STR:

1. View attached testcase (ideally on a display without pixel-scaling -- i.e. with HiDPI/Retina turned off in display settings).

EXPECTED RESULTS:
A black (or antialiased-to-gray) border of a square next to every number.

ACTUAL RESULTS:
I see a square down to 0.51
0.5 has the lower-right corner of a square.
Beyond that (scale values 0.49 and below), there is no square at all.

I suspect this is ~known outcome of our pixel-snapping behavior for borders. Pixel-snapping is nice for making sure that border-drawing is crisp, but when it makes borders fully-disappear, it's a pretty unfortunate outcome.

(Note that this is in the presence of explicit transform-scaling, where some fuzziness/smearing would be acceptable and arguably preferable)

This bug (I think it's the same bug?) results in e.g. attachment 9279462 [details] rendering without-any-borders-at-all for me in Print Preview, which is unfortunate.

(In reply to Daniel Holbert [:dholbert] from comment #0)

1. View attached testcase (ideally on a display without pixel-scaling -- i.e. with HiDPI/Retina turned off in display settings).

(Note: HiDPI / pixel-scaling helps a little here, letting us paint thinner borders, but it doesn't fix this; it just changes the threshold.)

(In reply to Daniel Holbert [:dholbert] from comment #0)

This bug (I think it's the same bug?) results in e.g. attachment 9279462 [details] rendering without-any-borders-at-all for me in Print Preview, which is unfortunate.

Here's a screenshot to demonstrate this, FWIW.

Looks like we've got bugs on file for this. bug 1258112 seems to be the oldest; duping.