* Add ignored_content to the Config structure.
* Use the GlobSet crate to parse the glob patterns into a matcher, which
  is created once at program initialization. If there are no patterns in
  ignored_content, an empty globber is created, which excludes no files.
  This is consistent with the existing behaviour of Gutenberg, before
  this feature was added.
* Bail if there are any errors in the glob patterns.
* Add a call to the globber in page.rs to actually do the filtering.
* Update documentation.

A note on the Config structure
------------------------------
* I had to remove the PartialEq derive from the Config structure as it
  does not work for the GlobSet type. No harm is done, Config does not
  need to be PartialEq anyway, since there is no need to sort Configs.
* The implementation follows the pattern of the existing config settings
  in that it uses an Option<...>. This would appear unnecessary, in that
  an empty vec could be used as the default, but it appears to be needed
  by the TOML parsing. A better approach would be to use a separate
  SerializableConfig and map to/from a Config struct. This would also
  allow the elimination of most, if not all, of the other Options in
  the Config structure, but that ought to be another PR.

And fix bug with taxonomies urls

This removes the dependency cycle between config and rendering that
causes 4 packages to be recompiled every time a change is made.

I just want to code fast!

And some tweaks as I write the docs

Fix #108

Fix #104