webref
webref copied to clipboard
w3c
Linter to prevent xref-ing unexported terms?
I think some generalized warning system for cases like https://github.com/w3c/input-events/issues/129 will be helpful, and webref is a good place for that to happen.
Or, if bikeshed never has this problem and it's actually ReSpec-specific, maybe @sidvishnoi could implement a linter there?
https://github.com/w3c/respec/issues/1360 is probably the relevant ReSpec issue here.
If Bikeshed also faces such issue, this might be better suited for spec-prod. Edit: Or implement here and spec-prod can use it like it does for webidl.
Webref contains the data needed to create the warnings. We've split Reffy's codebase into two parts not long ago: Reffy in itself being the crawler and Strudy being the analyser. The analyses that get produced have the right info but they are somewhat useless right now in that they don't create warnings and thus are ignored by basically everyone in practice:
- External links to private terms - 839 problematic external links found in 195 specifications in the latest report
- External links that neither targets definitions nor headings
- Broken external links (which contains lots of false positives because the tool does not like links to the single page version of the HTML spec.
Anyway, we are thinking about creating a more useful warning system that actually raises issues against the specs when needed, see https://github.com/w3c/strudy/issues/7
