Lars Gullik Bj?nnes wrote: > Jean-Marc Lasgouttes <lasgout...@lyx.org> writes: > > | Pavel Sanda <sa...@lyx.org> writes: > > > >> Abdelrazak Younes wrote: > >>> As for refering to a svn revision instead of a git branch, this is a > >>> different mental model indeed, but not not a loss of function IMHO. > >> > >> to me this depends on what kind of development model you use and given > >> the number of lyx developers and the way we proceed i think the centralized > >> way is the better one. > > > | For users that (try to) help us find bugs (and we need these people), > | saying "it did work at r1234" is easier that giving a hash (isn't this > | how a git state is defined? here I show my ignorance about it). > > I do not quite get that... who is it easer to say r1234 instead of > 23ae45?
if one wants to have unambigue commit pointer for future use 6-char hash prefix wont do. you also lose info about the order wrt to the commits around. > | To make things clearer, there is some merit to the increasingness of the > | svn revision numbers (in two different branches at a time). From what I > | understand about git, it is lost in the new world. > > well... otoh one change on the trunk will be named/numbered the same way > on branch... that is not the case with subversion. (even if you take > merge tracking into account) no, commits in branch are different due to status.16x changes. pavel
