http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49152
--- Comment #20 from Ralph Loader <suckfish at ihug dot co.nz> 2012-03-23 07:54:51 UTC --- Re comment 12 - as someone who regularly needs to understand gcc diagnostics, I disagree completely. Understanding a failure to look something up, the single most important thing to know is the key being looked up, and in this case the key is the function/operator name and the argument types. gcc used to output that information, and removing that was a regression, completely independently of whether the current diagnostics make any sense. Given that no significant progress has been made with fixing the current diagnostics in nearly a decade, it is puzzling to me why the obvious & simple fix of reverting to the previous, useful, diagnostics, has not been made.
