llvm dev - Nov 2012 - [LLVMdev] [cfe-dev] !!! 3.2 Release branch patching and the Code Owners

----- Original Message -----
> From: "Joe Abbey" <joe.abbey at gmail.com>
> To: "Nadav Rotem" <nrotem at apple.com>
> Cc: llvmdev at cs.uiuc.edu
> Sent: Saturday, November 17, 2012 1:25:04 PM
> Subject: Re: [LLVMdev] [cfe-dev] !!! 3.2 Release branch patching and the
Code Owners
> 
> 
> On Nov 17, 2012, at 12:57 PM, Nadav Rotem <nrotem at apple.com>
wrote:
> 
> > I think that the code owner process is becoming complicated and I
> > am not sure if it serves Chris's original intent. I don't
think
> > that we need to change every file nor do we need an automatic tool
> > to find the owner. I think that a simple text file, or a section
> > in the docs is enough.
> 
> ^^ this
Pawel, please correct me if I'm wrong, but it sounds like your underlying
problem is that people are sending you merge requests directly, and you're
not sure from whom you need to make sure to get approvals. This being the case,
you should stop accepting such requests. Requests should be sent directly to the
code owners (on list). Only those code owners should communicate directly with
you (either to instruct you to merge in certain patches, or better yet, to merge
in approved changes directly). I think this matches the original intent of the
system: it partitions the workload among domain experts instead of forcing you
to deal explicitly with many of the requests.

 -Hal
> 
> Joe
> 
> 
> _______________________________________________
> LLVM Developers mailing list
> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
> 
-- 
Hal Finkel
Postdoctoral Appointee
Leadership Computing Facility
Argonne National Laboratory

> ----- Original Message -----
>> From: "Joe Abbey" <joe.abbey at gmail.com>
>> To: "Nadav Rotem" <nrotem at apple.com>
>> Cc: llvmdev at cs.uiuc.edu
>> Sent: Saturday, November 17, 2012 1:25:04 PM
>> Subject: Re: [LLVMdev] [cfe-dev] !!! 3.2 Release branch patching and
the	Code Owners
>>
>>
>> On Nov 17, 2012, at 12:57 PM, Nadav Rotem <nrotem at apple.com>
wrote:
>>
>>> I think that the code owner process is becoming complicated and I
>>> am not sure if it serves Chris's original intent. I don't
think
>>> that we need to change every file nor do we need an automatic tool
>>> to find the owner. I think that a simple text file, or a section
>>> in the docs is enough.
>>
>> ^^ this
> 
> Pawel, please correct me if I'm wrong, but it sounds like your
underlying problem is that people are sending you merge requests directly, and
you're not sure from whom you need to make sure to get approvals. This being
the case, you should stop accepting such requests. Requests should be sent
directly to the code owners (on list). Only those code owners should communicate
directly with you (either to instruct you to merge in certain patches, or better
yet, to merge in approved changes directly). I think this matches the original
intent of the system: it partitions the workload among domain experts instead of
forcing you to deal explicitly with many of the requests.
Hal, this is exactly what is happening.
The problem for me is compounded by the fact
that we have new code owners and I am spending
a lot of time verifying the ownership and then
I need to determine whether this is *approved*
or not.

As I have mentioned in the initial message I have
stopped processing the requests. I am just queuing
them up until situation gets resolved.

Workflow looks simple:

Developer -> patch -> Code owner -> *approved* -> Release Manager

but currently it breaks for me when this happens:

Developer -> patch -> Release Manager

Release Manager -> is this *approved* ? -> Code owners(?)

Code owner -> *approved* -> Release Manager


> 
>  -Hal
> 
>>
>> Joe
>>
>>
>> _______________________________________________
>> LLVM Developers mailing list
>> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
>> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
>>
>

----- Original Message -----
> From: "Pawel Wodnicki" <pawel at 32bitmicro.com>
> To: "Hal Finkel" <hfinkel at anl.gov>
> Cc: "Joe Abbey" <joe.abbey at gmail.com>, llvmdev at
cs.uiuc.edu, "Nadav Rotem" <nrotem at apple.com>
> Sent: Saturday, November 17, 2012 2:04:10 PM
> Subject: Re: [LLVMdev] [cfe-dev] !!! 3.2 Release branch patching and the
Code Owners
> 
> > ----- Original Message -----
> >> From: "Joe Abbey" <joe.abbey at gmail.com>
> >> To: "Nadav Rotem" <nrotem at apple.com>
> >> Cc: llvmdev at cs.uiuc.edu
> >> Sent: Saturday, November 17, 2012 1:25:04 PM
> >> Subject: Re: [LLVMdev] [cfe-dev] !!! 3.2 Release branch patching
> >> and the	Code Owners
> >>
> >>
> >> On Nov 17, 2012, at 12:57 PM, Nadav Rotem <nrotem at
apple.com>
> >> wrote:
> >>
> >>> I think that the code owner process is becoming complicated
and I
> >>> am not sure if it serves Chris's original intent. I
don't think
> >>> that we need to change every file nor do we need an automatic
> >>> tool
> >>> to find the owner. I think that a simple text file, or a
section
> >>> in the docs is enough.
> >>
> >> ^^ this
> > 
> > Pawel, please correct me if I'm wrong, but it sounds like your
> > underlying problem is that people are sending you merge requests
> > directly, and you're not sure from whom you need to make sure to
> > get approvals. This being the case, you should stop accepting such
> > requests. Requests should be sent directly to the code owners (on
> > list). Only those code owners should communicate directly with you
> > (either to instruct you to merge in certain patches, or better
> > yet, to merge in approved changes directly). I think this matches
> > the original intent of the system: it partitions the workload
> > among domain experts instead of forcing you to deal explicitly
> > with many of the requests.
> 
> Hal, this is exactly what is happening.
> The problem for me is compounded by the fact
> that we have new code owners and I am spending
> a lot of time verifying the ownership and then
> I need to determine whether this is *approved*
> or not.
> 
> As I have mentioned in the initial message I have
> stopped processing the requests. I am just queuing
> them up until situation gets resolved.
> 
> Workflow looks simple:
> 
> Developer -> patch -> Code owner -> *approved* -> Release
Manager
> 
> but currently it breaks for me when this happens:
> 
> Developer -> patch -> Release Manager
> 
> Release Manager -> is this *approved* ? -> Code owners(?)
> 
> Code owner -> *approved* -> Release Manager
I think that the solution to your problem is simple:

 1. Enforce the intended workflow: If a developer sends a patch to you directly,
then send a message back instructing them to send it to the code owner (on
list).
 2. Trust the code owners: Part of the responsibility of being in the
CODE_OWNERS file is knowing for what you're responsible and for what
you're not. It is not (or should not be) your job to verify a code
owner's approval. All you need to verify is that the person giving the
approval is a code owner (of something). I don't think that we have a
rogue-code-owner problem, do we?

We all appreciate the work that you're putting into this.

Thanks again,
Hal
> 
> 
> 
> > 
> >  -Hal
> > 
> >>
> >> Joe
> >>
> >>
> >> _______________________________________________
> >> LLVM Developers mailing list
> >> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
> >> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
> >>
> > 
> 
> 
-- 
Hal Finkel
Postdoctoral Appointee
Leadership Computing Facility
Argonne National Laboratory