Header IDs should use same deduping as the contained anchors

[juliusv]

For example, in https://prometheus.io/docs/introduction/comparison/, half of the links are pointing at the wrong anchor.

[juliusv]

Also, on https://prometheus.io/docs/introduction/faq, the last link ("Why does Prometheus use a custom storage backend rather than [some other storage method]? Isn't the "one file per time series" approach killing performance?") is broken.

[grobie]

How is this related to the TOC? It just uses the ids defined on each heading.

[juliusv]

@grobie Ah, didn't notice that the IDs were not properly deduped either. Only the anchors themselves are! Ok, so this is not TOC-related. Will fix title.

[grobie]

I'd suggest to use the full heading path, so h2 "Prometheus vs. Graphite" -> h3 "Storage" becomes #prometheus-vs-graphite-storage.

[grobie]

This will break existing links, but users would still end up on the same page. This change will help to provide more descriptive links in the long run though, so I'd rather do that then extend on the *-{counter} model.

[juliusv]

Yeah, that'd be better. You still have to have deduping ontop of it just to be sure...