llvm-commits got removed from CC list when hitting reply-all to a Phab review

Tue Sep 15 18:24:32 PDT 2015

On Tue, Sep 15, 2015 at 5:13 PM, Duncan P. N. Exon Smith
<dexonsmith at apple.com> wrote:
>
>> On 2015-Sep-15, at 16:52, Manuel Klimek <klimek at google.com> wrote:
>>
>>
>>
>> On Tue, Sep 15, 2015 at 4:33 PM Duncan P. N. Exon Smith <dexonsmith at apple.com> wrote:
>> +llvm-commits
>>
>> > On 2015-Sep-15, at 15:49, Teresa Johnson <tejohnson at google.com> wrote:
>> >
>> > I just realized that none of the discussion starting with Duncan's
>> > review is going to llvm-commits, since llvm-commits is a subscriber in
>> > Phab, but the review by email doesn't actually update Phab or cc the
>> > subscribers (only the reviewers).
>>
>> And I just added two more emails :(.
>>
>> I'm not sure I understand how this happened.  Is it my fault somehow?  I just hit "reply-all" and I guess I didn't check the CC list.
>>
>> If there was an email that was not cc'ed to llvm-commits, llvm-commits was not put in as subsriber before creating the revision (the creator of the revision has to make sure, just like if you manually send an email)
>>
>
> This seems like a different case.  I've attached the original email as an attachment, but I'll quote the relevant part inline:
>
>> Begin forwarded message:
>>
>> From: Teresa Johnson <tejohnson at google.com>
>> Date: 2015 September 1 at 10:53:51 PDT
>> To: tejohnson at google.com, dexonsmith at apple.com, davidxl at google.com, mehdi.amini at apple.com
>> Cc: llvm-commits at lists.llvm.org
>> Subject: [PATCH] D12536: Function bitcode index in Value Symbol Table and lazy reading support
>> Reply-To: reviews+D12536+public+4f470f0312a65d2b at reviews.llvm.org
>> List-Id: reviews at reviews.llvm.org
>>
>> tejohnson created this revision.
>> tejohnson added reviewers: dexonsmith, davidxl, joker.eph.
>> tejohnson added a subscriber: llvm-commits.
>>
>
> Obviously I'm not a phab user, but it looks to me like this revision was created with llvm-commits as a subscriber.  Is there another way for llvm-commits to get dropped?
>

Yep, I was careful to add llvm-commits as subscriber when creating the
revision because I had been told that was necessary to get the right
email behavior. But it looks like it still didn't cc the subscribers
in all cases.

Looking back at the history of all the emails, it looks like the
trouble started when I pinged my patch. Emails shown below with some
of the relevant email header.

I first updated the patch via arc and that generated an expected email:

------------------
Date: Wed, 2 Sep 2015 17:02:57 +0000
To: tejohnson at google.com, dexonsmith at apple.com, davidxl at google.com,
mehdi.amini at apple.com
From: Teresa Johnson <tejohnson at google.com>
Cc: llvm-commits at lists.llvm.org
Reply-to: reviews+D12536+public+4f470f0312a65d2b at reviews.llvm.org
Subject: Re: [PATCH] D12536: Function bitcode index in Value Symbol
Table and lazy reading support

tejohnson updated this revision to Diff 33820.
tejohnson added a comment.

Update per Mehdi's review comments.
<snip>
------------------

I then did a reply-all to ping it, which for some reason only sends
email to reviews+D12536+public+4f470f0312a65d2b at reviews.llvm.org and
llvm-commits (just tried a draft reply-all on the above email to
confirm, not sure why reply-all is not including Duncan, Mehdi and
David who are all on the To: line):

------------------
Date: Tue, 8 Sep 2015 06:42:57 -0700
Delivered-To: tejohnson at google.com
Subject: Re: [PATCH] D12536: Function bitcode index in Value Symbol Table and
 lazy reading support
From: Teresa Johnson <tejohnson at google.com>
To: reviews+D12536+public+4f470f0312a65d2b at reviews.llvm.org
Cc: llvm-commits <llvm-commits at lists.llvm.org>

Ping.

Please let me know if you have any other comments or if it looks ready to go in.

Thanks,
Teresa

On Wed, Sep 2, 2015 at 10:02 AM, Teresa Johnson <tejohnson at google.com> wrote:
<snip>
------------------

This in turn resulted in an auto-generated email that included my
above Ping and sent to all the reviewers, but not llvm-commits (for
some reason also adding me as a subscriber):

------------------

Date: Tue, 8 Sep 2015 13:43:01 +0000
Return-Path: tejohnson at google.com
To: tejohnson at google.com, dexonsmith at apple.com, davidxl at google.com,
mehdi.amini at apple.com
From: Teresa Johnson <tejohnson at google.com>
Reply-to: reviews+D12536+public+4f470f0312a65d2b at reviews.llvm.org
Subject: Re: [PATCH] D12536: Function bitcode index in Value Symbol
Table and lazy reading support

tejohnson added a subscriber: tejohnson.
tejohnson added a comment.

Ping.

Please let me know if you have any other comments or if it looks ready to go in.

Thanks,
Teresa
<snip>
------------------

And the above email was the one Duncan responded to, starting our review thread:

------------------
Sender: dexonsmith at apple.com
Subject: Re: [PATCH] D12536: Function bitcode index in Value Symbol Table and
 lazy reading support
From: "Duncan P. N. Exon Smith" <dexonsmith at apple.com>
Date: Tue, 08 Sep 2015 11:40:34 -0700
Cc: tejohnson at google.com, davidxl at google.com,
 Mehdi Amini <mehdi.amini at apple.com>
To: reviews+D12536+public+4f470f0312a65d2b at reviews.llvm.org

> On 2015-Sep-08, at 06:43, Teresa Johnson <tejohnson at google.com> wrote:
<snip>
------------------

Manuel, any ideas on what went wrong here? It doesn't look like either
of us did anything particularly funky, but I don't understand why/when
Phab is sending emails to subscribers vs reviewers.

Thanks,
Teresa

-- 
Teresa Johnson | Software Engineer | tejohnson at google.com | 408-460-2413