On 2008/10/07 17:30, Girish Venkatachalam wrote: > On 13:15:53 Oct 07, irix wrote: > > Hello Ports, > > > > When you can answer about this port ? > > > > Wait. Don't rush. It will take a while but it will go in for sure. > > Without the broadcast patch. ;) > > -Girish >
On 2008/10/07 16:47, irix wrote: > Hello Ports, > > Why without broadcast patch ? Using FLAVOR for something that should be an option in the code doesn't make much sense. There must be some reason the upstream author doesn't want to do this by default. If you make it a proper option I'll consider it, maybe the upstream author will too. To do this you need simple changes to help() and usage(), and add a global variable in main() and ipguard.h (see how the other options work, e.g. "nofirst"). Add an unused character to the getopt() call and handle it - again, see how another option works. Then you can wrap the packet.c changes in some if/else testing for this variable, and update the docs. It also could use testing on more arch - a strict alignment arch (e.g. sparc64, arm), a big-endian arch (e.g. sparc64, powerpc), and an arch where char defaults to being unsigned (arm, powerpc). And this isn't the simplest thing to test since it involves changing network config on machines other than the one you're testing, and there's the chance of it knocking things out on the network you're testing on, so these machines may need direct console access rather than just SSH. This can be done, sometime, but it is much more work than the changes you need to make to your patch to make it optional...
