Re: [Rfc-markdown] [rfc-i] How do drafts get section links?

[Jeffrey Yasskin <[e-mail]>]

On Thu, Dec 7, 2017 at 4:47 PM Carsten Bormann <[e-mail]> wrote:

> Hi Jeffrey,
>
> On Dec 8, 2017, at 01:11, Jeffrey Yasskin <[e-mail]> wrote:
> >
> > What's the right way to get deep links to sections from kramdown-rfc2629
> source?
>
> (Sections of other documents, that is.)
>
> RFCXML version 2 has no way to do this, so kramdown-rfc doesn’t support it.
>

Yeah, I mostly mentioned kramdown-rfc to make sure any answers like "use
https://www.greenbytes.de/tech/webdav/rfc2629xslt/rfc2629xslt.html#ext-rfc2629.xref"
included instructions for how to get the right attributes into the XML from
.md source.

So why are you seeing links on tools.ietf.org (and only there) where there
> aren’t any?
> The .txt to .html converter at tools.ietf.org has some heuristics that
> allow it to identify references to sections of other RFCs.  These
> heuristics sometimes fail to notice (and famously sometimes point to the
> wrong document, leading to spurious errata about problems that aren’t even
> in the RFC), but they are better than nothing.
>

Ah, I see: https://tools.ietf.org/tools/rfcmarkup/ does this for URLs under
https://tools.ietf.org/html/, but the id/ URLs only use xml2rfc or
rfc2629.xslt and so don't get the benefit. That answers my question of how
to test that my source gives the right section links, and in fact
https://tools.ietf.org/html/draft-yasskin-http-origin-signed-responses-01#section-2
has decent deep links, while
https://tools.ietf.org/id/draft-ietf-httpbis-origin-frame-04.html#rfc.section.1
doesn't.

I'm happy to conclude that my mistake was in sending out the wrong links,
and that this list doesn't need to do anything until RFCXML v3 support
comes along.

Now how is the syntax for deep links into documents going to look like in
> kramdown-rfc when we add RFCXML v3 support?
> I don’t know yet.  But I note that essentially the same heuristics that
> work on .txt files on tools.ietf.org might work in a markdown source
> document (and, if they don’t, the author can still set a manual xref [er,
> relref, if that thing stays in v3]).
>

(https://github.com/rfc-format/draft-iab-xml2rfc-v3-bis/issues/26)

(CCing rfc-markdown, as the question is of interest there, too; please trim
> the CC list if you just want to vent about relref, as you probably should
> at this point.)


Thanks,
Jeffrey