On Tue, 25 Sep 2001 19:23:36 +0200, in perl.perl5.porters you wrote:
> [Disclaimer: I haven't read perlpodspec yet, so some of my questions may
> be answered there. Feel free to point me there.]
OK, here are my comments after having read perlpodspec.
> the concept of "paragraph" doesn't appear to get defined explicitly.
Ah, these appear to be better defined in perlpodspec.
> And you don't seem to say explicitly whether the format name on the =end
> has to match the format name used on the =begin
Perlpodspec mentions that this closes "the end of the region opened by
the matching "=begin formatname" region.".
> (it could conceivably just be documentation, so you could use =begin html
> / =end stuff, or perhaps even =begin html / =end (without format name on
> the =end, sort of like #ifdef foo matching #endif).
So I suppose the format name has to be specified on =end, and it has to
match a =begin command.
> Also, whether =begin/=end blocks may be nested, e.g. for specialisations
> of format, and if so, what constraints are placed be on the =end
> commands.
But perlpodspec does not appear to say anything about nested formats. To
contrive an example,
=begin html
<p>This is (generic) HTML</p>
=begin html4
<p>This is HTML that needs a parser for HTML 4.0 or higher</p>
=end html4
<p>And more generic HTML</p>
=end html
Or, more strangely,
=begin html
<p>This is (generic) HTML</p>
=begin html4
<p>This is HTML that needs a parser for HTML 4.0 or higher</p>
=end html
<p>This is, er, what?</p>
=end html4
(overlapping regions)
Hm, near the end there's an example using nested regions. But it doesn't
explicitly prohibit overlapping regions.
> Also, what's the format of an I<formatname>? /\S+/? /\w+/? /.*/?
> /[:\w]+/?
I see that perlpodspec has some advice (viz., m/^:?[-a-zA-Z0-9_]+$/s
[sic]).
> Any valid HTML entity name? Or only those that map to Latin-1? What
> about ™, for example? Or €?
Ah, perlpodspec implies "all of them", since it explicitly mentions that
some may only support those mapping to Latin-1.
> As I understand it, the intent is that exactly one whitespace character
> is deleted while rendering such multiple-bracket delimiters
perlpodspec vaguely implies that all whitespace following the <{2,}
delimiter and preceding the >{2,} delimiter is snipped by specifying the
format as "two or more "<"'s, one or more whitespace characters, any
number of characters, one or more whitespace characters, and ending with
the first matching sequence of two or more ">"'s". But it doesn't say
so.
Cheers,
Philip
