Re: Next word of alias to alias that end with <blank> is not checked for alias

[Daniel Douglas]

> Of course, all mention of aliases should really be removed from
> POSIX - aliases are even worse than "set -e".

It should be removed because the spec is full of inaccuracies and isn't
useful. People that know enough to define the standard properly have
other priorities clearly. I think many of us are not inclined to follow
the standard anyway because details of alias expansion easily force
unacceptable constraints on the implementation that are more important
than emulating historical shells.

This is one among many examples where it makes little sense to point
out one small difference without considering the overall design of the
parser. There are more where this came from. I could write about alias
bugs all day long if finding small inconsistencies is the only goal.