Re: recent argp changes

[Ben Pfaff]

Paul Eggert <address@hidden> writes:

> Bruno Haible <address@hidden> writes:
>
>> you are now assuming an in-place sorting algorithm. If a libc were
>> to implement qsort by use of mergesort (and temporary storage), such
>> a sorting algorithm would pass you two pointers to hol_entrys in
>> different array.
>
> A minor quibble: the C standard (both C89 and C99) does not allow such
> an implementation for qsort; it requires that the arguments to the
> comparison function be bona fide pointers into the
> originally-specified array.

But even if this is true (it does seem to be implied by the
specification in the standard, now that I've re-read it), it
doesn't mean that comparison based on address will force a stable
sort.  I can, for example, imagine an implementation of qsort
that performs a random shuffle on the contents of the array
before doing a quicksort, to reduce the probability of
pathological behavior.
-- 
Ben Pfaff
address@hidden