On Thu, Mar 22, 2007 at 10:41:48PM +0000, José Matos wrote: > On Thursday 22 March 2007 10:06:38 pm Angus Leeming wrote: > > Enrico Forestieri <[EMAIL PROTECTED]> writes: > > > The attached patch corrects an oversight in implementing rule 18a in > > > appendix G of the TeXbook. I attach a .lyx file demonstrating what the > > > problem is and three screenshots showing the appearance on screen before > > > and after the patch, as well as the dvi output. > > > > I do believe that you're enabling LyX to produce better looking stuff than > > dvi! Way to go, Enrico! > > I agree. :-)
Thanks for your appreciation. However, I must say that mine are finishing touches to an already well thought architecture. I applied a slightly improved patch: http://www.lyx.org/trac/changeset/17518 Unless I overlooked something, what remains to be done are small adjustments to the MathFontOld and MathUnknown insets, but they are really marginal. Now I am thinking to try some optimization. The position of super and subscript must be jointly computed and this is done in the dy01 method. However, dy01 is called by both dy0 and dy1, so it is unnecessarily called twice when a super and subscript are both present. Honestly, I must say that I did not notice a slow down with respect to the previous strategy, so I am asking for opinions. Do you think that such optimization is worth the trouble? -- Enrico
