Adjust line to account for exclusive upper bounds

[jbcrail]

When extending a line, I decremented the upper bounds for both x and y.

Fix #328

[jbednar]

Thanks! Can you explain the reasoning behind this fix? It's clear a change needs to be made, but changing xmax and ymax affects the slope of the line segment, not just its length, and I just want to make sure how you chose this fix rather than leaving the maxes as they are but then changing the comparison. I would guess the correct answer can be determined by hand-plotting some very short lines in small arrays, and if so it would be good to include those here to demonstrate why this is the correct fix. Or is that already what's in the tests (not sure where those matrices came from)?

[jbcrail]

I tried two strategies for this fix. I focused on extend_line since it was the only method that used the bounds, instead of just passing the value along. My intent with both strategies was to shorten the bounding box for the clipped area.

First, I changed the comparisons in _compute_outcode for both maximum bounds from > to >=. This seemed like the least intrusive approach while still accounting for the exclusive upper bounds. Unfortunately, the while loop in extend_line never exited for several tests.

Second, I decremented the maximum bounds prior to clipping the segment. This approach worked better and is the one in this PR. The modified tests accounted for the newly clipped segments at the exclusive upper bounds. For instance, one test had a top-left to bottom-right line segment that was clipped in the bottom-right corner as expected after the fix.

I missed the change to the slope of the line segment. Ideally, I'd prefer the first strategy but I don't know the code or the algorithms well enough to know what changes are needed to make it work.

[jbednar]

Let's go ahead and merge this to avoid the segfaults, but I'd still like to compare to an example worked out on paper before release.