On Thu, 19 Jul 2018, Martin Liška wrote: > I must admit that was my intention :) In my eyes it makes it more consistent > and > it gives consumers feedback about usage of an option that does nothing. > For x86_64 there's list of options that are Ignore and don't produce a > warning:
The design is meant to be that if the option is purely non-semantic - if it's an option for tuning/enabling/disabling some aspect of some optimization, and the implementation has changed so that the option no longer makes sense with the current set of optimizations, for example - then silent ignoring is appropriate. I think that generally applies to most of the options in your list. -- Joseph S. Myers jos...@codesourcery.com
