Re: Mitigating effects of Bug #750946

[Jonas Smedegaard <jonas@jones.dk>]

Quoting Kjetil Kjernsmo (2017-08-31 23:05:17)
> On Thursday, 31 August 2017 18:59:41 CEST gregor herrmann wrote:
> > AFAICS, in each of these cases HTML::HTML5::Parser is brought in as 
> > a transitive dependency via some other package.
> 
> Indeed, but it was the immediate dependency point that was difficult 
> to me.
>  
> > 
> > % reverse-depends libhtml-html5-parser-perl
> > Reverse-Recommends
> > ==================
> > * libhtml-html5-sanity-perl
> > * libhtml-html5-writer-perl
> 
> Aha, this was it that didn't make sense to me. Right, so even though 
> HTML::HTML5::Writer recommends HTML::HTML5::Parser in its metadata, it 
> is not used in code at all. A recommends is therefore too strong. If 
> there should be a relationship, suggests is more appropriate, i.e. 
> libhtml-html5-writer-perl suggests libhtml-html5-parser-perl Actually, 
> I think I would just remove the relationship alltogether, but that's 
> your call. :-)
> 
> The relationship between libhtml-html5-sanity-perl and libhtml-html5- 
> parser-perl is different. However, HTML::HTML5::Sanity clearly 
> enhances HTML::HTML5::Parser, but a recommends is too strong (I have 
> never used H:H:Sanity) and doesn't really reflect what it does:
> 
> My suggestion is therefore
> libhtml-html5-sanity-perl Enhances libhtml-html5-parser-perl 
> libhtml-html5-parser-perl Suggests libhtml-html5-sanity-perl
> 
> Jonas, do you want me to file bugs about this? 

No need.  Thanks to your analysis here I can quickly fix those package.


> > * libhttp-lrdd-perl
> 
> Hmmm, I'm not too familiar with this code... However, it doesn't 
> appear to ever instantiate an object of the HTML::HTML5::Parser 
> class... It might be able to be fixed, but the LRDD spec died, I 
> think, so I think we can conclude it is unmaintained.

I will request removal for that package, then - as I recall nothing else 
depends on it.


> > * librdf-rdfa-parser-perl
> 
> Soooo, this is the hard one... It can use several parsers, so 
> HTML::HTML5::Parser isn't strictly required, but it is the most 
> reasonable choice. One may argue that a recommends is appropriate, but 
> suggests is certainly not. I'm inclined to suggest leaving it as it 
> is, and hope something hops up to fix #750946.

I agree: Recommends is accurate here.


Thanks a lot for your close inspection - very helpful!


 - Jonas

-- 
 * Jonas Smedegaard - idealist & Internet-arkitekt
 * Tlf.: +45 40843136  Website: http://dr.jones.dk/

 [x] quote me freely  [ ] ask before reusing  [ ] keep private