Turtle Art blocks should autoexpand when necessary

[walter]

Any block that takes arguments has the potential of occlusion. This was addressed in part by #2330. I'll add the same expansion functionality to all the blocks that take more than 1 arg (e.g., setxy. setcolor, etc.) but it would be even nicer if this was an automated process.

[walter]

Works for everything except BOOLEAN_STYLE when docking (​http://git.sugarlabs.org/projects/turtleart/repos/mainline/commits/920ad48dbea89b4801259aac10e58efd9692038c)

Still need to do AND and OR blocks. Also need to consider whether to auto-expand when the size of the block changes while already docked. (Probably not at this point.)

[walter]

took care of last of the bugs referred to in #2330

[walter]

There is a still a little bit quirky behavior with the boolean blocks, but I think this is about as good as it is going to get this go-around. I think it is a big improvement over #2330.

[tonyforster]

blocks not expanding

[tonyforster]

math blocks do not always autoexpand see screenshot

[tonyforster]

On grow/shrink blocks, expansions are miscalculated. They can be corrected by redocking.

Overlap a little, maybe 20% when shrunk, leave gaps maybe 20% when expanded

[walter]

math blocks do not always autoexpand see screenshot

Hmmm. This is a general class that I don't currently handle at all: when blocks from the lower of the two connectors hang below. I'll have to think about whether it is worth tackling, as it means I need to examine a lot more branches every time one docks a block. (This can be worked around by #2330 or by adding a vertical space in the case of blocks that don't expand.)

On grow/shrink blocks, expansions are miscalculated.

This is probably due to the scaling factor not being properly applied to the gap. I'll look into this as well.

[tonyforster]

math blocks do not always autoexpand - I agree, probably not worth fixing. It does answer your question in #2330 though. Keep the expand collapse handles.

[walter]

On grow/shrink blocks, expansions are miscalculated. They can be corrected by redocking.

Fixed this one: ​http://git.sugarlabs.org/projects/turtleart/repos/mainline/commits/36523c9b6caad48219f3a0a94b652115aaa459f7