http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54588
--- Comment #5 from Manuel López-Ibáñez <manu at gcc dot gnu.org> --- (In reply to Marc Glisse from comment #4) > > or maybe it does. I'd be very careful about removing information from the > diagnostics. I find cases where we are missing information (PR53822 for I didn't say it was easy ;-) but clang shows that it is possible in some cases to elide things that don't matter and point out exactly the differences that matter. And there are even more clear-cut cases like PR43113 > instance) much worse than those where we need to ignore a number of lines to > find the right message. I can read a few more lines, I cannot invent what > isn't there. So if there is any chance it might be useful, please at least > keep it under some -fdetailed-diagnostic flag. Clang has some specific flags controlling this, I guess for the same reasons that you give. But in this specific case, how can the type of VeryComplicatedType change anything about the convertibility of A<...> to 'int'? Sorry if I lack imagination...
