On Sun, Mar 1, 2020 at 10:39 PM Zac Medico <zmed...@gentoo.org> wrote:
>
> On 2/20/20 9:29 PM, Matt Turner wrote:
> > @@ -564,7 +577,22 @@ def findPackages(
> >
> > # Exclude if binpkg exists in the porttree and not --deep
> > if not destructive and port_dbapi.cpv_exists(cpv):
> > - continue
> > + if not options['changed-deps']:
> > + continue
> > +
> > + uselist = bin_dbapi.aux_get(cpv, ['USE'])[0].split()
> > + all_equal = True
> > +
> > + for k in ('RDEPEND', 'PDEPEND'):
> > + binpkg_deps = bin_dbapi.aux_get(cpv, [k])
> > + ebuild_deps = port_dbapi.aux_get(cpv, [k])
> > +
> > + if not _deps_equal(binpkg_deps, ebuild_deps, cpv.eapi,
> > uselist):
> > + all_equal = False
> > + break
> > +
> > + if all_equal:
> > + continue
>
> If all_equal is True, then none of the other filters have an opportunity
> to add this package to the dead_binpkgs set. That's not good is it?
There are four cases we skip including packages: 1) exclude list, 2)
time limit, 3) non-destructive and package still exists (and now
optionally runtime deps haven't changed), 4) destructive and package
is installed. Cases (3) and (4) are non-overlapping.
If none of those cases are true, then we add the package to the
dead_binpkgs set. The logic looks right to me.
Maybe I'm not understanding.
With your other suggestion in place, the code looks like this, which
is hopefully clearer.
# Exclude if binpkg exists in the porttree and not --deep
if not destructive and port_dbapi.cpv_exists(cpv):
if not options['changed-deps']:
continue
keys = ('RDEPEND', 'PDEPEND')
binpkg_deps = bin_dbapi.aux_get(cpv, keys)
ebuild_deps = port_dbapi.aux_get(cpv, keys)
uselist = bin_dbapi.aux_get(cpv, ['USE'])[0].split()
if _deps_equal(binpkg_deps, ebuild_deps, cpv.eapi, uselist):
continue
Unfortunately I don't have any packages with changed-deps at the
moment for testing :(
