Re: Does PIL (3rd edition) repeatedly misuse the length operator on tables and invoke undefined behavior?

[Steven Degutis <sbdegutis@...>]

Quoth Tim Hill:

> No, I think the OP is correct insofar as what the reference manual says. My
> reading of the strict definition of sequence is that an empty table is NOT a
> sequence, and therefore the result of # is indeed undefined, regardless of
> what it actually returns in practice.

Given the manual wording in question:

"Unless a __len metamethod is given, the length of a table t is only
defined if the table is a sequence, that is, the set of its positive
numeric keys is equal to {1..n} for some integer n. In that case, n is
its length."

It says n may be any integer, and 0 is a valid integer.

The only two arguments against this are:

1. n must be >= 1 due to the use of {1..n} in the definition, and

2. that n may be a negative number (e.g. -14), thus making length = -14

My arguments:

1. The definition does not say that n must be >= 1, only that it must
be an integer; so we rely on whether 0 is a valid length, which we
hold as self-evident.

2. The concept of length precludes the possibility of a negative
length, excluding all negative values of n.

Thus no clarification in the manual is required, so long as these
principles are considered. Therefore I believe the current wording of
this manual quotation is satisfactory.

-Steven