Jean-Marc Lasgouttes wrote: > The original reason was a bad reason: that the session code had bad > interactions with your natbib code and loaded child documents twice, > therefore triggering bogus alerts. This has been or should be fixed > differently, so it is less relevant now.
This is actually fixed twice: By Bo who made sure that loaded child documents are ignored, and by my revised natbib code that does not loaad the child documents anymore. > Another idea I had for this case was to store in the buffer object the > file's last access time and check against that every time we do an > action that changes the buffer (like emacs does). This would solve > your particular problem and solve problems with files edited in two > places concurrently. I am not sure yet it is a good idea, so I did not > propose it. I think it is a good idea: Since both emacs and vim do something like that it must be good! The implementation is certainly more complicated. Do what you like, but please don't switch to the loaded buffer automatically. Georg
