Re: [gentoo-portage-dev] [PATCH] depgraph: eval disjunctive build deps earlier (bug 639346)

2017-12-04 Thread Brian Dolbec 
On Sun,  3 Dec 2017 15:19:44 -0800
Zac Medico  wrote:

> Since built-time deps tend to be a superset of run-time deps, evaluate
> disjunctive build-time deps before run-time deps, so that choices for
> build-time deps influence choices for run-time deps.
> 
> Also, fix OnlydepsMinimalTestCase to specify ambiguous_merge_order,
> since the merge order is affected by the order of evaluation.
> 
> Bug: https://bugs.gentoo.org/639346
> ---
>  pym/_emerge/depgraph.py| 12 ++-
>  .../resolver/test_disjunctive_depend_order.py  | 87
> ++ .../tests/resolver/test_onlydeps_minimal.py
> |  5 +- 3 files changed, 98 insertions(+), 6 deletions(-)
>  create mode 100644
> pym/portage/tests/resolver/test_disjunctive_depend_order.py
> 
> diff --git a/pym/_emerge/depgraph.py b/pym/_emerge/depgraph.py
> index f54acdc26..6e5ca6508 100644
> --- a/pym/_emerge/depgraph.py
> +++ b/pym/_emerge/depgraph.py
> @@ -3234,7 +3234,15 @@ class depgraph(object):
>   if ignore_hdepend_deps:
>   edepend["HDEPEND"] = ""
>  
> + # Since build-time deps tend to be a superset of
> run-time deps, order
> + # dep processing such that build-time deps are
> popped from
> + # _dep_disjunctive_stack first, so that choices for
> build-time
> + # deps influence choices for run-time deps (bug
> 639346). deps = (
> + (myroot, edepend["RDEPEND"],
> + self._priority(runtime=True)),
> + (myroot, edepend["PDEPEND"],
> + self._priority(runtime_post=True)),
>   (depend_root, edepend["DEPEND"],
>   self._priority(buildtime=True,
>   optional=(pkg.built or
> ignore_depend_deps), @@ -3243,10 +3251,6 @@ class depgraph(object):
>   self._priority(buildtime=True,
>   optional=(pkg.built or
> ignore_hdepend_deps), ignored=ignore_hdepend_deps)),
> - (myroot, edepend["RDEPEND"],
> - self._priority(runtime=True)),
> - (myroot, edepend["PDEPEND"],
> - self._priority(runtime_post=True))
>   )
>  
>   debug = "--debug" in self._frozen_config.myopts
> diff --git
> a/pym/portage/tests/resolver/test_disjunctive_depend_order.py
> b/pym/portage/tests/resolver/test_disjunctive_depend_order.py new
> file mode 100644 index 0..88f6dac2d --- /dev/null
> +++ b/pym/portage/tests/resolver/test_disjunctive_depend_order.py
> @@ -0,0 +1,87 @@
> +# Copyright 2017 Gentoo Foundation
> +# Distributed under the terms of the GNU General Public License v2
> +
> +from portage.tests import TestCase
> +from portage.tests.resolver.ResolverPlayground import (
> + ResolverPlayground,
> + ResolverPlaygroundTestCase,
> +)
> +
> +class DisjunctiveDependOrderTestCase(TestCase):
> +
> + def testDisjunctiveDependOrderTestCase(self):
> + ebuilds = {
> + 'virtual/jre-1.8': {
> + 'EAPI': '6',
> + 'SLOT' : '1.8',
> + 'RDEPEND' : '||
> ( dev-java/oracle-jre-bin:1.8 virtual/jdk:1.8 )',
> + },
> + 'virtual/jdk-1.8': {
> + 'EAPI': '6',
> + 'SLOT' : '1.8',
> + 'RDEPEND' : '|| ( dev-java/icedtea:8
> dev-java/oracle-jdk-bin:1.8 )',
> + },
> + 'dev-java/icedtea-3.6': {
> + 'SLOT' : '8',
> + },
> + 'dev-java/oracle-jdk-bin-1.8': {
> + 'SLOT' : '1.8',
> + },
> + 'dev-java/oracle-jre-bin-1.8': {
> + 'SLOT' : '1.8',
> + },
> + 'dev-db/hsqldb-1.8'   : {
> + 'DEPEND' : 'virtual/jdk',
> + 'RDEPEND' : 'virtual/jre',
> + },
> + }
> +
> + binpkgs = {
> + 'dev-db/hsqldb-1.8'   : {
> + 'DEPEND' : 'virtual/jdk',
> + 'RDEPEND' : 'virtual/jre',
> + },
> + }
> +
> + test_cases = (
> + # Test bug 639346, where a redundant jre
> implementation
> + # was pulled in because DEPEND was evaluated
> after
> + # RDEPEND.
> + ResolverPlaygroundTestCase(
> + ['dev-db/hsqldb'],
> + success=True,
> + mergelist=[
> +

Re: [gentoo-portage-dev] [PATCH] xtermTitle: support st (simple terminal)

2017-12-04 Thread Zac Medico 
On 12/04/2017 12:16 AM, Christoph Böhmwalder wrote:
> On Mon, Dec 04, 2017 at 06:27:39AM +0100, Christoph Böhmwalder wrote:
>> Maybe we should be using the 'stterm' symlinks instead?
>>
>> --
>> Regards,
>> Christoph
> 
> Uh, actually that doesn't make any sense, since the $TERM variable is
> still set to just 'st-...'.  It was a little too early when I wrote that
> I suppose :)
> 
> So yeah, I think we might need to just roll with 'st-256color' anyways
> and just see how it goes

Okay, merged:

https://gitweb.gentoo.org/proj/portage.git/commit/?id=a45f1cad89eb96e11fd198565f2a6f8ce6922e38
-- 
Thanks,
Zac

Re: [gentoo-portage-dev] [PATCH] xtermTitle: support st (simple terminal)

2017-12-04 Thread Christoph Böhmwalder 
On Mon, Dec 04, 2017 at 06:27:39AM +0100, Christoph Böhmwalder wrote:
> Maybe we should be using the 'stterm' symlinks instead?
> 
> --
> Regards,
> Christoph

Uh, actually that doesn't make any sense, since the $TERM variable is
still set to just 'st-...'.  It was a little too early when I wrote that
I suppose :)

So yeah, I think we might need to just roll with 'st-256color' anyways
and just see how it goes

--
Regards,
Christoph