Marpa::R2 to require HTML::Parser

12 views
Skip to first unread message

Jeffrey Kegler

unread,
Sep 30, 2012, 11:28:20 AM9/30/12
to Marpa Parser Mailing LIst
I've changed the latest developer's version of Marpa::R2 to require (and
not merely recommend) HTML::Parser and HTML::Entities. I did so because
my latest Marpa::R2 changes center on Marpa::R2::HTML, and apparently,
many cpantesters ignore recommended modules. Marpa::R2::HTML was
getting very little testing by cpantesters.

My current intention is to carry this requirement forward into the next
indexed release of Marpa::R2, but I would appreciate feedback.

Thanks, jeffrey

Ron Savage

unread,
Sep 30, 2012, 4:04:32 PM9/30/12
to marpa-...@googlegroups.com
Hi Jeffrey
Did you consider splitting off Marpa::R2::HTML into its own distro,
which would mandate HTML::*, but which would have the effect that
Marpa::R2 did /not/ need HTML::*?


--
Ron Savage
http://savage.net.au/
Ph: 0421 920 622

Jeffrey Kegler

unread,
Sep 30, 2012, 6:38:30 PM9/30/12
to marpa-...@googlegroups.com
Yes. Splitting out the distros was what I did with Marpa::{XS,PP} and
Marpa::HTML. I don't want to do that again because Marpa::R2::HTML
versions are tied tightly with Marpa::R2, and that kind of one-to-one
version tie-in does not play well with the way CPAN and cpantesters
works. CPAN assumes that you keep modules backward and forward
compatible, and implements other approaches as exceptions or not at all.
I learned this the hard way.

I don't mean to disparage CPAN's approach. Perhaps the single biggest
reason that Marpa was written in Perl was CPAN. CPAN's approach works
well in most cases.

-- jeffrey

Ron Savage

unread,
Sep 30, 2012, 7:52:57 PM9/30/12
to marpa-...@googlegroups.com
Hi Jeffrey

On 01/10/12 08:38, Jeffrey Kegler wrote:
> Yes. Splitting out the distros was what I did with Marpa::{XS,PP} and
> Marpa::HTML. I don't want to do that again because Marpa::R2::HTML
> versions are tied tightly with Marpa::R2, and that kind of one-to-one
> version tie-in does not play well with the way CPAN and cpantesters
> works. CPAN assumes that you keep modules backward and forward

True.

> compatible, and implements other approaches as exceptions or not at all.
> I learned this the hard way.

OK.

> I don't mean to disparage CPAN's approach. Perhaps the single biggest
> reason that Marpa was written in Perl was CPAN. CPAN's approach works
> well in most cases.
>
> -- jeffrey
>
> Ron Savage wrote:
>> Hi Jeffrey
>>
>> On 01/10/12 01:28, Jeffrey Kegler wrote:
>>> I've changed the latest developer's version of Marpa::R2 to require (and
>>> not merely recommend) HTML::Parser and HTML::Entities. I did so because
>>> my latest Marpa::R2 changes center on Marpa::R2::HTML, and apparently,
>>> many cpantesters ignore recommended modules. Marpa::R2::HTML was getting
>>> very little testing by cpantesters.
>>> My current intention is to carry this requirement forward into the next
>>> indexed release of Marpa::R2, but I would appreciate feedback.
>>
>> Did you consider splitting off Marpa::R2::HTML into its own distro,
>> which would mandate HTML::*, but which would have the effect that
>> Marpa::R2 did /not/ need HTML::*?
>>
>>
>
>
>


Reply all
Reply to author
Forward
0 new messages