Make complex scalar selection rectangles work, partial fix for #162.

This fixes the broken logic for computing the scalar selection rectangle on the engine side, as discussed in #162 (closed). It doesn't fix the scalar bboxes which are obviously wrong on the engine side, but at least the selection rectangles are properly shown now even if the canvas scaling involves negative scaling factors and/or nonzero origins.

Mentioned in issue #162 (closed)

Added 2 commits:

• 818d512d - Fix up scalar bbox calculation of array plots, fixes #162 (closed).
• 82cc5883 - Fix some gcc warnings and dubious constructs.

Compare with previous version

Added 1 commit:

• 2154d843 - Cosmetic changes.

Compare with previous version

Accidentally committed the wrong edit of g_scalar.c to this branch, fixed in commit rev. 2154d843.

Hm, here's what I get when compiling from your branch (an attempt to select the red thingy in the middle of the score:

Works for me (this is the version of the patch included with purr-data):

Works also with the version of the patch from Pd 0.47.1:

This is on Arch. I'll do a build on Ubuntu to see whether that makes a difference...

Are you sure you got the entire shebang? Just to make sure, here are the diffs against your current HEAD:

fix-scalar-selection.diff

I get the same (correct) results on Ubuntu 14.04. Selecting any of the scalars, or just selecting all of them, works fine for me. For the record, I'm using nw.js 0.18.5 on all of my systems.

Beats me. All that I can say is that it works without a hitch for me, and I've tried a bunch of different patches now from both the 4.data.structures and the 4.data.structures/pd-l2ork/ds-tutorials section, clicking on any of the scalars in there, they all work for me.

Thanks. It turned out to be something in git stash that I didn't notice.

It works fine now. Garrays seems to have no regressions, so I'm merging...

Status changed to merged

Cool, thanks a bunch. So once the next rc is out, I will actually be able to do some nice graphical stuff with my students in the notation course, that's great!

Ok, now that this is sorted out, I can hopefully look at the multi-graph stuff (#157 (closed)) soon.