On 3/1/06, Davide Viti <[EMAIL PROTECTED]> wrote: > >> Not exactly. We have installer/build/needed-characters, that's all. > > > >It has crossed my mind several times to use that info for font > >stripping, but I never did it. > >The main problem is the fact that consecutive codes should be > >condensed in ranges and I didn't figured yet how to do this. > > > >OTOH, now I remember that I was told at some point that the > >information from those files is no longer used and this info is > >actually gathered from the msgstrs from the po files. Is this true? > > yesterday I managed to extend the spellchecker as to give infos about > the codepoints used for each language on each level; it still needs some > minor fixes and has not been committed yet.
Cool :-) Do you create a list of unicodes which are used and leave it like that, or do you compress consecutive codes in ranges? (This is important when passing long sequences like the ones resulting for CJK, Indic and RTL languages, in order not to have memeory consumption set too high). > The good thing about it is that > the > informations are up to date with the very latest version od the po files. ... and that we can strip as much as we can ;-) -- Regards, EddyP ============================================= "Imagination is more important than knowledge" A.Einstein
