Re: [OE-core] [PATCH 11/28] ifupdown: update 0.8.22 -> 0.8.35

[Randy MacLeod]

On 1/9/20 9:36 AM, Alexander Kanavin wrote:
On Thu, 9 Jan 2020 at 14:55, Richard Purdie 
> wrote:



That may be but I think this is an oversight, not an intention and we
should support busybox-free images...


Right, I'll fix the build, but the implementation of runtime testing is 
up to people who need this kind of feature.


Alex



We use a sysvinit image with ifupdown and without busybox so
we'll write a test. I've created a defect to track the work:
   https://bugzilla.yoctoproject.org/show_bug.cgi?id=13736

--
# Randy MacLeod
# Wind River Linux
--
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [OE-core] [PATCH 11/28] ifupdown: update 0.8.22 -> 0.8.35

[Alexander Kanavin]

On Fri, 10 Jan 2020 at 16:00, Adrian Bunk  wrote:

> Is musl runtime tested at all?
>
> ifupdown is one of these recipes where a musl build error was "fixed"
> with a patch that makes something do the wrong thing at runtime when
> using musl.
>

musl runtime is tested, just not with ifupdown, it's replaced by busybox in
all of the testing scenarios.

Alex
-- 
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [OE-core] [PATCH 11/28] ifupdown: update 0.8.22 -> 0.8.35

[Adrian Bunk]

On Thu, Jan 09, 2020 at 05:12:58PM +0100, Alexander Kanavin wrote:
> On Thu, 9 Jan 2020 at 16:55, Tom Rini  wrote:
> 
> > I have to admit I'm guessing at what level of automated run-time testing
> > we normally have today.  We sanity test a sysvinit+busybox image with
> > networking, yes?  If so, we also do that on systemd and I assume with
> > what I put for the systemd+busybox-free we also do there, yes?
> >
> 
> I believe we test sysvinit+busybox, and systemd+busybox at runtime via poky
> and poky-alt distro configs.
> 
> We also test systemd+busybox-free core-base-utils (without ifupdown), but
> only as a build (no runtime tests). See
> lib/oeqa/selftest/cases/imagefeatures.py.

Is musl runtime tested at all?

ifupdown is one of these recipes where a musl build error was "fixed" 
with a patch that makes something do the wrong thing at runtime when
using musl.

> Alex

cu
Adrian
-- 
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [OE-core] [PATCH 11/28] ifupdown: update 0.8.22 -> 0.8.35

[Alexander Kanavin]

On Thu, 9 Jan 2020 at 16:55, Tom Rini  wrote:

> I have to admit I'm guessing at what level of automated run-time testing
> we normally have today.  We sanity test a sysvinit+busybox image with
> networking, yes?  If so, we also do that on systemd and I assume with
> what I put for the systemd+busybox-free we also do there, yes?
>

I believe we test sysvinit+busybox, and systemd+busybox at runtime via poky
and poky-alt distro configs.

We also test systemd+busybox-free core-base-utils (without ifupdown), but
only as a build (no runtime tests). See
lib/oeqa/selftest/cases/imagefeatures.py.

Alex
-- 
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [OE-core] [PATCH 11/28] ifupdown: update 0.8.22 -> 0.8.35

[Tom Rini]

On Thu, Jan 09, 2020 at 04:33:58PM +0100, Alexander Kanavin wrote:
> On Thu, 9 Jan 2020 at 16:01, Tom Rini  wrote:
> 
> > On Thu, Jan 09, 2020 at 03:47:24PM +0100, Alexander Kanavin wrote:
> > > On Thu, 9 Jan 2020 at 15:44, Tom Rini  wrote:
> > >
> > > > The problem here would have been caught with systemd-free and
> > > > busybox-free being a check in lib/oeqa/selftest/cases/imagefeatures.py
> > > > yes?
> > >
> > > Yes, and you'd also need to do -c testimage on the image to make sure it
> > > boots and has the network up and working.
> >
> > Yes, but I'm thinking in terms of what to change / add so that it would
> > be part of the standard suite.  busybox-free but systemd is today, yes?
> 
> I can't parse your question :)

I have to admit I'm guessing at what level of automated run-time testing
we normally have today.  We sanity test a sysvinit+busybox image with
networking, yes?  If so, we also do that on systemd and I assume with
what I put for the systemd+busybox-free we also do there, yes?

-- 
Tom


signature.asc
Description: PGP signature
-- 
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [OE-core] [PATCH 11/28] ifupdown: update 0.8.22 -> 0.8.35

[Alexander Kanavin]

On Thu, 9 Jan 2020 at 16:01, Tom Rini  wrote:

> On Thu, Jan 09, 2020 at 03:47:24PM +0100, Alexander Kanavin wrote:
> > On Thu, 9 Jan 2020 at 15:44, Tom Rini  wrote:
> >
> > > The problem here would have been caught with systemd-free and
> > > busybox-free being a check in lib/oeqa/selftest/cases/imagefeatures.py
> > > yes?
> >
> > Yes, and you'd also need to do -c testimage on the image to make sure it
> > boots and has the network up and working.
>
> Yes, but I'm thinking in terms of what to change / add so that it would
> be part of the standard suite.  busybox-free but systemd is today, yes?
>

I can't parse your question :)

Alex
-- 
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [OE-core] [PATCH 11/28] ifupdown: update 0.8.22 -> 0.8.35

[Tom Rini]

On Thu, Jan 09, 2020 at 03:47:24PM +0100, Alexander Kanavin wrote:
> On Thu, 9 Jan 2020 at 15:44, Tom Rini  wrote:
> 
> > The problem here would have been caught with systemd-free and
> > busybox-free being a check in lib/oeqa/selftest/cases/imagefeatures.py
> > yes?
> 
> Yes, and you'd also need to do -c testimage on the image to make sure it
> boots and has the network up and working.

Yes, but I'm thinking in terms of what to change / add so that it would
be part of the standard suite.  busybox-free but systemd is today, yes?

-- 
Tom


signature.asc
Description: PGP signature
-- 
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [OE-core] [PATCH 11/28] ifupdown: update 0.8.22 -> 0.8.35

[Alexander Kanavin]

On Thu, 9 Jan 2020 at 15:44, Tom Rini  wrote:

> The problem here would have been caught with systemd-free and
> busybox-free being a check in lib/oeqa/selftest/cases/imagefeatures.py
> yes?
>

Yes, and you'd also need to do -c testimage on the image to make sure it
boots and has the network up and working.

Alex
-- 
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [OE-core] [PATCH 11/28] ifupdown: update 0.8.22 -> 0.8.35

[Tom Rini]

On Thu, Jan 09, 2020 at 03:36:08PM +0100, Alexander Kanavin wrote:
> On Thu, 9 Jan 2020 at 14:55, Richard Purdie <
> richard.pur...@linuxfoundation.org> wrote:
> 
> >
> > That may be but I think this is an oversight, not an intention and we
> > should support busybox-free images...
> >
> 
> Right, I'll fix the build, but the implementation of runtime testing is up
> to people who need this kind of feature.

The problem here would have been caught with systemd-free and
busybox-free being a check in lib/oeqa/selftest/cases/imagefeatures.py
yes?

-- 
Tom


signature.asc
Description: PGP signature
-- 
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [OE-core] [PATCH 11/28] ifupdown: update 0.8.22 -> 0.8.35

[Alexander Kanavin]

On Thu, 9 Jan 2020 at 14:55, Richard Purdie <
richard.pur...@linuxfoundation.org> wrote:

>
> That may be but I think this is an oversight, not an intention and we
> should support busybox-free images...
>

Right, I'll fix the build, but the implementation of runtime testing is up
to people who need this kind of feature.

Alex
-- 
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [OE-core] [PATCH 11/28] ifupdown: update 0.8.22 -> 0.8.35

[Richard Purdie]

On Thu, 2020-01-09 at 14:54 +0100, Alexander Kanavin wrote:
> On Thu, 9 Jan 2020 at 14:47, Alexander Kanavin <
> alex.kana...@gmail.com> wrote:
> > I had a closer look at how and where ifupdown recipe is used, and
> > from what I see it isn't:
> > every configuration is relying on busybox to provide the same
> > utilities. That's why the failure is seen only in world builds :)
> > 
> > Should we drop the ifupdown recipe?
> > 
> 
> A followup: there is a busybox-free set put together in meta/recipes-
> extended/packagegroups/packagegroup-core-base-utils.bb
> which includes ifupdown only if systemd is not in use.
> 
> However that packagegroup is not used in any images, and the only
> test for it is in lib/oeqa/selftest/cases/imagefeatures.py
> which does enable systemd, so that doesn't pull in ifupdown either.
> 
> This means ifupdown isn't runtime-tested now, and probably hasn't
> been tested for a while.

That may be but I think this is an oversight, not an intention and we
should support busybox-free images...

Cheers,

Richard

-- 
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [OE-core] [PATCH 11/28] ifupdown: update 0.8.22 -> 0.8.35

[Alexander Kanavin]

On Thu, 9 Jan 2020 at 14:47, Alexander Kanavin 
wrote:

>
> I had a closer look at how and where ifupdown recipe is used, and from
> what I see it isn't:
> every configuration is relying on busybox to provide the same utilities.
> That's why the failure is seen only in world builds :)
>
> Should we drop the ifupdown recipe?
>

A followup: there is a busybox-free set put together in
meta/recipes-extended/packagegroups/packagegroup-core-base-utils.bb
which includes ifupdown only if systemd is not in use.

However that packagegroup is not used in any images, and the only test for
it is in lib/oeqa/selftest/cases/imagefeatures.py
which does enable systemd, so that doesn't pull in ifupdown either.

This means ifupdown isn't runtime-tested now, and probably hasn't been
tested for a while.

Alex
-- 
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [OE-core] [PATCH 11/28] ifupdown: update 0.8.22 -> 0.8.35

[Alexander Kanavin]

On Wed, 8 Jan 2020 at 22:23, Richard Purdie <
richard.pur...@linuxfoundation.org> wrote:

> On Wed, 2020-01-08 at 14:27 +0100, Alexander Kanavin wrote:
> > Remove inet-6-.defn-fix-inverted-checks-for-loopback.patch as
> > it is difficult to rebase and not clear if still necessary.
> >
> > Signed-off-by: Alexander Kanavin 
> > ---
> >  ...efn-fix-inverted-checks-for-loopback.patch | 403 
> > --
> >  ...{ifupdown_0.8.22.bb => ifupdown_0.8.35.bb} |  11 +-
> >  2 files changed, 5 insertions(+), 409 deletions(-)
> >  delete mode 100644 meta/recipes-core/ifupdown/files/inet-6-.defn-
> > fix-inverted-checks-for-loopback.patch
> >  rename meta/recipes-core/ifupdown/{ifupdown_0.8.22.bb =>
> > ifupdown_0.8.35.bb} (83%)
>
>
> http://errors.yoctoproject.org/Errors/Details/305656/
>

I had a closer look at how and where ifupdown recipe is used, and from what
I see it isn't:
every configuration is relying on busybox to provide the same utilities.
That's why the failure is seen only in world builds :)

Should we drop the ifupdown recipe?

Alex
-- 
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [OE-core] [PATCH 11/28] ifupdown: update 0.8.22 -> 0.8.35

[Richard Purdie]

On Wed, 2020-01-08 at 14:27 +0100, Alexander Kanavin wrote:
> Remove inet-6-.defn-fix-inverted-checks-for-loopback.patch as
> it is difficult to rebase and not clear if still necessary.
> 
> Signed-off-by: Alexander Kanavin 
> ---
>  ...efn-fix-inverted-checks-for-loopback.patch | 403 
> --
>  ...{ifupdown_0.8.22.bb => ifupdown_0.8.35.bb} |  11 +-
>  2 files changed, 5 insertions(+), 409 deletions(-)
>  delete mode 100644 meta/recipes-core/ifupdown/files/inet-6-.defn-
> fix-inverted-checks-for-loopback.patch
>  rename meta/recipes-core/ifupdown/{ifupdown_0.8.22.bb =>
> ifupdown_0.8.35.bb} (83%)


http://errors.yoctoproject.org/Errors/Details/305656/

Cheers,

Richard

-- 
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core