[dpdk-dev] [dpdk-stable] [PATCH] examples/rxtx_callbacks: fix port ID format specifier

Tyler Retzlaff roretzla at linux.microsoft.com
Thu May 6 00:45:21 CEST 2021


On Wed, May 05, 2021 at 11:39:23PM +0200, Thomas Monjalon wrote:
> 05/05/2021 18:00, Tyler Retzlaff:
> > On Tue, May 04, 2021 at 09:48:22AM +0300, Dmitry Kozlyuk wrote:
> > > > 
> > > > what was the -Wformat clang on windows complaint?
> > > 
> > > PRIx16 would work, but I noticed that in other places where port ID is
> > > printed, the pattern above is used. IMO uniform approach is better.
> > 
> > ah, consistency. yes i'll have some of that. maybe one day in the future
> > we can change them all in one shot, but not today.
> > 
> > Acked-By: Tyler Retzlaff <roretzla at linux.microsoft.com>
> 
> I think PRIu16 is more appropriate.

works for me.

> Consistency doesn't matter if an approach is better.

well, there is some value in having consistency even if it is something
sub-optimal if we intend to do a mechanical change of all instances in
the future.

as a side question, what is the projects stance on getting more warnings
clean? there are a few not enabled that i'd really like to see e.g.
format, conversion, truncation etc..

i looked at lib/eal previously and there are... hundreds? of instances so
it's a non-trivial task. the problem i see is somehow getting to a
warnings clean state where we can enable -Werror in the CI pipeline but
at the same time figuring out how to prevent new instances from
appearing until we do.


More information about the dev mailing list