On 27 March 2014 23:37, Daniel Gutson <daniel.gut...@tallertechnologies.com> wrote: >> >> There are some known bugs in -Wconversion and nobody working on them, >> so if you are still interested in helping, I can give you some hints >> on where your help will be very welcome. > > Could you please go ahead and show me the failing cases? > I only worked in the backend, but I'd like to take a look to the front-end.
Two "easy" ones: http://gcc.gnu.org/PR55096 about converting 'false' to bool *. http://gcc.gnu.org/PR60591 about unsafe enum conversions There are more, but they are a bit more complicated. If you feel brave enough, let me know. > There's also all the other commenters of my post regarding the =- and > =+ looking-like operators, which should take ws into account. > I think it would be a good idea to have a warning for =- and =+ saying something like: test.c:4:8: warning: suggested space between '=' and '-' to avoid confusion with operator '-=' x =- 4; ^ Clang already warns about this: test.c:4:7: warning: use of unary operator that may be intended as compound assignment (+=) x =+ 4; ^~ I think this should be easy to implement during parsing for both C and C++, so you may want to start with this one. Cheers, Manuel.
