current s390x libstdc++ failures analysis

[Paolo Carlini]

Hi,
> 1) strict-aliasing rule breaking, warning
> I believe there is already a PR filed about this, remember reading some
> exchange between Paolo and Richard G trying to figure out what is going
> on. (I remember thinking the exchange was quite funny.) Paolo, can you
> help me remember what the number is? I cannot find it with searching.
> This should be marked as a 4.4 regression, if it is not already.
>   
I don't know, if you mean middle-end/38477 is already fixed.

Paolo.