Basic doc/docview cache for better performance (#105)

Currently, DocView:get_col_x_offset() is an expensive operation when
used on long lines. Plugins that use it extensively like: drawwhitespace,
colorpreview or selectionhighlight, cause the editor to hang or not
respond when dealing with huge lines.

Some example of files that cause the editor to hang are minified css
or javascript files. While no one would casually edit these files it
is totally normal to open them while inspecting a codebase.

For this reason a basic cache system is introduced caching the values
returned by such functions when necessary. Another slow function can
be DocView:get_x_offset_col(), so this change leverages the caching
performed on col_x_offset to perform faster x_offset_col and keep it
simple.

Beside adding the data cache, this work also introduces the ability of
getting each token starting from a specific column on the line. This
feature reduces the amount of iterations needed to calculate a column
position since we can resume from a specific column.

* Stop tokenizer if doc referenced from closed view

This stops the tokenizer when a doc belongs to a docview that got
closed. Tokenization of really long lines could make the editor a bit
unresponsive and this change gives the user the power of ending the
tokenization inmmedietly when the associated docviews are closed.

* Take into consideration utf-8

0 files changed, 0 insertions, 0 deletions