Re: [Xen-devel] [PATCH 2/2] MAINTAINERS: use https for git trees

[Wei Liu <wei.liu2@xxxxxxxxxx>]

On Tue, Jul 10, 2018 at 11:36:33AM +0100, George Dunlap wrote:
> On 07/10/2018 11:30 AM, Wei Liu wrote:
> > On Tue, Jul 10, 2018 at 11:28:34AM +0100, George Dunlap wrote:
> >> On 07/10/2018 11:23 AM, Ian Jackson wrote:
> >>> Wei Liu writes ("Re: [PATCH 2/2] MAINTAINERS: use https for git trees"):
> >>>> On Tue, Jul 10, 2018 at 02:36:49AM -0600, Jan Beulich wrote:
> >>>>> On 10.07.18 at 10:15, <wei.liu2@xxxxxxxxxx> wrote:
> >>>>>> Signed-off-by: Wei Liu <wei.liu2@xxxxxxxxxx>
> >>>>>
> >>>>> What's wrong with git:// ? I think the commit message should be non-
> >>>>> empty here.
> >>>>
> >>>> git: is not encrypted, while https: is. At this time of age, it is
> >>>> better to use encryption as much as possible.
> >>>
> >>> I agree with this change, so
> >>>
> >>> Acked-by: Ian Jackson <ian.jackson@xxxxxxxxxxxxx>
> >>>
> >>>
> >>> Let me expand on Wei's reasons:
> >>>
> >>> The git protocol is not just unencrypted, but also unauthenticated.
> >>> In theory it is possible to verify the signed tags for actual
> >>> releases, but that is a cumbersome process which I very much doubt
> >>> anyone really does.
> >>>
> >>> As for the various branch tips, there is currently no way (unless you
> >>> have a shell account on xenbits) to get any kind of authenticated
> >>> value.
> >>>
> >>> Conversely, if you use an https url, you get some cryptographic
> >>> authentication of what you are cloning.  The crypto there is far from
> >>> perfect but it is massively better than nothing.
> >>
> >> I agree with this logic, but it should have been in the commit message.
> > 
> > Alright. I took it for granted that everyone would think the more
> > encryption the better.
> > 
> > I will put what Ian wrote into the commit message.
> 
> Well in general, the more things are encrypted, the less conspicuous
> encrypted traffic looks.  But on the other hand, there may be other
> costs with switching from git to https -- more server computation time,
> longer download time, &c.  If it were just a general "make more
> encrypted traffic to make encryption of actual secrets more safe", I
> don't think it would be worth degrading performance / increasing server
> compute time.  But for an extra level of authentication, I think it's
> worth it.
> 
> And in any case, I think it's almost always worth at least a brief line
> for the archaeologists.  Imagine 10 years down the road someone wants to
> know why it changed -- was it because we shut down the git servers?  Was
> it because https was measured as being faster?  Was it to get around
> firewalls?  Or was is just to improve authentication?  It may matter.

I don't fully agree what you said above but I'm not going to argue
because I've got what I wanted. :-)

Wei.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel