- Avoid fighting import cycles using a factory function
- Collapse useless subpackages into simple modules
- Move flask plugins / infrastructure in own module
- Refactor how blueprints are used to localize information about
URL routing to app factory
- Support for blockquotes
- Support for lists (ol, ul)
- Make tertiary buttons more obvious by using an underline
- Extend demo to showcase new features
We don’t have proper dependency tracking for those in Make, and
I don’t care to invest a lot into getting this right, given that
the build for all of them finishes in <1s.
The raw values now map to names describing their hue. Only when
mapping them to variable names, we assign semantic meaning. This
allows easy re-use of colour values between different semantic
meanings and changing e.g. accent vs. primary without affecting
the use as warning.
- Fix etag attaching (add_etag is actually a coroutine which hashes
the data payload)
- Add expires header (with now + 1800s default) so that we don’t
get hit with an avatar request on each load -- also helps with
page responsiveness.
- Proper handling for HEAD requests.
- CSP to prevent funny SVG attacks.
If a session cookie is set, but prosody doesn’t know about the
session anymore, we could get into fun states. This patch fixes
them by requiring the session to be tested with a ping request
on each HTTP request.
This allows us to include _theme to get all the variables. The
idea is that we may, in the future, be able to use a few different
CSS files if we have complex (= large) features which are only
needed rarely.
- Create a colour palette
- Create a sizing schema for paddings and fonts
- Implement basic form controls
- Create a theme demo page
- Apply the theme to the existing pages.
Still TODO is the final font selection.
