From mboxrd@z Thu Jan 1 00:00:00 1970 From: Olivier MATZ Subject: Re: [PATCH 1/3] stringfns: remove rte_snprintf Date: Wed, 25 Jun 2014 09:55:47 +0200 Message-ID: <53AA8083.8000900@6wind.com> References: <20140624090253.140206a7@nehalam.linuxnetplumber.net> <59AF69C657FD0841A61C55336867B5B02CEE3359@IRSMSX103.ger.corp.intel.com> <20140624105427.166c21bf@nehalam.linuxnetplumber.net> <59AF69C657FD0841A61C55336867B5B02CEE345B@IRSMSX103.ger.corp.intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: "dev-VfR2kkLFssw@public.gmane.org" To: "Richardson, Bruce" , Stephen Hemminger Return-path: In-Reply-To: <59AF69C657FD0841A61C55336867B5B02CEE345B-kPTMFJFq+rELt2AQoY/u9bfspsVTdybXVpNB7YpNyf8@public.gmane.org> List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces-VfR2kkLFssw@public.gmane.org Sender: "dev" Hi Bruce, On 06/24/2014 08:00 PM, Richardson, Bruce wrote: >> I want to get it out now rather than some 2 year life cycle. >> The issue was discussed and marking it as deprecated breaks the build. >> Alternate is removing all instances and adding: >> >> #define rte_snprintf snprintf >> >> in header file for user compatiablity. > > We can remove it from all our apps and then mark as deprecated and all our code would still build. > An interesting point I'd never thought of is, is it right for us to force user apps to treat all warnings as errors? Perhaps we should also consider removing -Werror from the CFLAGS when using rte.extapp.mk. I don't understand why removing a function like rte_snprintf() in a new version of DPDK would be a problem. Yes, it would break the build of external applications, but that's the best way to fix the problem in the external code. Adding a compatibility layer just delays the issue. Maybe having a line about this in a "release note" or a "porting guide" would be enough? It could even references the commit id, showing how to solve the issue. About the -Werror flag, I would say that removing it is not a good idea. From my experience, many issues are pointed out by warnings, and forcing the compiler to stop on warning helps to have better code quality. Regards, Olivier