Re: Speed up array allocation

[Tom Tromey <tromey at redhat dot com>]

Tom> I didn't want to do that since it wastes space.

Jeff> Really?  If you require all arrays to have double alignment, I
Jeff> figure you waste at most 4 bytes per array (on x86).  Is that
Jeff> really significant to some apps?

I doubt that it is.

It also makes the code in gcj/array.h conceptually cleaner, but I
guess that isn't all that big of a constraint.  We can always add
voluminous explanatory comments.

Bryce> I'd agree that for speed and simplicity it would be desirable
Bryce> to do this, if nobody comes up with a better
Bryce> solution. Presumably it would require changes to the compiler
Bryce> though?

Changes to the compiler and to gcj/array.h.  Right now the differing
alignment requirement comes from how arrays are reflected into C++.

Tom