> I don't see the logic here. It is certainly not that hard to fix the > code to make this assertion trivial, simple loop through all insns in > sched_init() will do it.
Yes, but apparently nobody is willing to do that at the moment so we'll need to kludge until then and add calls to recog_memoized when we stumble on insns that have not been recognized. But our laziness shouldn't impact the user if we are confident that this cannot result in wrong code. > There is a bug and it should get fixed rather than suppressed. AFAICS nobody has submitted a real fix for this bug. I'm ready to approve Adam's patch but only if the assertion is downgraded to ENABLE_CHECKING. -- Eric Botcazou
