Abdelrazak Younes wrote: > Georg Baum wrote: >> - unicode transition: this is not completed, people do even add new >> instances of utf8/file system encoded strings > > What's missing?
Basically all std::string instances for internal use carrying non-ASCII stuff, ie many places where from_utf8/to_utf8, filesystemEncoding, local8bit etc are used. The only allowed place for utf8 or any other 8bit encoding is communication with the outside world (files, programs, clipboard etc). Then mathed is completely unicode ignorant (in terms of LaTeX output). >> - getStatus/dispatch machinery: The old design is partly destroyed, a new >> one is not in sight. > > I didn't know it was destroyed. I see nothing changed up to now... but I > might be wrong. IIRC it is not guaranteed anymore that no dispatch method is called without the corresponding getStatus() before. IIRC there is no single dispatch entry point anymore either, so if I need to call some dispatch() it is completely unclear which one, and it is also unclear in which dispatch() I need to add some new stuff. >>>> but you write "I don't think the bug >>>> will be solved soon", >>> Please re-read what I said: "we should do the same for *branch* as I >>> don't think the bug will be solved soon. >> >> I see no difference. > > *branch*, I said *branch*. Yes, I understood that you wanted to put the workaround in branch. Nevertheless the reason you gave does also tell that the bug will probably not be fixed soon, and I did not read _that part_ as being restricted to branch, because IMHO it would be strange not to backport a fix to branch if one is found. Georg
