Hello Armel, I was thinking of speed, 4 bytes per cell (2+2), not 3.
I wouldn't mind 1 chr + 3 attributes. That would be great! -Anthony Monday, June 19, 2006, 3:58:38 PM, you wrote: >>> is there any chance to have two style bytes for one character? One style >>> for backColor and one for other style properties. The purpose is to >>> easily distinguish embedded languages by backColor independently on >>> "fore" style. >> >> I think that it would be much simpler to just have 2 bytes for each, >> UTF-16 for the characters and 16 bits for the attribute - simply >> extending the attribute from 8 to 16 bits. > why UTF-16? what advantage? it has the same properties as UTF8 but is bigger > most of the time at least for Western languages. translation from UTF-8 to > UTF-16 is low cost and purely functional, keeping UTF-8 seems a better > idea... moreover _all_ the code assumes 'char*', going to UTF16 would > involve changes in _all_ the code... > extended the style stuff is feasible (I did it), but it has a big > disadvantage: the extra memory cost. 1 style byte makes a x2 on memory > (around x6 during insertions) but 4 styles bytes involves (x12 !!!)... so as > say the song: Hey think twice... > this solution is far from perfect :( > Armel > _______________________________________________ > Scintilla-interest mailing list > [email protected] > http://mailman.lyra.org/mailman/listinfo/scintilla-interest -- Best regards, Anthony mailto:[EMAIL PROTECTED] _______________________________________________ Scintilla-interest mailing list [email protected] http://mailman.lyra.org/mailman/listinfo/scintilla-interest
