Abdelrazak Younes <[EMAIL PROTECTED]> writes: | Lars Gullik Bjønnes wrote: | > Abdelrazak Younes <[EMAIL PROTECTED]> writes: | > | > The dispatch function called is not the same dispatch function | > | > anymore. There might be a reason why we have decided to go through the | > | > top most dispatch and let that handle the traversement of the dispatch | > | > hierarchy. | > | | And what is that reason? I thought it was just a remnant of some | > old | > | code practice. | > This code, the dispatch hierarchy is not very old. | > And you have to talk to the creators of this "modern" dispatch thing | > to get the reason, if any. | | So, that's what I am doing right now, right? Or is there somebody else?
If it is me you are thinking of, then now. Andre and J-M was certainly part of that rework. | > | > If this is an ivalid reason or not I cannot say, but there | > | > you have it.... | > | | I think it is invalid yes. If the traversal could lead to a | > different | > | LyXText then I think we are in serious trouble. If there someone | > | knowing more than that about that stuff. | > It is not a different lyxtext that is the problem, but a different | > dispatch. With this change the dispatch hierarchy is entered not at | > the top anymore. | | I know, I wrote the patch remember? ;-) | | More seriously, why is it a problem? The dispatch hierarchy will | inevitably end-up there so why not take the shortcut? IMHO the code | looks a lot more cleaner this way. Yeah, but it also means that some LFUNS, if called from lyxtext must still use bv.owner()->dispatch(...), else you will not be able to reach all available LFUNS. (Hmm... or do we call back up to the top if we don't find the lfun?) -- Lgb