Docs.krita.org has been switched to sphinx!

[Boudewijn Rempt]

On zondag 27 mei 2018 15:09:21 CEST Wolthera wrote:
> Hey everyone,
> 
> https://docs.krita.org/ has been moved to sphinx, this technology switch
> means that we'll be able to provide offline copies, translation
> possibilities and in general the pages will be faster and safer due to
> being static with a little whiff of javascript for using the search.
> 
> You can thus now go to https://docs.krita.org/fr/KritaFAQ.html for the
> French version of the FAQ, and for the english language epub you can
> currently go to https://docs.krita.org/en/epub/KritaManual.epub
> 
> Now, there's a lot of things that need discussing. To the point I am not
> sure if the meeting of tomorrow will be sufficient. Things I cannot figure
> out by my own:
> 
> 1. 404 pages. All the old links won't work anymore, and we don't even have
> a custom 404 page. How should we go about resolving this.

I guess we can easily get one, that explains the move. 

> 2. How should we decide upon a draft/proofingreading workflow?

Let's use phabricator for that, in principle? I guess there are people who 
should just commit and push, like you, but others probably want to have their 
changes looked at. Maybe Irina likes to do some proof reading?

> 3. How should we decide the translation process.
>    The translation process is to generate pot files with gettext, then to
> generate po files, have those translated, and then sphinx can generate mo
> files and use those for a language specific build.
>    3.1. Where should the PO files go?

I think we should put that question to the i18n-doc list.

>    3.2. Maybe we should remove certain pages(contributor's manual) from the
> POT files with some kind of script?

Good point...

>    3.3. Where do we want the epubs to go, and should we try to name them
> according to language?

Maybe even more importantly: where shall we make them available for download?

> 4. Protocol for removing pages.

That should always be a diff on phabricator, I think. It shouldn't be 
something people can do on a whim.

> 5. EPUB files are build by jenkins too. I've configured the conf.py so that
> we get git commit hashes in the EPUB identifier and the html footer when
> you build locally, but Jenkins doesn't have access to the git commit for
> technical reasons(only the build url). How do we figure out which version
> of the manual the epub is build on when its made by jenkins?

That I don't know...

> 
> Anyway, those are some of the basic points I am a little confused about.


-- 
Boudewijn Rempt | https://www.valdyas.org | https://www.krita.org