On Tue, 19 Nov 2019, Richard Biener wrote:
> > +/* For nested functions disqualify ones taking VLAs by value
> > + from inlining since the middle-end cannot deal with this.
> > + ??? We should arrange for those to be passed by reference
> > + with emitting the copy on the ca
On Fri, 8 Nov 2019, Richard Biener wrote:
>
> The following works around a middle-end limitation not being able
> to deal with by value-passing of VLAs transparently during inlining
> (but only DECL_BY_REFERENCE is handled) in the C frontend by marking
> said functions as not inlinable. This avo
The following works around a middle-end limitation not being able
to deal with by value-passing of VLAs transparently during inlining
(but only DECL_BY_REFERENCE is handled) in the C frontend by marking
said functions as not inlinable. This avoids ICEs later.
Bootstrapped and tested on x86_64-u
