On Monday 13 June 2005 1:27 pm, Christian Lohmaier wrote: > Hi *, > > On Fri, Jun 10, 2005 at 09:49:12AM -0700, rajeev joseph sebastian wrote: > > [...] > > Any ideas where I can look to see where > > the layout is actually done for latin ? > > Thanks to issue 50571 I now have a pointer for you :-) > > Have a look at vcl/source/glyphs/gcach_layout.cxx > one of the comments reads: > > // get layout engine matching to this script > // no engine change necessary if script is latin Yes I had already seen that. I changed the line with latnScriptCode to USCRIPT_CODE_LIMIT, so that it would use ICU to layout latin. Unfortunately, it didnt work.
I am also testing ICU layout engine seperately at the moment (in icu 2.6 you can see the following directory icu/source/test/letest/), just in case there is something wrong with the underlying library. That is the method I used for testing ICU's Indic shaping engine. For latin though, my GPOS tables are *not* used for some odd reason: odd because it does seem to be searching the gpos tables, and for mark records, and even though I have defined the marks properly in Fontforge, they are not loaded (I verified this by opening the generated .ttf font in fontforge and checked for the anchors). Another wierd thing is, it searches for GSUB tables for "akhn" features !!! akhn (akhand) is supposed to be used only for Indic, but it is being searched for shaping a latin sequence. I am using a slightly changed version of letest, which prints the width used, but I am not able to see the right width for a sequence in latin (which is, a with macron). For these reasons, I didn't want to speak about these tests earlier, since I am not very sure about my results (or my experiments) yet. As far as I have heard, ICU *does* position latin correctly, but for some reason, I do not see those results. Rajeev J Sebastian --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
