[Editorial] 'policy' is a local term, not global.

[mikewest]

@antosart WDYT of this? In short, we're linking to the wrong term as for because I misunderstood what Bikeshed did with local-lt.

---

Preview | Diff

[antosart]

This seems to break all internal links. E.g. on line 486 [=policy/disposition=] doesn't resolve anymore to the local spec, while [=content security policy object/disposition=] does.

It's not really clear to me how local-lt behaves. Maybe it's easiest to just give up and use the long link everywhere?

[mikewest]

That surprises me. :/ @tabatkins might have insight into the expected behavior, but I'll just use the long name for the moment (which looks like it produces the expected output).

[tabatkins]

Note that the for value on dfns/links is purely an opaque namespacing mechanism. It has no inherent connection to any other definition, except by convention. (IDL attribute/method definitions, for example, expect their for values to be the name of the interface, etc.) If a term like "content security policy object" has some associated definitions that you want to link together (and hide from the global namespace), you can do so with whatever for value you desire; Bikeshed does not care whether there is an actual dfn with that term, "local" or not.

[mikewest]

Ok. I'll land this, and rethink the way I'm using lt here and elsewhere... :/