On Feb 25, 2010, at 10:01 AM, Tim Bunce wrote:
>> That's two unacceptable alternatives, you need to find a third one.
>> I think most people will have no trouble settling on "do not update
>> to Safe 2.2x" if you don't offer a better solution than these.
>=20
> I believe the next version of Safe will revert to Safe 1.19 behaviour
> because the side effects of the change in 2.20 are too severe for it to
> be left enabled by default.
Which means losing sort $a <=3D> $b again, alas. Such was always the case i=
n the past, so that might be an okay tradeoff to get recursive calls workin=
g again, but I certainly hope that Safe can be updated in the near future t=
o give us both.
There seem to be no good answers here.
David