Re: [Buildroot] [autobuild.buildroot.net] Your daily results for 2022-05-20

Re: [Buildroot] [autobuild.buildroot.net] Your daily results for 2022-05-20

[Thomas Petazzoni via buildroot]

Hello Ramon,

On Sat, 21 May 2022 17:23:04 +0300
Ramon Fried <rfried.dev@gmail.com> wrote:

> > Build failures related to your packages:
> >
> >     arch     |             reason             |                                       url
> > -------------+--------------------------------+---------------------------------------------------------------------------------
> >     m68k     |          bitwise-0.43          | http://autobuild.buildroot.net/results/9e3b9a148bc84004fde4dfe9b389ca2b359a92e2
> >   riscv64    |          bitwise-0.43          | http://autobuild.buildroot.net/results/348afc2a0fde85c098a12b58e8e2dcc3780bbb6b
> >
> >
> > Thanks for your contribution to Buildroot!
> >
> > --
> > http://autobuild.buildroot.net  
> Hi thomas.
> as I mentioned in an earlier mail, it looks like it is probably in dependencies.
> I took riscv64 as an example and built the qemu
> qemu_riscv64_virt_defconfig + bitwise locally.
> It is built successfully.
> 
> How was the config file for building the above created ?

If you follow the links above, you end up in a directory that contains
the full .config file (config) as well as the corresponding defconfig
(named "defconfig"). These are the configurations that triggered the
build issue.

You can also reproduce a build issue by downloading the script
https://git.busybox.net/buildroot-test/plain/utils/br-reproduce-build
and run:

 ./br-reproduce-build 9e3b9a148bc84004fde4dfe9b389ca2b359a92e2

Where 9e3b9a148bc84004fde4dfe9b389ca2b359a92e2 is the identifier of the
failed build, as visible in the autobuild.buildroot.net URL.

Let me know if this helps!

Thomas Petazzoni
-- 
Thomas Petazzoni, co-owner and CEO, Bootlin
Embedded Linux and Kernel engineering and training
https://bootlin.com
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [autobuild.buildroot.net] Your daily results for 2022-05-20

[Ramon Fried]

On Sat, May 21, 2022 at 5:44 PM Thomas Petazzoni
<thomas.petazzoni@bootlin.com> wrote:
>
> Hello Ramon,
>
> On Sat, 21 May 2022 17:23:04 +0300
> Ramon Fried <rfried.dev@gmail.com> wrote:
>
> > > Build failures related to your packages:
> > >
> > >     arch     |             reason             |                                       url
> > > -------------+--------------------------------+---------------------------------------------------------------------------------
> > >     m68k     |          bitwise-0.43          | http://autobuild.buildroot.net/results/9e3b9a148bc84004fde4dfe9b389ca2b359a92e2
> > >   riscv64    |          bitwise-0.43          | http://autobuild.buildroot.net/results/348afc2a0fde85c098a12b58e8e2dcc3780bbb6b
> > >
> > >
> > > Thanks for your contribution to Buildroot!
> > >
> > > --
> > > http://autobuild.buildroot.net
> > Hi thomas.
> > as I mentioned in an earlier mail, it looks like it is probably in dependencies.
> > I took riscv64 as an example and built the qemu
> > qemu_riscv64_virt_defconfig + bitwise locally.
> > It is built successfully.
> >
> > How was the config file for building the above created ?
>
> If you follow the links above, you end up in a directory that contains
> the full .config file (config) as well as the corresponding defconfig
> (named "defconfig"). These are the configurations that triggered the
> build issue.
>
> You can also reproduce a build issue by downloading the script
> https://git.busybox.net/buildroot-test/plain/utils/br-reproduce-build
> and run:
>
>  ./br-reproduce-build 9e3b9a148bc84004fde4dfe9b389ca2b359a92e2
>
> Where 9e3b9a148bc84004fde4dfe9b389ca2b359a92e2 is the identifier of the
> failed build, as visible in the autobuild.buildroot.net URL.
>
> Let me know if this helps!
Yes.
I managed to reproduce.
How do I know if it's a regression, or a new configuration that was tested ?
How was the configuration generated ?
I'm asking because the autotools fails to detect libreadline.
libreadline is installed in the filesystem.
Something is different in libreadline in this setup. I can't point my
finger at the problem.
>
> Thomas Petazzoni
> --
> Thomas Petazzoni, co-owner and CEO, Bootlin
> Embedded Linux and Kernel engineering and training
> https://bootlin.com
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [autobuild.buildroot.net] Your daily results for 2022-05-20

[Thomas Petazzoni via buildroot]

Hello Ramon,

On Sat, 21 May 2022 19:47:24 +0300
Ramon Fried <rfried.dev@gmail.com> wrote:

> > Let me know if this helps!  
> Yes.
> I managed to reproduce.
> How do I know if it's a regression, or a new configuration that was tested ?
> How was the configuration generated ?

Configurations are randomly generated. So it can either be a
configuration that was never tested before, or some other package that
was updated, and which has an impact on bitwise.

If you look at:

  http://autobuild.buildroot.net/?reason=bitwise-0.43

You can see that the problem started appearing on May 1, 2022. It
affects only riscv64 and m68k/5208, which is extremely strange (which
only those architectures ?).

What can be tried is take the first of those build issues (from May 1),
reproduce it. And then checkout progressively older commits of
Buildroot (before May 1), and see when it started failing.

Best regards,

Thomas
-- 
Thomas Petazzoni, co-owner and CEO, Bootlin
Embedded Linux and Kernel engineering and training
https://bootlin.com
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [autobuild.buildroot.net] Your daily results for 2022-05-20

[James Hilliard]

On Mon, May 23, 2022 at 9:47 AM Thomas Petazzoni via buildroot
<buildroot@buildroot.org> wrote:
>
> Hello Ramon,
>
> On Sat, 21 May 2022 19:47:24 +0300
> Ramon Fried <rfried.dev@gmail.com> wrote:
>
> > > Let me know if this helps!
> > Yes.
> > I managed to reproduce.
> > How do I know if it's a regression, or a new configuration that was tested ?
> > How was the configuration generated ?
>
> Configurations are randomly generated. So it can either be a
> configuration that was never tested before, or some other package that
> was updated, and which has an impact on bitwise.
>
> If you look at:
>
>   http://autobuild.buildroot.net/?reason=bitwise-0.43
>
> You can see that the problem started appearing on May 1, 2022. It
> affects only riscv64 and m68k/5208, which is extremely strange (which
> only those architectures ?).

Well it likely started getting hit due to the full randomization I
recently started
running on my autobuilders, if you notice only my two autobuilders are hitting
it(which are the only ones running the non-default full randomization rather
than just package configuration randomization).

>
> What can be tried is take the first of those build issues (from May 1),
> reproduce it. And then checkout progressively older commits of
> Buildroot (before May 1), and see when it started failing.

If it's only my autobuilders it likely just wasn't getting good coverage before,
there were also a lot of other bugs that full randomization uncovered that
were masking other issues by causing super early build failures, so good
chance it's not actually a regression but an issue that just never showed up
before due to poor test coverage of certain toolchain configurations.

>
> Best regards,
>
> Thomas
> --
> Thomas Petazzoni, co-owner and CEO, Bootlin
> Embedded Linux and Kernel engineering and training
> https://bootlin.com
> _______________________________________________
> buildroot mailing list
> buildroot@buildroot.org
> https://lists.buildroot.org/mailman/listinfo/buildroot
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot