Ticket #308 (closed patch: fixed)
[PATCH] [imcc] Allow for 128-bit long doubles
|Reported by:||doughera||Owned by:||rurban|
This patch allows imcc to handle FLOATVALS with size == 16. (It already handled sizes of 8 and 12.)
I've also changed the fall-through case from a misspelled useless run-time warning to an error. If I understood why any of this was needed in the first place, I might have an informed opinion about whether or not any error at all should be issued. Meanwhile, this is better than the current state of affairs.
With this patch, Configure.pl --floatval='long double' should now build on 64-bit systems that have 128-bit long doubles. There are still some test failures with this change, but they look to me like errors in the test suite. In any case, there are fewer errors than without this change.