On 15-08-25 05:29 PM, Khaled Hosny wrote: >> but I'll work something out to match Uniscribe to the extent possible. >> > I suppose the later connection shall win, ie. gid8 stays on baseline and >> > everything else adjusted to follow. > Yes, that is what Uniscribe is doing and my “expected” result.
So this essentially becomes an immediate-mode alignment, whereas the RTL bit decides which side of the join moves... I'll see if I can shoehorn that into our model. In HarfBuzz we defer vertical adjustments for cursive to until after all lookups are done. It makes it more efficient to do it that way, but makes things like this harder... _______________________________________________ HarfBuzz mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/harfbuzz
