Right now dictionaries are not buffered, but are rather read in their entirety before being used to instantiate a User:
Offending helper method that reads dictionaries.
Here the result is used to instantiate a User.
This is a massive resource hit and could likely be done better but I'm not quite sure how. Hopefully we don't need to wait until the web-extensions API exposes the actual browser spell checker (which would remove the need for nspell entirely) but that may be the case.
Right now dictionaries are not buffered, but are rather read in their entirety before being used to instantiate a User:
Offending helper method that reads dictionaries.
Here the result is used to instantiate a User.
This is a massive resource hit and could likely be done better but I'm not quite sure how. Hopefully we don't need to wait until the web-extensions API exposes the actual browser spell checker (which would remove the need for nspell entirely) but that may be the case.