[PATCH 1/2] Use an explicit initial default patch state

Dirk Wallenstein halsmit at t-online.de
Mon Apr 30 22:01:17 EST 2012

On Mon, Apr 30, 2012 at 12:18:53PM +0200, Wolfgang Denk wrote:
> Dear Jeremy,
> In message <4F9E5D3D.5050504 at ozlabs.org> you wrote:
> > 
> > The patches Dirk sent don't do this; they allow the state to be set on 
> > the *original mail*, when the patch is first parsed.
> I understand thisnow.  Sorry, I was expecting too much :-(
> > The issue with allowing the state to be updated is that there isn't a 
> > method to authenticate the sender, and ensure that they're a maintainer 
> > of the project. It's possible to look at the From: field, but that's 
> > easily faked. In effect, I could send mail to 
> > incoming at patchwork.ozlabs.org, and set all of the buildroot patches to 
> > Rejected.
> I would be willing to accept this.  I don't expect much misuse here,
> and for me the benefit appears to be greater than the risk.
> > If anyone has useful ideas on how patchwork could do proper 
> > authentication, then I'd be happy to implement this.
> Sorry, but I don't know of an easy _and_ reliable way for email-based
> authentication.
> > It if suits, you could achieve the same thing with a git hook, which 
> > sets the patchwork state (using pwclient) when it's applied to your tree.
> I'm using this already, but this is not what I'm looking for.
> I want to be able to change the status even without git interaction;
> for example, when asking the submitter to rework his patch, I would
> like to add a "X-Patchwork-State: Changes Requested" header; when I
> see it falls into the bailiwick of a specific custodian, I would also
> set the appropriate "X-Patchwork-Delegate" header.
> As is, I always have to use a second tool to perform this action -
> which means additional efforts, which means patchwork is more of a
> pain than a tool that makes my work more efficient.

Actually, that might also be a nice thing for occasional contributors
without an account -- in particular revoking patches (although 
a patch sign language would probably be better in this case, like
[PATCH REVOKED] in the subject).
Anyway, Jeremy, I got the nice automated update notifications this
morning.  Maybe those could be considered sufficient for reassurance?


More information about the Patchwork mailing list