I think the Effect may make it difficult/impossible to apply syntax
highlighting to static text

The paint() function seems like the simplest way of rendering
highlighted text, and it's only available on Screen or Canvas objects.
This is why I'm not using an Effect for a code block.

We are no longer calling the _code() function, so we don't need to store
the data in the same way. The code element tells us what to render, and
the row will tell us where to place it on the slide.

_code() will be removed later.

From this function we should be able to get the language from the block,
then highlight and paint line by line

Once we get to the restart page, the animation loop keeps running and we
get an error if we try to set `cur_slide` normally.

This solves the character rendering problems

I don't like modifying slide.code_blocks in the get_effects function.
The reason I'm doing it is because that is where we determine what row
the code block will go on.

I don't want to get too carried away with this type of thing for this
PR. I'm doing this to hopefully make it more obvious where code_blocks
is coming from.

This is just a bit of cleanup for organization & clarity.

I tested this change by putting a fireworks effect on the slide and
checking to see if I got the ValueError.

This feels more in line with the Scene/Effect structure of the project.

This doesn't change the functionality - I'm just doing this since there
isn't a particular reason to remove them, and I want to make the PR diff
smaller.

A user might use a code block and intentionally leave out the language
for displaying logs or other things.