Re: [Tzdist-bis] Last Call: <draft-murchison-tzdist-tzif-13.txt> (The Time Zone Information Format (TZif)) to Proposed Standard

[Paul Eggert <[e-mail]>]

tom petch wrote:
> I would expect the
> index to be
> [0, 'typecnt'-1)
> or
> [1, 'typecnt')
> but not
> [0, 'typecnt')
> as the I-D says, unless zero has a special meaning but then it is not
> really an index.

Alhough the [lo, hi) notation is well-established (as Tim Parenti mentioned), it 
wouldn't hurt to reword this to [0, 'typecnt'-1] to avoid the confusion you ran 
into.
> I infer, and would rather be told explicitly, that
> "The total number of octets is specified by the 'charcnt' field ..
> excludes the NUL that terminates the list.

charcnt includes the NUL; it does not exclude it. Since you inferred otherwise 
we should make this explicit. Also, come to think of it, we should say that 
charcnt MUST NOT be zero, as this constraint (which is a logical implication of 
the already-documented constraint that typecnt must not be zero) will help avoid 
further confusion in this area.
> In s.1,
> "the source the time zone data  metadata, identifiers, ..."
> would appear to be light of an 'of'.

Thanks, good catch.

Proposed patch attached. It attempts to address all the above issues.