>>>>> "michael" == michael gerz <[EMAIL PROTECTED]> writes:
michael> Hi folks, I would like to inform you about two issues that I michael> am going to address this evening. If you disagree, please michael> complain loudly. Otherwise, I will continue the CT cleanup... michael> 1. Change time This looks OK. Please encapsulate the policy telling which changes are equal so that we can modify it later. Another solution would be to have all the changes in a same session be the same, or all changes by a same author be the same. michael> 2. Cleverer reject-changes michael> Presently, if you delete an inset, all its content is deleted michael> recursively. This is unfortunate. Imagine that your colleague michael> makes changes in an existing inset. He sends it to you and michael> you come to the conclusion that the inset should go michael> completely. If you erase it, LyX will delete all text within michael> the inset, i.e., the CT'ed insertions of your colleague are michael> lost and there is no way to reanimate them later. I am going michael> to change this behaviour in a way that the inset's content michael> remains untouched if you delete the inset. Moreover, michael> reject-Changes will not reject changes in nested insets michael> unless the inset itself is set to UNCHANGED. What sounds like michael> extra work in fact leads to less code, because I can remove michael> several methods at the same time. But currently, one can delete an INSERTED change even if it is done by someone else, I think. We need a global solution. JMarc
