On Thu, Sep 06, 2007 at 11:25:04AM -0700, Alan Irwin wrote: > Hi Andrew: > > I have been doing some extensive comparisons between -dev psc (text=0, i.e > Hershey fonts), -dev psttfc, > and -dev pscairo for ./x07c, and the results have been quite interesting. > > First I have sorted out a number of issues with the Hershey to unicode > translation table. > > Second, I found and fixed a bug in cairo.c concerning the font choice. > It now chooses the same fonts as psttf.cc rather than what I think > were emergency fonts chosen by fontconfig when it didn't get the right > information. > > Third, there is a bug in the vertical position of psttf. pscairo gives the > same nice vertically centred position as the Hershey fonts, but psttf > results
It appears that this is not actually a psttf bug. It also appears with the ps driver if you plot postscript symbols instead of hershey symbols (off by default). Problem is that the baseline is set to be 1 for plotting symbols. Some drivers (like ps, psttf) tried to use this, which resulted in the incorrect placing. Others (like gd) ignore it and so work. I've now set base = 0 in plhrsh. It works for the drivers I have tested - please check with your favourites! > are uniformly too high. > > Fourth, for default fonts on my system (with libLASi built against a modern > pango stack), I get the following segfault: > > [EMAIL PROTECTED]> ./x07c -dev psttfc -o test07.psttfc1 > PANGO is returning a glyph index of 10002980 > Segmentation fault <Details deleted> I can't reproduce this with Ubuntu feisty. I get the glyph error (correctly) but no segmentation fault. Andrew ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2005. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Plplot-devel mailing list Plplot-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/plplot-devel
