Multiple values. Was: Re: 5.1

[Rici Lake <lua@...>]

On 18-Oct-05, at 1:24 AM, Glenn Maynard wrote:

> The problem is that design of the language, in this area, is 
> unintuitive
> and unnatural to me.  The multiple-results documentation in PiL reads 
> like
> a checklist of special cases to be memorized.  I'm not saying it's 
> something
> that should or could be changed; it just feels forced.

PiL says:

> 1) Lua always adjusts the number of results from a function to the 
> circumstances of the call.

> 2) When we call a function as a statement, Lua discards all of its 
> results.

> 3) When we use a call as an expression, Lua keeps only the first 
> result.

> 4) We get all results only when the call is the last (or the only) 
> expression in a list of expressions.

It seems pretty clear to me. The short summary (leaving out the obvious 
case of function call statements) is:

1) Lua expects an expression to yield exactly one value except when the 
expression is the last (or only) expression in an expression-list, in 
which case it yields all of its values.

2) Expression lists (as a whole) are adjusted to the context of an 
expression list, by either discarding extra results, or by appending a 
sufficient number of nils.

Rule 2 means that:

a, b = foo()

will adjust foo() to 2 return values, since clearly two values are 
expected (and since foo() is the only expression in an expression 
list.)

The one apparent exception to this is numeric for statements, because 
"a,b,c" in:

  for i = a, b, c

is not grammatically an expression list. (It is <expr> "," <expr> ["," 
<expr>] )

Normal value adjustment would result in the last value being nil rather 
than 1 if the rule were applied literally, but possibly it would be 
less surprising if

  for i = ...

were valid, with the third value being treated as 1 if it were nil (and 
possibly the second value being treated as the first value.)