Re: get_maintainers.pl subsystem output

["Duda, Sebastian" <sebastian.duda@fau.de>]

On 2019-07-19 08:50, Joe Perches wrote:
> On Fri, 2019-07-19 at 07:35 +0000, Duda, Sebastian wrote:
>> Hi Joe,
>> 
>> I'm conducting a large-scale patch analysis of the LKML with 1.8 
>> million
>> patch emails. I'm using the `get_maintainer.pl` script to know which
>> patch is related to which subsystem.
> 
> The MAINTAINERS file is updated frequently.
> 
> Are you also using the MAINTAINERS file used
> at the time each patch was submitted?

Yes, for each patch we use the MAINTAINERS file from the current (by the 
time the patch was submitted) release (candidate).

>> I ran into two issues while using the script:
>> 
>> 1. When I use the script the trivial way
>> 
>>      $ scripts/get_maintainer.pl --subsystem --status --separator ,
>> drivers/media/i2c/adv748x/
>>      Kieran Bingham <kieran.bingham@ideasonboard.com> 
>> (maintainer:ANALOG
>> DEVICES INC ADV748X DRIVER),Mauro Carvalho Chehab <mchehab@kernel.org>
>> (maintainer:MEDIA INPUT INFRASTRUCTURE
>> (V4L/DVB)),linux-media@vger.kernel.org (open list:ANALOG DEVICES INC
>> ADV748X DRIVER),linux-kernel@vger.kernel.org (open list)
>>      Maintained,Buried alive in reporters
>>      ANALOG DEVICES INC ADV748X DRIVER,MEDIA INPUT INFRASTRUCTURE
>> (V4L/DVB),THE REST
>> 
>> the output is hard to parse because the status `Maintained` is 
>> displayed
>> only once but related to two subsystems.
>> 
>> I'd prefer a more table like representation, like this:
>> 
>>      Kieran Bingham <kieran.bingham@ideasonboard.com> 
>> (maintainer:ANALOG
>> DEVICES INC ADV748X DRIVER),linux-media@vger.kernel.org (open
>> list:ANALOG DEVICES INC ADV748X DRIVER),ANALOG DEVICES INC ADV748X
>> DRIVER,Maintained
>>      Mauro Carvalho Chehab <mchehab@kernel.org> (maintainer:MEDIA 
>> INPUT
>> INFRASTRUCTURE (V4L/DVB)),MEDIA INPUT INFRASTRUCTURE
>> (V4L/DVB),Maintained
>>      linux-kernel@vger.kernel.org (open list),THE REST,Buried alive in
>> reporters
>> 
>> 
>> 2. I want to analyze multiple patches, currently I am calling the 
>> script
>> once per patch. When calling the script with multiple files the files
>> output is merged
>> 
>>      $ scripts/get_maintainer.pl --subsystem --status --separator ','
>> drivers/media/i2c/adv748x/ include/uapi/linux/wmi.h
>>      Kieran Bingham <kieran.bingham@ideasonboard.com> 
>> (maintainer:ANALOG
>> DEVICES INC ADV748X DRIVER),Mauro Carvalho Chehab <mchehab@kernel.org>
>> (maintainer:MEDIA INPUT INFRASTRUCTURE
>> (V4L/DVB)),linux-media@vger.kernel.org (open list:ANALOG DEVICES INC
>> ADV748X DRIVER),linux-kernel@vger.kernel.org (open
>> list),platform-driver-x86@vger.kernel.org (open list:ACPI WMI DRIVER)
>>      Maintained,Buried alive in reporters,Orphan
>>      ANALOG DEVICES INC ADV748X DRIVER,MEDIA INPUT INFRASTRUCTURE
>> (V4L/DVB),THE REST,ACPI WMI DRIVER
>> 
>> I'd like to run the script with all files but separated output, like
>> this:
>> 
>>      $ scripts/get_maintainer.pl --subsystem --status --separator ','
>> --separate-files drivers/media/i2c/adv748x/ include/uapi/linux/wmi.h
>>      Kieran Bingham <kieran.bingham@ideasonboard.com> 
>> (maintainer:ANALOG
>> DEVICES INC ADV748X DRIVER),Mauro Carvalho Chehab <mchehab@kernel.org>
>> (maintainer:MEDIA INPUT INFRASTRUCTURE
>> (V4L/DVB)),linux-media@vger.kernel.org (open list:ANALOG DEVICES INC
>> ADV748X DRIVER),linux-kernel@vger.kernel.org (open list)
>>      Maintained,Buried alive in reporters
>>      ANALOG DEVICES INC ADV748X DRIVER,MEDIA INPUT INFRASTRUCTURE
>> (V4L/DVB),THE REST
>> 
>>      platform-driver-x86@vger.kernel.org (open list:ACPI WMI
>> DRIVER),linux-kernel@vger.kernel.org (open list)
>>      Orphan,Buried alive in reporters
>>      ACPI WMI DRIVER,THE REST
>> 
>> 
>> My Questions are:
>> 1. How can I make get_maintainer's output to be more table-like?
> 
> I suggest adding --nogit --nogit-fallback --roles --norolestats

Unfortunately, this doesn't change the output:
     $ scripts/get_maintainer.pl --subsystem --status --separator , 
drivers/media/i2c/adv748x/
     Kieran Bingham <kieran.bingham@ideasonboard.com> (maintainer:ANALOG 
DEVICES INC ADV748X DRIVER),Mauro Carvalho Chehab <mchehab@kernel.org> 
(maintainer:MEDIA INPUT INFRASTRUCTURE 
(V4L/DVB)),linux-media@vger.kernel.org (open list:ANALOG DEVICES INC 
ADV748X DRIVER),linux-kernel@vger.kernel.org (open list)
     Maintained,Buried alive in reporters
     ANALOG DEVICES INC ADV748X DRIVER,MEDIA INPUT INFRASTRUCTURE 
(V4L/DVB),THE REST

     $ scripts/get_maintainer.pl --subsystem --status --separator , 
--nogit --nogit-fallback --roles --norolestats 
drivers/media/i2c/adv748x/
     Kieran Bingham <kieran.bingham@ideasonboard.com> (maintainer:ANALOG 
DEVICES INC ADV748X DRIVER),Mauro Carvalho Chehab <mchehab@kernel.org> 
(maintainer:MEDIA INPUT INFRASTRUCTURE 
(V4L/DVB)),linux-media@vger.kernel.org (open list:ANALOG DEVICES INC 
ADV748X DRIVER),linux-kernel@vger.kernel.org (open list)
     Maintained,Buried alive in reporters
     ANALOG DEVICES INC ADV748X DRIVER,MEDIA INPUT INFRASTRUCTURE 
(V4L/DVB),THE REST

>> 2. How can I make get_maintainer.pl to separate each file's output?
> 
> Run the script with multiple invocations. once for each file
> modified by the patch.

This is the way I'm doing it right now but this is very slow. I thought 
calling the script only once for many files could speed up the analysis.

Thank you
Sebastian