > can be taken: > > http://patchwork.quagga.net/patch/1249/ > > It strikes me that in v4, 'sho ip route' is system wide, and then there > are per-protocol commands, so 'sho ip mroute' should in theory be about > all mcast routes that zebra has, separate from where they come from. So > this seems backwards. > > I thought it was backwards too, but this thread convinced me otherwise: https://lists.quagga.net/pipermail/quagga-dev/2015-May/012536.html Hence the state of the patch I put forward. > > > http://patchwork.quagga.net/patch/1240/ > > This looks like it is gratuitously dropping a lot of const, and I don't > see the point. Maybe that's because I didn't read the missing commit > message. > > I sure wish this one was easier. I spent a non-trivial amount of time attempting to keep the const's there and came to the conclusion that execv's arguments are not your friend when attempting to remove the compiler warnings( so we could turn on -Werror internally). All in all, I'm not really sure what a point of a const argument to an execv call is, as that execv replaces the current process with the new one. > > minor update to be sent, then could be merged: > > http://patchwork.quagga.net/patch/1248/ -> Please, Donald? > > looks ok to me. > I'll send the new patch in a bit here. donald
_______________________________________________ Quagga-dev mailing list [email protected] https://lists.quagga.net/mailman/listinfo/quagga-dev
