Road to Lektor 2.0

by Armin Ronacher on Thursday, March 10, 2016

It's great to see what people have been building with Lektor so far. It also shows us what still needs to be built. For a lot of things we improved for Lektor 2.0 which is going to release very soon indeed. To give you some ideas what's going to change here is a current brief look into the changelog and what has been changed so far.

Page Discovery

One of the biggest frustrations users have voiced is the inability to control discoverability of pages. In Lektor 1.0 you can only hide pages entirely in which case the build process skips over them but there was no way to hide them by default from queries. While you could always hide pages by changing the query, this did not work for pagination and it also required you to be quite careful with the queries you are writing in templates.

In Lektor 2.0 we introduced the system _discoverable attribute which allows you to easily hide pages from any query. Queries can explicitly include undiscoverable pages but you do not need to take care of this yourself. This makes it possible to automatically hide drafts from blogs for instance. Lektor will still build it but without knowing the URL you cannot see it on the overview.

Virtual Paths

The biggest change in Lektor 2.0 is the introduction of virtual paths and sources. This is somewhat of an under the hood change but it has big implications on what is possible with Lektor plugins. Each page can have virtual resources below it that can be provided by plugins. These virtual resources are separated from the page through what is called a virtual path which is indicated by the at-sign (@). For instance this blog here uses a plugin which provides a blog archive available at /blog@blog-archive and the year 2015 is available at /blog@blog-archive/2015 etc.

Next / Previous Page

Lektor 2.0 implements sibling support through it's virtual path system which allows you to refer to the next or previous record easily. This is for instance used by this blog here to link between blog posts.

Improved Alternative Support

Alternatives were heavily improved. Individual fields that are absent in content files now fall back to the primary content file. In addition it is now possible to ask Lektor about which alternatives exist for a given source or in total. This simplifies handling of internationalized pages greatly but more work will be done in that field.

Improved Plugin Support

Plugins now have the ability to do a few more things they could not do before:

  • custom field types (want to render reStructuredText? You can now build a plugin)
  • custom build programs: because of the virtual source and path support you can now build custom build programs that build things that do not exist in the source tree. For instance you can build feeds, blog archives etc.