r226178 - PR 20146

Nathan Sidwell nathan at acm.org
Wed Jan 21 10:04:12 PST 2015

On 01/21/15 12:35, Joerg Sonnenberger wrote:
> On Thu, Jan 15, 2015 at 04:45:53PM -0000, Nathan Sidwell wrote:

> This breaks code using "volatile void" to mark functions as no-return,
> an oldish GCC extension. This is still somewhat popular in old code. Do
> we really want to / have to create a hard error for this?

It did seem a very picky, but unrejected, bug report  And I'd forgotten about 
the gcc extension you mention.  Perhaps it should be a pedantic error and 
warning otherwise?


More information about the cfe-commits mailing list