[Freeipa-devel] Linking tickets in the commit messages

[Alexander Bokovoy]

On Thu, 17 Sep 2015, Jan Cholasta wrote:
>On 17.9.2015 14:08, Alexander Bokovoy wrote:
>>On Thu, 17 Sep 2015, Martin Kosek wrote:
>>>On 09/17/2015 02:00 PM, Alexander Bokovoy wrote:
>>>>On Thu, 17 Sep 2015, Martin Kosek wrote:
>>>>>On 09/17/2015 01:47 PM, Tomas Babej wrote:
>>>>>>Hi fellow developers,
>>>>>>
>>>>>>more or less we tend to stick to the tradition of linking Trac tickets
>>>>>>to the commit messages of the patches we send to the list.
>>>>>>
>>>>>>However, every now and then, a patch lands on the list, which is
>>>>>>either
>>>>>>linked to a BZ or does not contain any link at all. Admittedly, I am
>>>>>>also guilty of this mishap. This poses certain problems, as we're
>>>>>>trying
>>>>>>to automate the bookkeeping and pushing-related processes with
>>>>>>ipatool [1].
>>>>>>
>>>>>>Nevertheless, this useful habit is not formally agreed upon by
>>>>>>developers nor documented in our wiki [2]. I'd suggest we add it
>>>>>>there,
>>>>>>if we come to such consensus.
>>>>>>
>>>>>>This would mean:
>>>>>>* Patches fixing an issue described only in BZ (rare issue) would need
>>>>>>to create a Trac ticket referencing the BZ
>>>>>
>>>>>+1
>>>>>
>>>>>>* Patches fixing an issue not tracked in Trac nor BZ would need to
>>>>>>file
>>>>>>a ticket in Trac and reference it
>>>>>
>>>>>I am not sure we are there yet. For typos and small fixes, I do not
>>>>>think we
>>>>>need to create a hard requirement for a Trac ticket. But for patches
>>>>>that you
>>>>>want to be considered for say backports to downstream releases, it
>>>>>is better to
>>>>>have the ticket with the right metadata and collection of the right
>>>>>hashes that
>>>>>the downstream release can digest.
>>>>Yes, please do a ticket per a changeset.
>>>
>>>Are you agreeing now to me, i.e. do not require tickets for trivial
>>>fixes or to
>>>Tomas' proposal - require ticket for *all* patches?
>>I think we should have tickets for reasonable pieces of work. The
>>problem is always in identifying what does 'reasonable' mean. A
>>single-line fix may be an important CVE fix or a key to an important
>>bugfix. Still, there should be a context to fit, thus a ticket.
>
>Speaking as IPA package maitainer in RHEL, I would like to have ticket 
>link in every commit in maintenance branches. If a commit goes to the 
>master branch only, I'm OK with it not having a ticket link. So that's 
>where I would draw the line - if a commit goes into a maintenance 
>branch, it is a reasonable piece of work.
Good suggestion, thanks. We actually have the same with Samba -- *any*
backport to released branches requires a new bug to be opened and
mentioned in the commit message.
-- 
/ Alexander Bokovoy