bagder/http2-explained

Headers in wrong order

gaopinghuang0 opened this issue · 3 comments

The headers of the gitbook are in wrong order, as shown in the screenshot below:

image

I took a look into this because it was bugging me, and it looks like the situation is this:

  1. the book was written the OSS Gitbook (3.2.3), which supports multiple languages at top level and uses a SUMMARY.md inside each language to determine chapter order
  2. it's being hosted on gitbook.com, which uses "Gitbook 4", which is a significant rewrite and which is not backwards-compatible; specifically, it no longer supports multiple languages, and it requires a .gitbook.yaml at root level pointing to a project-wide SUMMARY.md

(Re (2), see my two-language example from this branch.)

So there's a few ways to fix it depending on how @bagder wants to go forward:

  • commit to using Gitbook.com, update the book to Gitbook 4, and have a single SUMMARY.md which covers all languages, and chapters in a language are effectively subchapters of the language "chapter"
  • If @bagder has a legacy.gitbook.com account, host the book as-is on that instead of on "new" gitbook.com, but that's gonna run out of time someday, and you can't sign up new accounts there now
  • Continue using OSS gitbook, hosting the book somewhere other than on gitbook.com, which would probably involve build tooling.

I really appreciated having this book available and would be happy to do the work to get the formatting fixed up, once I know which direction to head in!

Thanks for this @wlonkly. I wish gitbooks would have been up front and clear with this when they pushed me into switching away from the legacy version into this. ☹️

I don't think going back to legacy is a way forward, and I don't think hosting it myself is the short term better answer, so let's go with option A) - fix the SUMMARY.md file to get all the languages into a single one! (that seems like such a stupid design decision but hey, I don't make those calls).

Your help will be much appreciated! (The same problem exists in http3 explained)

Will do -- I'll get a PR up. Once we're happy with how this one turns out, I'll take a look at http3.