Michael Gerz wrote: > Georg Baum schrieb: >> Of course it has to be different. The problem is that these outputs come >> from exactly the same file. It has >> >> \output_changes true >> >> but 1.4 does not output any dvipost stuff because there is no unaccepted >> change in the file. I don't know whether that is a bug in 1.4 or in 1.5 >> or in none of them. >> > So you are telling me that 1.4 checks whether there is any change in the > doc and if not, it does not include dvipost?
I don't know exactly what it does (my "because" was just a guess), since I don't know change tracking very well. What I observed can be seen in the diff. > That would be a big surprise to me! Could someone with a 1.4 on > his/her(*) machine please check this? This is confirmed by Bo and by me, I would guess that this is enough ;-) > Anyway, I think that user are more or less aware that output may be > different when CT output is activated. If there is some discrepancy > between 1.4 and 1.5 in CT output mode, I don't consider this as a bug. I do consider that a bug, because users may not even be aware that CT is on in their 1.4 documents. Did you know that it is on in all but one file in our own documentation (both 1.4 and 1.5)? $ grep 'output_changes true' lib/doc/*.lyx | wc -l 59 $ grep 'output_changes' lib/doc/*.lyx | wc -l 60 This is no surprise since lyx2lyx adds "\output_changes true" when you convert from 1.3 to 1.4, so this will affect _a lot_ of files. Georg
