Johan Corveleyn wrote on Wed, Sep 23, 2015 at 11:03:43 +0200: > Just one small addition on the fundamental part: I still think (like > we discussed during breakfast last Friday in Berlin :-)) there is no > problem in having / preserving / exposing null-text-changes to paths > in Subversion.
I tend to agree: in general, allowing the vacuous case tends to be useful. Johan's example of hanging a non-empty log message on a revision that lacks a non-empty text-delta is a perfect example of that. Another example is creating padding revisions. When we migrated to ASF, we had to incur a revision numbers offset of 840074. It would have been reasonable to create 26 padding revisions to make the offset a round number. Those revisions wouldn't have needed any content. If Subversion forbade revisions with no changed-paths, we would have had to create dummy changes — but, at that point, we would be working around a limitation of the tool. Cheers, Daniel
