Messages in this thread | | | Date | Mon, 26 Mar 2012 14:51:46 -0700 | From | Andrew Morton <> | Subject | Re: [PATCH v4] x86, olpc: add debugfs interface for EC commands |
| |
On Mon, 26 Mar 2012 14:31:46 -0700 "H. Peter Anvin" <hpa@zytor.com> wrote:
> On 03/26/2012 02:29 PM, Andrew Morton wrote: > >> > >> v4: really fix sign-off tags > > > > s/fix/break/? "Originally-from" is not a recognised tag. If this code > > is based upon an earlier version from Paul then Signed-off-by: is > > correct. > > > > What's going on here? What are you trying to communicate? > > I have been using "Originally-by:" (not Originally-from:) to indicate > content originally written by someone else which has then been > dramatically redone. It seems to communicate, but probably should be > written into a spec somewhere. >
Well, if a person had any contribution at all, we should seek their Signed-off-by:. Otherwise they could say "hey, you admitted using my code but I did not authorise its use", or any other range of bad IANAL things?
If we want to provide additional details on the authorship trail then fine, that can be done in English in the changelog. And it can be done far more clearly than in an ad-hoc tag whose meaning is unclear.
It's pretty unusual to see a patch which was authored by person A to have "From: B". Usually this comes about because B made a mistake. I will always query the authorship on such patches. Sometimes it is deliberate, but there's no consistent pattern in the reasoning. For this reason it is always best to fully explain the authorship alteration in the changelog text.
If we want to add more tag types then OK, we can discuss that, clearly define them, raise a patch against Documentation/SubmittingPatches and start to use them. But let's not invent new ones without explaining to anyone else what they mean.
Generally if I see anything other than "Signed-off-by:", "Cc:", "Tested-by:" "Reported-by:", "Acked-by:" or "Reviewed-by:", I consider them to be undefined and will zap 'em, sometimes transferring the intent into the changelog body instead.
| |