[dpdk-dev] [PATCH 1/2] scripts: add checkpatch wrapper

Bruce Richardson bruce.richardson at intel.com
Fri Oct 30 17:16:40 CET 2015


On Thu, Oct 29, 2015 at 02:57:42PM +0100, Thomas Monjalon wrote:
> 2015-10-29 14:54, David Marchand:
> > On Thu, Oct 29, 2015 at 2:48 PM, Bruce Richardson wrote:
> > > On Thu, Oct 29, 2015 at 02:34:32PM +0100, Thomas Monjalon wrote:
> > > > I agree with both of you.
> > > > I could suggest something but I'm afraid it will be difficult to have a
> > > > consensus between a "quiet tool" and a "double check verbose tool".
> > > > As it is a really critical piece of code, I think we should have a
> > > meeting
> > > > with a technical steering comittee ;)
> > > > ... or we can add an option: -q or -v ? Debate is open :D
> > > 
> > > Yes, the whole future of the project could hinge on this decision :-)
> > 
> > Eheh :-)
> > 
> > > Ok, my suggestion is both!
> > > 1) Have the default (in case of no errors), be a single line print out at
> > > the end
> > > stating number of files scanned
> > > 2) If "-q" flag specified, skip this
> > > 3) If "-v" flag specified, do current behaviour with a line per file.
> > 
> > Ok for me.
> 
> I'm really happy we can have a sane consensus to this difficult question,
> with just few emails!
> Thanks guys :)
> 
> PS: I will send a v2 when the easy task of RC1 integration will be done ;)

Another request, can you perhaps also fix the script for situations where 
checkpatch.pl is not in the kernel tree. I've used this script now to check a
couple of patchsets, which came back clean, but it turns out that because I was
using checkpatch.pl outside the kernel directory, it is passing things it
shouldn't. [Thanks to Sergio for pointing this out]. 

Testing with a known-broken patch, this script indicates all ok, and only
reports an error with the --no-tree added to the options inside the script. :-(

/Bruce


More information about the dev mailing list