Hello, On Wed 19 Aug 2020 at 11:16AM +01, Ian Jackson wrote:
> I think Sean has been under the impression that the meaning of the > flags that follow --server can be found by reading the manual. > Certainly I was under that impression. > >> Now, it's interesting to note that the 'u' here does not reflect the >> client's '-u' option. > > This is the key thing I was missing. Er, yes, me too. >> I don't know how the inclusion of "uid/gid 0 in the id map" can break >> things, but maybe I'm overlooking something. However, if we indeed >> agree that things can break here, then it seems to me that a new bug >> should be filed against rsync, IMO. > > Sean was probably thinking -u here meant "skip files that are newer on > the receiver". That's what I was thinking. > > I think we can have two general approaches to these undocumented > command line options: > > (1) UTSL to find out what each flag means, and decide if we like it. > I certainly didn't do that right at the beginning. If we do this we > should really review all the existing ones. > > (2) Trust rsync upstream not to get this wrong, and assume that if > the rsync client contrives to pass these options as part of --server, > that they aren't dangerous. > > I'm in favour of (2), which would imply immediately applying Sergio's > patch. Sean, what do you think ? Agreed. -- Sean Whitton
signature.asc
Description: PGP signature
