> -----Original Message----- > From: Legacy, Allain [mailto:allain.leg...@windriver.com] > Sent: Friday, March 3, 2017 12:14 PM > To: Dumitrescu, Cristian <cristian.dumitre...@intel.com>; Yuanhan Liu > <yuanhan....@linux.intel.com>; Richardson, Bruce > <bruce.richard...@intel.com> > Cc: dev@dpdk.org; Jolliffe, Ian (Wind River) <ian.jolli...@windriver.com> > Subject: RE: [dpdk-dev] [PATCH 1/5] cfgfile: configurable comment character > > > > -----Original Message----- > > From: Dumitrescu, Cristian [mailto:cristian.dumitre...@intel.com] > > Both approaches can support this. Therefore, IMO the separator char is > not > > enough to justify approach 1. I would only go for approach 1 if there are > > some other parameters that we could consider adding to the load function > > now or later. Do you see any? > > No, I don't have any future parameters in mind but that doesn't mean that > none will arise eventually. IMO, the comment character should be specified > as an actual "char" in the rte_cfgfile_params. Specifying it as a flag is a > bit > kludgy - I don't like overloading a flag/enum to specify something that > already has a type that can be used (char). Also, I don't think we need to > control which comment characters are valid. If the app wants to use a 'X' > then that's their choice. > >
I disagree here. I think we must control the set of allowed separators to avoid confusion.
