Re: [meta-oe][PATCH V2 6/7] blacklist.bbclass: Move to meta-angstrom

[Koen Kooi <koen@dominion.thruhere.net>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Op 05-08-12 23:37, Chris Larson schreef:
> On Sun, Aug 5, 2012 at 4:41 AM, Koen Kooi <koen@dominion.thruhere.net>
> wrote:
>>> doent sound logical to me. Would it be acceptable if angstrom used
>>> this feature from OE-Core as well ? then it will remove the usecase 
>>> completely.
>> 
>> I keep waiting on the person that pushed it to oe-core to provide
>> patches for distro layers using the old functionality, see above.
> 
> This doesn't make much sense to me. It's impossible to know what all 
> layers use a given piece of functionality, particularly ones that aren't
> published in public places, so expecting every person changing something
> in oe-core will go and submit changes to every layer that exists is
> entirely unreasonable. If you maintain a layer, it's your responsibility
> to make sure your layer retains compatibility with upstream. It's not
> upstream's responsibility to make sure everyone using their functionality
> stays compatible.

In this specific instance the class was moved from meta-angstrom to meta-oe
as is and then moved to oe-core and changed to make it "less angstrom
specific".
I don't see how you can  argue that the person making that change was/is
unaware of the existence of the angstrom layer. Or of the SHR usage of that
class.

But yes, some people don't know that layers exist, there was a nice example
last week when I asked if someone checked *all* the layers out there on the
impact of the patch in question and the response was "Yes, I checked both
oe-core and meta-intel".

I agree with your generic point that changes in oe-core can not and should
not require updating of all layers by the submitter. On the other hand: if
you know you're moving functionality between layers while breaking it in the
process it is anti-social to not fix up the user(s) of that functionality
that you know of.

This is similar to asking a oe-core patch to be held back for a few days
because a number of layers have a matching bbappend that will need updating.
Especially if those layers have maintainers where the patch review latency
is measured in days instead of hours (e.g. me with non-denzil based stuff).
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)
Comment: GPGTools - http://gpgtools.org

iD8DBQFQH3W3MkyGM64RGpERAlbgAKCBvZTQD38H44mB+JzjQ2t92ttMzQCfVxeJ
nX2o+AfM7mrFEZnBT7YAV5s=
=v8kK
-----END PGP SIGNATURE-----