From: <pl...@pi...> - 2007-05-25 15:41:31
|
On Thu, 24 May 2007 18:10:50 +0200, Ethan Merritt <merritt@u.washington.edu> wrote: > On Thursday 24 May 2007 08:00, pl...@pi... wrote: >> > Try this as a test page: >> > >> > http://gnuplot.sourceforge.net/demo_svg/utf8text.html >> >> That page seems to come up just about the same in Opera and firefox for >> me >> (running Gentoo linux untested on win) except that Opera scales fonts >> continously and smoothly. Opera maintains appropriate font size for the >> axes lables as display gets bigger. FFx quickly makes a mess of it with >> huge fonts that overwrite each other and spread onto the graph itself. > > I agree that the rendering in Firefox is worse than either Opera > or Konqueror (ksvg). However both Opera and ksvg have at least one > annoying bug that can be seen in the vertical positioning of the > expression > inside the integral on the test page. It should share a baseline with > the integral sign itself but instead it shows up level with the > superscript '∞'. The bug is that if a text chunk > contains positioning information, but the text consists only of > spaces ('\040'), then the vertical positioning is ignored. > This leaves the "current position" in the wrong place vertically > for any following text chunks. > > Opera also fails to render transparent pattern-fill correctly. > Hi , looking at your utf8 test page and any output from webify.pl , I dont see any definition of background colour or opacity. Neither do I see any comment about the background of a plot in gnuplot help function. Where does the very light grey b/g I see using ksvg come from and why is Opera's transparant redering wrong? TIA. |