On Mon, Nov 18, 2013 at 12:51 PM, Saurabh Shah <ssaur...@vmware.com> wrote:
> > Ben wrote: > > Microsoft documentation is just hilarious: > vsnprintf is included for compliance to the ANSI standard; ... > But I guess this wrapper won't help much because the MS libc doesn't > support, for example, the 'z' modifier. What's your plan for dealing > with that? > > > On Mon, Nov 18, 2013 at 09:45:48AM -0800, Saurabh Shah wrote: > > I was hoping we can just print with a larger length specifier. '%z' > will there have to be disallowed and an appropriate signed/unsigned > specifier should be used instead. Do you think that would be OK? > > > It's going to cause trouble not just now but in the future. Everyone > knows and uses 'z' for size_t, and new uses of 'z' will therefore > continue to be introduced. We need to come up with some way to scan > source code for it to avoid introducing new uses, or to avoid the > problem entirely. > > > By cause trouble, do you mean it will introduce bugs? I agree it will be a > little annoying for people who know and use 'z'. And, I can definitely > document the reason for not using the modifier where I define > ovs_vsnprintf. > vsnprintf() is only a tiny part of the problem. Very little code directly calls vsnprintf(), although a significant amount of code calls it indirectly. A quick grep shows that here are at least 143 uses of the 'z' format modifier in OVS. Documentation can help in some cases, but I don't think it will help in this case. No one reads the documentation for printf-like functions when they use them, because everyone already knows how printf format specifiers work. What would be a better alternative in your opinion? > I can see two alternatives. One is to add a rule to the Makefiles that reports uses of the 'z' format specifier. Another is to add a correct implementation of *printf() and related functions to the tree, and set up the build so that this implementation gets used transparently (either all the time or just when the C library's implementation is buggy). > > If not, there are only a handful of places where we use > vsnprintf()'s, so I could #ifdefs the code. > > > Why is vsnprintf the only issue? *printf() has many variants, and all > of them are potential problems. > > Sure > -- "I don't normally do acked-by's. I think it's my way of avoiding getting blamed when it all blows up." Andrew Morton
_______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
