Documentation: search result links are broken...

python-hyper / h2

HTTP/2 State-Machine based protocol implementation

https://h2.readthedocs.io/en/stable

MIT License

963 stars 151 forks source link

Documentation: search result links are broken... #1216

Closed dimaqq closed 4 years ago

dimaqq commented 4 years ago

TL;DR The search results have links with extra .rst in them.

For example:

go to https://python-hyper.org/projects/h2/en/stable/index.html
search for "increment" in the search box
you are sent to https://python-hyper.org/projects/h2/en/stable/search.html?q=increment&check_keywords=yes&area=default
click the first result, "Advanced usage"
you are taken to https://python-hyper.org/projects/h2/en/stable/advanced-usage.rst.html?highlight=increment which is a 404
the correct link is https://python-hyper.org/projects/h2/en/stable/advanced-usage.html?highlight=increment

Gallaecio commented 4 years ago

The whole documentation seems broken at the moment.

Lukasa commented 4 years ago

Yeah, this feels like an RTD thing? Going directly to the RTD page (https://python-hyper.org/en/latest/) seems to get a 403. The 403 we see on python-hyper.org is just a reflection of the other one.

sethmlarson commented 4 years ago

I don't get a 403 for https://hyper-h2.readthedocs.io/en/latest/ though?

Gallaecio commented 4 years ago

Maybe the documentation link in the README is what is wrong?

Lukasa commented 4 years ago

I think it’s specifically the central page that is broken.

vmagamedov commented 4 years ago

@Lukasa I'm wondering how custom domain and SSL certificate is configured. There is an issue that you can't use Cloudflare to Cloudflare proxying: https://github.com/readthedocs/readthedocs.org/issues/4395 - many project with custom domains were affected. They even noted this issue in the docs: https://docs.readthedocs.io/en/stable/custom_domains.html. Maybe this is the reason?

Kriechi commented 4 years ago

This should be fixed now - please report any issues in a new ticket. I updated a whole set of links and references - so search engines might take a while before they pick up the new correct URLs. The remaining broken links are fixed in the "lastest" docs, the "stable" docs are based on the most recent package release - so I'm planning a new release soon to also bump the docs.