[OE-core] [PATCH 5/7] base.bbclass: Report only the licenses that are incompatible for a package

[Peter Kjellerstedt]

> -----Original Message-----
> From: Quentin Schulz <quentin.schulz at streamunlimited.com>
> Sent: den 9 december 2019 11:20
> To: Peter Kjellerstedt <peter.kjellerstedt at axis.com>
> Cc: openembedded-core at lists.openembedded.org
> Subject: Re: [OE-core] [PATCH 5/7] base.bbclass: Report only the licenses
> that are incompatible for a package
> 
> Hi Peter,
> 
> On Sat, Dec 07, 2019 at 06:38:13AM +0100, Peter Kjellerstedt wrote:
> > Instead of reporting ${LICENSE} when a package is identified as using
> > an incompatible license, report the license(s) that are actually
> > incompatible.
> >
> > Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt at axis.com>
> > ---
> [...]
> > diff --git a/meta/lib/oeqa/selftest/cases/incompatible_lic.py
> b/meta/lib/oeqa/selftest/cases/incompatible_lic.py
> > index ad878571b5..9d1e801117 100644
> > --- a/meta/lib/oeqa/selftest/cases/incompatible_lic.py
> > +++ b/meta/lib/oeqa/selftest/cases/incompatible_lic.py
> > @@ -4,7 +4,7 @@ from oeqa.utils.commands import bitbake
> >  class IncompatibleLicenseTests(OESelftestTestCase):
> >
> >      def lic_test(self, pn, pn_lic, lic):
> > -        error_msg = 'ERROR: Nothing PROVIDES \'%s\'\n%s was skipped: it
> has an incompatible license: %s' % (pn, pn, pn_lic)
> > +        error_msg = 'ERROR: Nothing PROVIDES \'%s\'\n%s was skipped: it
> has incompatible license(s): %s' % (pn, pn, pn_lic)
> >
> >          self.write_config("INCOMPATIBLE_LICENSE += \"%s\"" % (lic))
> >
> > @@ -27,12 +27,12 @@ class IncompatibleLicenseTests(OESelftestTestCase):
> >      # license cannot be built when INCOMPATIBLE_LICENSE contains this
> SPDX
> >      # license
> >      def test_incompatible_spdx_license_alias(self):
> > -        self.lic_test('incompatible-license-alias', 'GPLv3', 'GPL-3.0')
> > +        self.lic_test('incompatible-license-alias', 'GPL-3.0', 'GPL-
> 3.0')
> >
> >      # Verify that a package with an alias (from SPDXLICENSEMAP) to an
> SPDX
> >      # license cannot be built when INCOMPATIBLE_LICENSE contains this
> alias
> >      def test_incompatible_alias_spdx_license_alias(self):
> > -        self.lic_test('incompatible-license-alias', 'GPLv3', 'GPLv3')
> > +        self.lic_test('incompatible-license-alias', 'GPL-3.0', 'GPLv3')
> >
> 
> We currently have 4 tests:
>  - GPL-3.0 in LICENSE field in recipe and GPL-3.0 in INCOMPATIBLE_LICENSE
>  - GPL-3.0 in LICENSE field in recipe and GPLv3 in INCOMPATIBLE_LICENSE
>  - GPLv3 in LICENSE field in recipe and GPL-3.0 in INCOMPATIBLE_LICENSE
>  - GPLv3 in LICENSE field in recipe and GPLv3 in INCOMPATIBLE_LICENSE
> 
> They test that the alias system (SPXLICENSEMAP) still works, which is
> rather important since we never enforced the use of licenses from
> SRC_DISTRIBUTE_LICENSES in LICENSE in recipes and allowed the use of
> aliases (hence SPXLICENSEMAP).
> 
> If I'm not mistaken this change induces the following 4 tests:
>  - GPL-3.0 in LICENSE field in recipe and GPL-3.0 in INCOMPATIBLE_LICENSE
>  - GPL-3.0 in LICENSE field in recipe and GPLv3 in INCOMPATIBLE_LICENSE
>  - GPL-3.0 in LICENSE field in recipe and GPL-3.0 in INCOMPATIBLE_LICENSE
>  - GPL-3.0 in LICENSE field in recipe and GPLv3 in INCOMPATIBLE_LICENSE
> 
> which duplicates two tests.

Actually not. Please note that different recipes are used for the 
first two tests (incompatible-license.bb) and the last two tests 
(incompatible-license-alias.bb).

> The questions are: which error warranted this code change?

Before, the LICENSE string was reported as is, whereas after my changes, 
the canonical SPDX names are reported.

> Are SPDXLICENSEMAP aliases still working? Can it still be checked?

It is still working and checked.

> If the original tests are indeed not needed anymore, we need to explicit
> why (and we can actually just remove them).
> 
> I'm still advocating for new tests with wildcards :)

Ah, now I think I get your comment from the previous mail. Yes, I 
should probably add some tests that validate the wildcard support.

> Thanks,
> Quentin

//Peter