Phil, The only change between the two examples is the name of the font. The testing code is pretty simple, and is derived from a swing tutorial example. In order to test a different font or size I have to change the code and rebuild. Literally the only change between the two is the name of the font. I'll look at capturing the set of GlyphVectors from these runs and presenting them separately, as a small sample.
The most interesting thing is that the hyphen seems to "anchor" the end of the text. I notice that as I resize and re-lay the text, the distance between characters within a word varies. The individual GlyphVectors are calculated once; the line breaking algorithm then simply makes use of their measurements. I expected that the letter spacing would not vary in presentation, but something in swing is deciding to fiddle with the character spacing. Peter On Fri, 2006-06-23 at 11:19 -0700, Phil Race wrote: > I don't know of any problems measuring italic fonts. > I suspect its a bug in your code but have no idea what. > Also the zero width spaces should be mapped to an invisible glyph id > but I don't know how you are creating the glyph vectors. > I can't imagine anything other than a *small* complete sample program > will help > understand this. > > -phil. > > Peter B. West wrote: > > >How do I measure the boundary of oblique text? When I layout upright > >text using a combination of logical and visual bounds, I get great > >results. When I use the same algorithms on the oblique version of the > >font, my measurements are all over the place like a dog's breakfast. > >It's not just the "lean" on the last character of the line. I'm missing > >one or two characters completely. > > > >The curious thing is that where I end a line with a hyphen, the > >measurement is OK. > > > >The line is constructed from a series of GlyphVectors, one per word or > >word fragment, where a word contains a hyphenation point. (The artifacts > >in the text are the font's representation of a ZWSP, much to my > >disappointment.) > > > >Any suggestions? > > =========================================================================== To unsubscribe, send email to [EMAIL PROTECTED] and include in the body of the message "signoff JAVA2D-INTEREST". For general help, send email to [EMAIL PROTECTED] and include in the body of the message "help".
