Prolog: I had a very coarse look at this discussion and therefore I certainly see only a part of the picture. I am not going to take part in the discussion either. But sometimes a view from outside is useful, and therefore I wrote the comments below. Please ignore them if you think that they are not appropriate.
Jürgen Spitzmüller wrote: > I'm not sure about that. The discussions revealed that people do not trust > the approach, whether you accept that or not. And I think such big changes > are risky. I don't want to have another 1.5.0 where we had to make a rush, > hectic emergency release (1.5.1) because a very much unneeded and untested > switch of the compression feature resulted in serious dataloss and made > 1.5.0 an embarrassing, dangerous release. Such changes need to be tested > carefully, and it's simply too late for that. Yes. I do not understand why you do not simply decide that embedding will not be part of 1.6.0. Period. No discussion. "You" means in this case whoever nowadays has the power to do so, be it the release manager, the majority of the whole team, or whoever else. Then, those who want to discuss or even implement this feature _now_ may do so, but nobody is forced to take part, and the final decision how exactly it will look like will take place in the next development cycle. After a break it is hopefully possible to come to a compromise. As it looks to me this is quite impossible now. > (And being the one that basically cleaned up the listings implementation > and fixed dozens of critical bugs entailed by that feature -- the last > crash just some weeks ago -- I have to say that "more or less done" has to > be more than "works for me".) Indeed. The listings stuff contained at least one bad hack (usage of the deprecated part of the InsetCommandParams interface which made the code quite complicated) and was not ready for 1.5.0. Georg
