Merge upstream v8.5.6 #171
Conversation
also updates the setup-graphviz action, but I'm not sure this will get rid of the deprecation warning in the workflow runs about using "Node.js 12 actions"
| {document.tags && document.tags.map(tag => ( | ||
| <span class="md-tag">{tag}</span> | ||
| ))} | ||
| {document.tags && ( |
There was a problem hiding this comment.
I guess we could see about supporting tags?
There was a problem hiding this comment.
I'm far from an expert on the search mods. Now that I think about it, that is something I didn't check before pushing.
There was a problem hiding this comment.
supporting the tags in search would be more work. But there is also a mechanism to just display the tags on the page that would be simpler to support --- we would just need a way to specify them as part of the file metadata. Maybe without search integration they aren't very useful though.
There was a problem hiding this comment.
I'm leaving this as is for now. If someone asks, then we can discuss it again (or just create a new issue as a reminder).
|
Thanks for doing this upgrade! |
|
If you don't mind I'd like to keep this open for testing and further needed documentation about newer features. I'm about to push changes to docs about customizing the admonition icons. I also intend to patch the |
Sure, that's why I didn't merge it myself. |
The metadata keys using hyphens was apparently broken; I'm not sure when it was broke. So, this fixes the `:hide-*:` metadata keys by using an underscore instead: `:hide_*:`.
|
I'm now ready to merge this. I'll let it stew for a day so you have another chance to review the documentation changes (namely in customization.rst and admonitions.rst). |
|
The I'm not sure how to document the cookies portion of the feature. The JS side seems obvious to me for custom cookies, but I'm not sure the upstream default cookies are enacted. Here's the config I used to test it: html_theme_options = {
"consent": {
"title": "Welcome!",
"description": "We don't use cookies, so your privacy is safe here.",
"actions": ["accept"],
}
}and then manually remove the I think I'll just add the necessary src code to use it and leave the documentation for another time. |
2bndy5
force-pushed
the
merge-upstream-v8.5.6
branch
from
5833139
to
35fcb8e
Compare
includes external links to RST doctree specs fix indent in example snippet
approved workaround to sphinx-doc/sphinx#10932 is to use a blank line at start of doc before metadata
2bndy5
force-pushed
the
merge-upstream-v8.5.6
branch
from
26679e6
to
85b5a9c
Compare
|
Thanks, looks good to me. |
|
Great, I'll merge this and start/continue squashing open issues. I suppose we could tag this merger, but I'm unreasonably excited about code annotations (despite the implementation's current limits), and want to solve #169 before the next PyPI release. |
| @@ -661,14 +676,18 @@ | |||
|
|
|||
| // Show navigation link as title | |||
| > .md-nav__link { | |||
| position: sticky; | |||
| top: 0; | |||
| z-index: 1; | |||
There was a problem hiding this comment.
This z-index caused a regression as reported in #189
fixes #189 by using the inline CSS variable with fallback value inherited from upstream
patch_info.zip
This wasn't as painful as I had expected (due to the long time since last update from upstream), or maybe I'm just getting better at it (wishful thinking 🤣 )
Aside from the usual JS tweaks after the merge script did it's thing, I also
I had to tweak the merge script for files that are explicitly excluded (in
MKDOCS_EXCLUDE_PATTERNS) and are no longer in the upstream base/target git indexes. There's still room for improvement because I had to manually add the new files from upstream since they weren't in this repo's index. The additions were mostly language translations and a couple features implemented on upstream only (like builtin site feedback and comments). Maybe there's a way to port the new functionality to sphinx builds, but I'm not overly enthusiastic about them (currently content with directing to other sites like github for a more dynamic user interactive experience).I did discover 1 regression about mermaid graphs and filed an issue upstream (squidfunk/mkdocs-material#4517). We can merge in the simple fix for that later though.