Re: [PATCH] checkkconfigsymbols.py: add option -i to ignore files

[Valentin Rothberg]

On Fri, May 1, 2015 at 10:49 PM, Paul Bolle <pebolle@xxxxxxxxxx> wrote:
> [Added Russell, because I, sort, of drop his name.]
>
> Valentin Rothberg schreef op vr 01-05-2015 om 22:13 [+0200]:
>> On Fri, May 1, 2015 at 9:31 PM, Paul Bolle <pebolle@xxxxxxxxxx> wrote:
>> > Valentin Rothberg schreef op wo 29-04-2015 om 16:58 [+0200]:
>> >> Sometimes a user might be interested to filter certain reports (e.g.,
>> >> the many defconfigs).
>> >
>> > Is this actually useful outside of filtering out defconfigs?
>>
>> It's a regex, so we can filter entire paths as well (e.g., -i
>> 'arch/.*' to ignore all issues in arch/).  Until now, I only used it
>> to get rid of all the defconfigs.
>
> So, perhaps we're better off by just skipping defconfigs?

I remember Greg writing that some people do actually clean up some
defconfigs, so I'd like to keep the door open.

When I do the daily ' $ checkkconfigsymbols.py --diff
yesterday..today-next ' I see regularly that patches that remove a
Kconfig option do not remove references in defconfigs.  Ideally (i.e.,
in my ideal world), people check their patch with checkkconfigsymbols
and see that some defconfigs are dirty and also fix that.

Kind regards,
 Valentin

>> As far as I know, it's really hard to manually configure certain
>> boards.  With defconfigs, only few people have to go through the fire.
>> Two years ago I tried to manually select a kernel configuration for my
>> Nexus 7 and failed desperately since some feature constraints are just
>> not visible/understandable from the menu.  Not that I am an ARM
>> developer, but there I understood the need to have a defconfig : )
>
> See https://lkml.org/lkml/2012/1/18/355 . Manually configuring from
> scratch is, I think, simply not doable. About the only advice I'd dare
> to give someone would be: somehow get a .config that works for your
> machine, however old that .config might be, and use it as your base.
> Probably by doing
>     yes "" | make oldconfig >/dev/null
>
> So I guess my question is: is a defconfig to be considered a ".config
> that works for your machine"? And, yes, I realize "works" is a very
> broad goal.
>
>
> Paul Bolle
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/