5:51:12slyrusjackdaniel: I agree with scymtym. Nice work on the bounding regions and the text stuff in general!
5:53:41beachslyrus: Oh, today I need to write my report on Jim Newton's thesis.
5:53:56beachAnd my favorite coauthor is coming over for lunch.
5:54:17beachSo I should finish the report (it's in French) and let her read it before I ship it.
5:55:42slyrusCool! Looking forward to hearing how it went.
6:04:48jackdanieloff-by-one issues are interesting when you take glyph metrics
7:19:11jackdanielfor instance: when we compute bbox for a multiline text (and as mentioned earlier, bounding rectangle is *always* computed on left-top aligned text)
7:19:41jackdanieland we take that for right-aligned text to adjust bounding rectangle, we may have left-bearing from line which is far from left edge computed into bbox
7:20:21jackdanielfor instance "T" may have left-bearing = -1
7:22:03jackdanielso text-bounding-rectangle (and font-text-extents) either have to take alignment into account or we must artificially assemble results line-by-line taking alignment into account
7:22:31jackdanielor we may ignore this error knowing, that only snobs will notice ;-)
7:24:41jackdanielthis http://i.imgur.com/FSiUrvR.png illustrates the issue. top bounding box is perfect for certains values of perfection
7:25:25jackdanielbottom - not so much, it is too big on the left by one pixel. that is because the second line, where T grows it by one pixel
7:26:03jackdaniel(glyphs look ugly because they are hacked to have their bounding box in the pixarray)