> -----Original Message----- > From: Alexander Akulich <[email protected]> > On Thu, Feb 6, 2020 at 4:52 PM Alex Blasche <[email protected]> wrote: > > > > Considering that our naming convention for error() signals is inconsistent > anyway, I favour an approach that highlights API changes early. > > The convention can not be inconsistent. It can either do not exist ("we have > no > convention, so we're agreed to do whatever we want and have inconsistency in > the API") or it can exist, so we want to follow it by definition.
Sorry, I should clarify. The current application of the convention is very inconsistent across the Qt modules already. I do not advocate to drop the convention especially when considering new API's. However sometimes there are extenuating circumstances which may provide reasons to not follow them. Under my proposal this would be such a case when weighing the two arguments. -- Alex _______________________________________________ Development mailing list [email protected] https://lists.qt-project.org/listinfo/development
