>>>>> "Andre" == Andre Poenitz <[EMAIL PROTECTED]> writes:
Andre> I fiddled around with "font insets" for mathed lately, and by Andre> now I am fairly convinced that this is the Right Thing to do. Andre> It is much more "logical Markup", it is basically what LaTeX Andre> does, and it is what Docbook does (if that's what José said). I Andre> don't think it is what "conventional word processors" do, but Andre> to be honest, I don't care. The problem is probably to see what kind of reasonable user interface we can give to this. >> ANother candidate would be Paragraph::insertChar, which is called a >> lot during parsing (and maybe other places) for inserting at the >> end of paragraph. Obviously in this case there is no need to >> search/update the fonttable/insettable. Luckily enough for 1.2.0 >> release date, I will not have time to look at it :) Andre> This problem, e.g. is magically solved by simply doing font Andre> related stuff only if drawing is requested. Just push_back() Andre> the char to the container... Huh? You have to fill in the data structure to indicate where font changes are, isn't it? Drawing is not relevant here. Andre> For the 'speed of string comparison matter': Somewhere in one Andre> of my older projects I had some class wrapping a 'const char *' Andre> and a hash values of the string for exactly that purpose. I Andre> seem to remember a speedup of four or five or so. So if this is Andre> really crucial, I can dig this out (or re-invent the wheel Andre> which is probably better for the stomach contents than looking Andre> at six years old code...) Note that we are talking about lots of compares of rather small (less than 10 chars) strings. JMarc
