Re: Table length evaluations --> Lua 5.5 vs. Lua 5.4

The comment in ltable.c line 1291 above luaH_getn:

** Try to find a border in table 't'. (A 'border' is an integer index
** such that t[i] is present and t[i+1] is absent, or 0 if t[1] is absent,
** or 'maxinteger' if t[maxinteger] is present.)

The "or 0 if t[1] is absent" makes it sound like it took precedence over
the first definition.

OTOH the documentation is correct and even mentions if there's more
than one border (non-sequence table) then any can be returned.

On Sun, 15 Feb 2026 at 15:48, Luther Thompson <[email protected]> wrote:

> On 2/15/26 07:04, Błażej Roszkowski wrote:
> > It seems there are still bugs. Comment in C source says "or 0 if t[1]
> > is absent" but
> > for {nil, 1, 1, 1} the # yields 4 (due to *lenhint being 2 so 1 is
> > never checked).
>
> I don't know where that comment is, but if it's like the documentation
> at https://www.lua.org/manual/5.5/manual.html#3.4.7, it's merely part of
> the definition of a border. Since the # result can be any border, it
> won't always be 0 if there are other borders in the table.
>
> Luther
>
> --
> You received this message because you are subscribed to the Google Groups
> "lua-l" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To view this discussion visit
> https://groups.google.com/d/msgid/lua-l/2b9ef29a-a816-40dd-b758-afe01c07ca07%40gmail.com
> .
>

-- 
You received this message because you are subscribed to the Google Groups "lua-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
To view this discussion visit https://groups.google.com/d/msgid/lua-l/CAOc0etHokkAP6W7xkrqevtRX%2BVn2eVGMFqUCbj4ohiPd5DY_VA%40mail.gmail.com.