This is a Rails application, so you can view the site locally by running rails server. Easy.
Note that the app does not require a database; content is stored in snip files under the soups directory. See below for details.
Set TINIFY_API_KEY and run bundle exec rake assets:compress_images.
When commits are pushed to any branch on the GitHub remote, the specs are run in a GitHub Action workflow.
When commits are pushed to the main branch on the GitHub remote and the specs pass in the GitHub Action workflow, the app is automatically deployed to Heroku.
So that exceptions are reported to Rollbar. You will need to create a new project within Rollbar in order to obtain the API key, which is project-specific.
$ heroku config:add ROLLBAR_ACCESS_TOKEN=<rollbar-access-token>
$ rails secret | pbcopy
$ heroku config:set SECRET_KEY_BASE=`pbpaste`
A snip is just a text file within the "soup", like "here-is-my-stuff.snip". You can write whatever you like in the file, and it will be rendered as is when you visit http://yoursite/here-is-my-stuff.
By default snips will be rendered as HTML. Snip files with a .markdown extension will be rendered using the kramdown gem. You can use ERB tags within any snip. Metadata is stored as YAML at the bottom of a snip.
This site uses multiple soups to make it easier to organise the content. Within the soups directory there are snips, and also sub-directories; These contain other snips, but don't result in nested URLs. In other words, if a snip called hello.snip is in soups/blog, it is still served from http://yoursite/hello, and not http://yoursite/blog/hello. The URL space is flat.
The layout for a snip will default to 'application' unless the render_as attribute is set to 'Blog' in which case it will default to 'blog'. You can override the layout by setting the layout attribute.
Only snips with kind set to 'blog' or 'show-and-tell' are included in the RSS/Atom feed.
Setting the draft attribute to true will mean the snip is not included in any lists or in the RSS/Atom feed.
The site map is generated on the fly and includes any page with the is_page attribute set to true.
If you're not sure, just look at some of the other snips in soups/blog.
There are a set of Rake tasks under the 'regression' namespace which allow generation and comparison of a set of normalized HTML/XML artefacts. Note that these Rake tasks depend on HTML Tidy 1 which can be installed via brew install tidy-html5 on OSX.