[LLVMdev] Phabricator and private reviews

David Blaikie dblaikie at gmail.com
Wed Jun 25 14:16:28 PDT 2014


On Wed, Jun 25, 2014 at 1:43 PM, Alp Toker <alp at nuanti.com> wrote:
>
> On 25/06/2014 23:35, Reid Kleckner wrote:
>
>> On Wed, Jun 25, 2014 at 11:57 AM, Nick Lewycky <nicholas at mxc.ca
>> <mailto:nicholas at mxc.ca>> wrote:
>>
>>     PS. I caused a few of these recently, my email in phab was set to
>>     an account that isn't subscribed to llvm-commits, so my reviews
>>     were held for moderation. Fixed on my end.
>>
>>
>> I think this is probably a fairly common cause of lost review mail.  I
>> updated the Phab documentation in r211731 to try to clarify this.  Feel free
>> to improve on that if I missed something.
>>
>> Alp, does r211731 help address your concerns?
>
>
> Cool, that makes sense.
>
> Is there a container image of the llvm.org Phabricator setup available?

Not sure.

> I'd like to have a shot at enforcing this or at least present a hint in the
> UI of the web app but I couldn't figure out where the container is hosted.

Just to be clear, the user experience here isn't any worse because of
Phab - a user should get the list mail telling them that their mail
(sent by Phab on their behalf) is awaiting moderation. The situation
in "[PATCH] cmake: NDEBUG needlessly defined in non-Release builds",
if Reid's theory is correct, would've been exactly the same even if
Phab were not involved at all. User emails <developer> and <mailing
list>, <developer> receives the mail while the list doesn't as the
mail awaits review. Developer replies and then we're exactly where we
were, every other developer sees a reply without the original mail and
is confused. Eventually the moderator accepts the email and it comes
through.

(I'm still a tad confused, though, as the ML archive seems to show the
original NDEBUG patch mail, but I don't think I've seen it come
through to my inbox - so I'm not sure if there are toher (Phab or
non-Phab) issues at work here)

- David



More information about the llvm-dev mailing list