On Nov 21, 2009, at 12:33 AM, Greg Stein wrote: > I don't recall ANY discussion of rewriting history. > > I'm somewhat peeved.
Sorry 'bout that; I didn't interpret this as rewriting history. I'm happy to change these edits back if we feel that they aren't appropriate. As far as whether or not the old svn:author properties should be updated to reflect changed ASF usernames, I believe they should, for the following reasons: * The contribulyzer wrongly interprets the old usernames, since they are no longer in COMMITTERS. (Arguably, the contribulyzer could be fixed, but it'd be a hack.) * We've rewritten the revision numbers as part of the migration to the ASF repos. Updating past usernames to reflect new accounts logically follows. * Usernames are just pointers back to people. Since psamuelson doesn't exist in the new repository, it's effectively a dangling pointer. These dangling references should be fixed, as a step toward *maintaining* history, not rewriting it. * As Lieven points out in another thread, since we don't have it recorded anywhere, it won't be long before the link of psamuelson<->peters is lost to newcomers to the project. (Even worse for non-obvious translations like dionisos<->ehu.) * In the future, these dangling usernames could be assigned to somebody *else* at the ASF, leading to even more confusion. Given the above, I think it reasonable to update the svn:author property on the imported revisions to reflect the new ASF usernames for our committers. [ Sidebar: I'm just a little concerned that discussions such as this are happening on d...@s.a.o, when we haven't yet switched off the old list. I don't want to cross-post, since there may be many people subscribed to both, but I don't know how many. We've now got two dev lists, and there is likely to be confusion about which one decisions are made on. Let's turn off d...@s.t.o RSN. ] -Hyrum
