On 13-09-27 05:03 PM, Jonathan Kew wrote: > On 27/9/13 19:14, Behdad Esfahbod wrote: >> Thanks Jonathan. Can you test my alternate approach? Attached. > > Close, but not quite. Move the statement > > j++; > > to the very end of the loop -- if you increment it too early, you'll end up > propagating the first glyph to all the following positions, which is not at > all what we want! > > With that change, it appears to work as required.
Thanks. Will do. > IMO, it's a bit trickier to understand than my version; it should be a bit > more efficient in the case where (a) space is not available, and (b) there are > multiple stray ignorables, but that should be such a rare situation that I > didn't think the performance issue was worth worrying about. > > Anyhow, I'll leave it up to you; either patch should solve the issue. I'm allergic to O(n^2) code in the shaper... > JK > > -- behdad http://behdad.org/ _______________________________________________ HarfBuzz mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/harfbuzz
