We are explicitly using the site's RSS feed, but after reading the
Hugo RSS docs I think it's better if we use the contextual version
of the RSSlink. This will make the RSS plumbing show a feed for the
current category, tag, section, etc. The example code also adds a
bit of markup to help browsers find the content more easily.
See: https://gohugo.io/templates/rss/
Instead of writing my own logic for metadata and Twitter cards, I
should be using Hugo's own. This makes my theme code cleaner, but
means I have to adjust some of the configuration conventions that
I have been using. An added bonus is that my theme will work with
more sites now, as long as they are using Hugo's config conventions.
Sites only need to add their Google Analytics tracking ID to their
config:
googleAnalytics = "UA-123-45"
This uses the async version of the Google Analytics code.
I had added Google Webmaster verify support but it got lost when I
was re-organizing the head meta stuff. I'm adding it back now plus
support for Bing and Yandex.
In a multilingual context a post's title could be "Title" in English
but «Title» in Bulgarian, and if we apply the Markdown filter to the
title tag, then the language's Black Friday configuration is used,
in this example it would be for the angled quotes.
We already do this everywhere else we can in the post's content, so
let's match it in the page title.
Hugo's own pagination stuff works just as well and probably covers
even more corner cases. Also, this is one less config variable to
have to check because you can just use Paginate (Default: 10).
Add categories/tags to front matter and they will be displayed on
summary and post pages, with links to taxonomy pages. Example:
+++
date = "2016-09-24T21:28:31+03:00"
title = "Post title"
categories = ["Nature"]
tags = ["Wetland", "Oasis"]
+++
For some reason if you add tags singularly, like:
categories = "Nature"
you get index errors from Hugo. Not sure if I need to parse the
tags differently or just add more logic to test if the terms are
singular or not.
Check if the params.sidebar is defined before checking to see if
params.sidebar.hide is true or false. New sites might not have
their config set up properly, so this avoids an error in the case
that params.sidebar isn't defined yet.
Advertises the XML sitemap and allows all user agents, but is only
enabled if the site sets this property in their config:
enableRobotsTXT = true
See: https://gohugo.io/extras/robots-txt/
By default they are on, but you can disable them by adding the
following configuration value:
[params.sharingicons]
hide = true
If this parameter is undefined (or "false") then the icons will be
shown. Furthermore, you can disable specific icons by setting their
value to "false", ie:
[params.sharingicons]
facebook = false
Any icons not named will be shown. Currently there a four icons
configureg: facebook, googleplus, linkedin, and twitter.
Uses basic data like title, description, author, and date that we
are using with existing vanilla meta tags, but extends them to OG
and Twitter Card tags. See [0] and [1].
For the Twitter summary cards specifically, you can optionally add
attribution for your username by adding something like this to your
configuration:
[params.social]
twitter_username = "@mralanorth"
... and for posts with images, you can specify an "image" in the
post's front matter like so:
image = "/2016/09/IMG_20160916_174409.jpg"
And then the theme will opt to use Twitter's "large" summary card.
[0] http://ogp.me/
[1] https://dev.twitter.com/cards/getting-started
Still need to figure out how to determine which page we're on so we
can set the "active" class accordingly (for Home we have the IsHome
variable the we can check).
Both work effectively the same for my use case, but the "type" is
more obvious when looking at the code. See the documentation for
Hugo's page variables for more info[0].
[0] http://gohugo.io/templates/variables/#page-variables
Use the post's author from frontmatter, or else use the author from
the site's config. You MUST have one of these set or else you will
get an error during site generation. I think it's better to force
the user to define an author tag rather than only printing it if
it is defined because it is a good practice to help bots understand
content.
Use the post's meta description if it exists, otherwise use the
one from the site's config. Set them using the "description" key
in frontmatter or site config.
Note: this means there is no way to NOT have a description. You
must have *at least* a site-wide description and *optionally* a
description for each page/post's front matter.
The W3C's HTML5 documentation says that header strength (h1–h6) is
only important per section, but their validator[0] recommends only
using one h1.
[0] https://validator.w3.org/
By default it shows the latest five posts, but you can set this
property in your config to override it:
[params.sidebar]
num_recent_posts = 7
Still no way to disable it, as I don't understand golang's HTML
templating stuff yet. The Hugo docs are a bit confusing, but it
seems like conditionals are a bit tricky because both false and
0 return a boolean false... hmm.
See: https://gohugo.io/templates/go-templates/
Add something like the following to your config to enable the about
block in the sidebar:
[params.sidebar]
about = "*My site* is really cool. Adios!"
Can be formatted in Markdown.
I need to figure out how to generate a list of posts grouped by
month, ie "August 2016", as well as a list template to show the
posts for each month for when the user clicks the link.