[PATCH 2 of 2 V3] patchbomb: includes a unique series ID in email header

[Pierre-Yves David]

On 05/09/2014 07:54 AM, Augie Fackler wrote:
> On Wed, May 07, 2014 at 10:52:11AM -0700, pierre-yves.david at ens-lyon.org wrote:
>> # HG changeset patch
>> # User Pierre-Yves David <pierre-yves.david at fb.com>
>> # Date 1399355480 25200
>> #      Mon May 05 22:51:20 2014 -0700
>> # Node ID 4523e13d2ce15bcb1b95ccf2ca2d0caa9db150ef
>> # Parent  ac17b5ae7f23ecc7c23742d34941b59e2c1a4e21
>> patchbomb: includes a unique series ID in email header
>>
>> Now that we have patch index and series size information, having a unique series
>> identifier will helps tool to glue all email back together without any
>> additional logic.
>
> Can I persuade you to do something other than the message-id of the
> parent message? Maybe a UUID, or the sha1 of the tip of the series
> plus a timestamp?

Ho yes, the current id have been quickly picked to make the topic moving 
forward.

> Somehow using the replied-to message id feels messy here - like if the
> same message is replied to more than once it'll comingle the
> identifiers.

No it will not. The ID is put in special header by patchbomb, so human 
reply to this email will not includes the header. And the value of this 
header is always the first patch in the series so a resend of a second 
patchbomb will use a new series ID even if their reply to the initial thread

The question is: is this identifier a good enough one?

I'm not super fan of it and we could use a bit more deterministic 
content. My currently best option would be to use a hash of: all 
revision hash + --flag content.

However I'm not sure the "better" ID is worth the extra complexity.

-- 
Pierre-Yves