Announcing meta-musl OE/Yocto layer supporting musl C library

Announcing meta-musl OE/Yocto layer supporting musl C library

[Khem Raj]

Hello All

musl <http://www.musl-libc.org/> is a newer implementation licensed
under MIT licence
A layer to support musl to supply your system C library as an
alternative to uclibc and eglibc/glibc  is now available at

https://github.com/kraj/meta-musl

README describes way to get going with it.

Few things to note

* Supports gcc 4.9 in OE not 4.8
* Current musl recipe is 1.1.4 based
* You need latest master of OE-Core
(106305227003761c3fc562c21bb859a5256f2b36) or          newer and
bitbake
* Its only able to build core-image-minimal
* core-image-minimal boots on qemuarm, qemuppc, qemumips, qemux86, qemux86-64

I will send a request to add it to layerindex soon and then you should
be able to find it using
layerindex.opernembedded.org as well

If you try it out on your hardware and/or have feedback I am
interested. Fork it from github, file bugs
send patches ....

Happy hacking

-Khem

Re: Announcing meta-musl OE/Yocto layer supporting musl C library

[Diego Sueiro]

[-- Attachment #1: Type: text/plain, Size: 1587 bytes --]

Wow!!!!

This is a really great job. Nice work.

I'm gonna test it.


Regards,

--
*dS
Diego Sueiro

Administrador do Embarcados
www.embarcados.com.br
<http://www.embarcados.com.br/?utm_source=assinatura_diego&utm_medium=e-mail&utm_campaign=Assinatura%20Email%20Diego>

/*long live rock 'n roll*/


On Sat, Aug 16, 2014 at 9:50 PM, Khem Raj <raj.khem@gmail.com> wrote:

> Hello All
>
> musl <http://www.musl-libc.org/> is a newer implementation licensed
> under MIT licence
> A layer to support musl to supply your system C library as an
> alternative to uclibc and eglibc/glibc  is now available at
>
> https://github.com/kraj/meta-musl
>
> README describes way to get going with it.
>
> Few things to note
>
> * Supports gcc 4.9 in OE not 4.8
> * Current musl recipe is 1.1.4 based
> * You need latest master of OE-Core
> (106305227003761c3fc562c21bb859a5256f2b36) or          newer and
> bitbake
> * Its only able to build core-image-minimal
> * core-image-minimal boots on qemuarm, qemuppc, qemumips, qemux86,
> qemux86-64
>
> I will send a request to add it to layerindex soon and then you should
> be able to find it using
> layerindex.opernembedded.org as well
>
> If you try it out on your hardware and/or have feedback I am
> interested. Fork it from github, file bugs
> send patches ....
>
> Happy hacking
>
> -Khem
> --
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.openembedded.org/mailman/listinfo/openembedded-core
>

[-- Attachment #2: Type: text/html, Size: 2563 bytes --]

Re: Announcing meta-musl OE/Yocto layer supporting musl C library

[atulkumar singh]

[-- Attachment #1: Type: text/plain, Size: 1722 bytes --]

Hi Khem,

It's really something new to test.
And as it's mention that will support only for core-image-minimal, so when
can i expect it to support for meta-toolchain.
As of now i will test it for core-image-minimal but i need it for
meta-toolchain as well.
So please update me regarding the same and in the meanwhile will test and
update for core-image-minimal in case found any issue.
I am really excited to test the same.

Thanks and regards,
Atul
On 17 Aug 2014 06:22, "Khem Raj" <raj.khem@gmail.com> wrote:

> Hello All
>
> musl <http://www.musl-libc.org/> is a newer implementation licensed
> under MIT licence
> A layer to support musl to supply your system C library as an
> alternative to uclibc and eglibc/glibc  is now available at
>
> https://github.com/kraj/meta-musl
>
> README describes way to get going with it.
>
> Few things to note
>
> * Supports gcc 4.9 in OE not 4.8
> * Current musl recipe is 1.1.4 based
> * You need latest master of OE-Core
> (106305227003761c3fc562c21bb859a5256f2b36) or          newer and
> bitbake
> * Its only able to build core-image-minimal
> * core-image-minimal boots on qemuarm, qemuppc, qemumips, qemux86,
> qemux86-64
>
> I will send a request to add it to layerindex soon and then you should
> be able to find it using
> layerindex.opernembedded.org as well
>
> If you try it out on your hardware and/or have feedback I am
> interested. Fork it from github, file bugs
> send patches ....
>
> Happy hacking
>
> -Khem
> --
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.openembedded.org/mailman/listinfo/openembedded-core
>

[-- Attachment #2: Type: text/html, Size: 2425 bytes --]

Re: Announcing meta-musl OE/Yocto layer supporting musl C library

[Paul Barker]

[-- Attachment #1: Type: text/plain, Size: 2918 bytes --]

On Sat, Aug 16, 2014 at 05:50:38PM -0700, Khem Raj wrote:
> Hello All
> 
> musl <http://www.musl-libc.org/> is a newer implementation licensed
> under MIT licence
> A layer to support musl to supply your system C library as an
> alternative to uclibc and eglibc/glibc  is now available at
> 
> https://github.com/kraj/meta-musl
> 
> README describes way to get going with it.
> 
> Few things to note
> 
> * Supports gcc 4.9 in OE not 4.8
> * Current musl recipe is 1.1.4 based
> * You need latest master of OE-Core
> (106305227003761c3fc562c21bb859a5256f2b36) or          newer and
> bitbake
> * Its only able to build core-image-minimal
> * core-image-minimal boots on qemuarm, qemuppc, qemumips, qemux86, qemux86-64
> 
> I will send a request to add it to layerindex soon and then you should
> be able to find it using
> layerindex.opernembedded.org as well
> 
> If you try it out on your hardware and/or have feedback I am
> interested. Fork it from github, file bugs
> send patches ....
> 
> Happy hacking
> 

core-image-minimal works with my configuration on qemuarm. It also successfully
builds what will soon be released as opkg-0.2.3 and the recipe for toybox-0.4.9
which I've recently created.

I tried building core-image-full-cmdline as I think that would be good next step
from core-image-minimal. I'm seeing failures in:

  meta/recipes-devtools/e2fsprogs/e2fsprogs_1.42.9.bb, do_compile
  meta/recipes-extended/tcp-wrappers/tcp-wrappers_7.6.bb, do_compile
  meta/recipes-extended/libtirpc/libtirpc_0.2.4.bb, do_compile
  meta/recipes-extended/iptables/iptables_1.4.21.bb, do_compile
  meta/recipes-connectivity/nfs-utils/libnfsidmap_0.25.bb, do_compile
  meta/recipes-devtools/perl/perl_5.20.0.bb, do_compile
  meta/recipes-extended/sysklogd/sysklogd_1.5.bb, do_compile
  meta/recipes-extended/ethtool/ethtool_3.14.bb, do_compile
  meta/recipes-extended/psmisc/psmisc_22.21.bb, do_compile
  meta/recipes-extended/pax/pax_3.4.bb, do_compile
  meta/recipes-extended/procps/procps_3.2.8.bb, do_compile
  meta/recipes-extended/net-tools/net-tools_1.60-25.bb, do_compile
  meta/recipes-extended/findutils/findutils_4.4.2.bb, do_compile
  meta/recipes-extended/logrotate/logrotate_3.8.7.bb, do_compile
  meta/recipes-extended/cronie/cronie_1.4.11.bb, do_compile
  meta/recipes-devtools/elfutils/elfutils_0.148.bb, do_compile
  meta/recipes-extended/at/at_3.1.14.bb, do_compile
  meta/recipes-extended/iputils/iputils_s20121221.bb, do_compile
  meta/recipes-extended/gzip/gzip_1.6.bb, do_compile
  meta/recipes-extended/sudo/sudo_1.8.10p3.bb, do_compile
  meta/recipes-extended/mc/mc_4.8.12.bb, do_compile

I can share detailed logs if anyone wants them. I'll have a look and see if I
can find some low hanging fruit amongst these which are easy to fix. 

Thanks,

-- 
Paul Barker

Email: paul@paulbarker.me.uk
http://www.paulbarker.me.uk

[-- Attachment #2: Type: application/pgp-signature, Size: 484 bytes --]

Re: Announcing meta-musl OE/Yocto layer supporting musl C library

[Khem Raj]

On Mon, Aug 18, 2014 at 11:14 AM, Paul Barker <paul@paulbarker.me.uk> wrote:
> On Sat, Aug 16, 2014 at 05:50:38PM -0700, Khem Raj wrote:
>> Hello All
>>
>> musl <http://www.musl-libc.org/> is a newer implementation licensed
>> under MIT licence
>> A layer to support musl to supply your system C library as an
>> alternative to uclibc and eglibc/glibc  is now available at
>>
>> https://github.com/kraj/meta-musl
>>
>> README describes way to get going with it.
>>
>> Few things to note
>>
>> * Supports gcc 4.9 in OE not 4.8
>> * Current musl recipe is 1.1.4 based
>> * You need latest master of OE-Core
>> (106305227003761c3fc562c21bb859a5256f2b36) or          newer and
>> bitbake
>> * Its only able to build core-image-minimal
>> * core-image-minimal boots on qemuarm, qemuppc, qemumips, qemux86, qemux86-64
>>
>> I will send a request to add it to layerindex soon and then you should
>> be able to find it using
>> layerindex.opernembedded.org as well
>>
>> If you try it out on your hardware and/or have feedback I am
>> interested. Fork it from github, file bugs
>> send patches ....
>>
>> Happy hacking
>>
>
> core-image-minimal works with my configuration on qemuarm. It also successfully
> builds what will soon be released as opkg-0.2.3 and the recipe for toybox-0.4.9
> which I've recently created.
>
> I tried building core-image-full-cmdline as I think that would be good next step
> from core-image-minimal. I'm seeing failures in:
>
>   meta/recipes-devtools/e2fsprogs/e2fsprogs_1.42.9.bb, do_compile
>   meta/recipes-extended/tcp-wrappers/tcp-wrappers_7.6.bb, do_compile
>   meta/recipes-extended/libtirpc/libtirpc_0.2.4.bb, do_compile
>   meta/recipes-extended/iptables/iptables_1.4.21.bb, do_compile
>   meta/recipes-connectivity/nfs-utils/libnfsidmap_0.25.bb, do_compile
>   meta/recipes-devtools/perl/perl_5.20.0.bb, do_compile
>   meta/recipes-extended/sysklogd/sysklogd_1.5.bb, do_compile
>   meta/recipes-extended/ethtool/ethtool_3.14.bb, do_compile
>   meta/recipes-extended/psmisc/psmisc_22.21.bb, do_compile
>   meta/recipes-extended/pax/pax_3.4.bb, do_compile
>   meta/recipes-extended/procps/procps_3.2.8.bb, do_compile
>   meta/recipes-extended/net-tools/net-tools_1.60-25.bb, do_compile
>   meta/recipes-extended/findutils/findutils_4.4.2.bb, do_compile
>   meta/recipes-extended/logrotate/logrotate_3.8.7.bb, do_compile
>   meta/recipes-extended/cronie/cronie_1.4.11.bb, do_compile
>   meta/recipes-devtools/elfutils/elfutils_0.148.bb, do_compile
>   meta/recipes-extended/at/at_3.1.14.bb, do_compile
>   meta/recipes-extended/iputils/iputils_s20121221.bb, do_compile
>   meta/recipes-extended/gzip/gzip_1.6.bb, do_compile
>   meta/recipes-extended/sudo/sudo_1.8.10p3.bb, do_compile
>   meta/recipes-extended/mc/mc_4.8.12.bb, do_compile
>
> I can share detailed logs if anyone wants them. I'll have a look and see if I
> can find some low hanging fruit amongst these which are easy to fix.
>


Thanks cool. Can you open a ticket on github with this information ?

Re: Announcing meta-musl OE/Yocto layer supporting musl C library

[Paul Barker]

[-- Attachment #1: Type: text/plain, Size: 3771 bytes --]

On Mon, Aug 18, 2014 at 11:26:26AM -0700, Khem Raj wrote:
> On Mon, Aug 18, 2014 at 11:14 AM, Paul Barker <paul@paulbarker.me.uk> wrote:
> > On Sat, Aug 16, 2014 at 05:50:38PM -0700, Khem Raj wrote:
> >> Hello All
> >>
> >> musl <http://www.musl-libc.org/> is a newer implementation licensed
> >> under MIT licence
> >> A layer to support musl to supply your system C library as an
> >> alternative to uclibc and eglibc/glibc  is now available at
> >>
> >> https://github.com/kraj/meta-musl
> >>
> >> README describes way to get going with it.
> >>
> >> Few things to note
> >>
> >> * Supports gcc 4.9 in OE not 4.8
> >> * Current musl recipe is 1.1.4 based
> >> * You need latest master of OE-Core
> >> (106305227003761c3fc562c21bb859a5256f2b36) or          newer and
> >> bitbake
> >> * Its only able to build core-image-minimal
> >> * core-image-minimal boots on qemuarm, qemuppc, qemumips, qemux86, qemux86-64
> >>
> >> I will send a request to add it to layerindex soon and then you should
> >> be able to find it using
> >> layerindex.opernembedded.org as well
> >>
> >> If you try it out on your hardware and/or have feedback I am
> >> interested. Fork it from github, file bugs
> >> send patches ....
> >>
> >> Happy hacking
> >>
> >
> > core-image-minimal works with my configuration on qemuarm. It also successfully
> > builds what will soon be released as opkg-0.2.3 and the recipe for toybox-0.4.9
> > which I've recently created.
> >
> > I tried building core-image-full-cmdline as I think that would be good next step
> > from core-image-minimal. I'm seeing failures in:
> >
> >   meta/recipes-devtools/e2fsprogs/e2fsprogs_1.42.9.bb, do_compile
> >   meta/recipes-extended/tcp-wrappers/tcp-wrappers_7.6.bb, do_compile
> >   meta/recipes-extended/libtirpc/libtirpc_0.2.4.bb, do_compile
> >   meta/recipes-extended/iptables/iptables_1.4.21.bb, do_compile
> >   meta/recipes-connectivity/nfs-utils/libnfsidmap_0.25.bb, do_compile
> >   meta/recipes-devtools/perl/perl_5.20.0.bb, do_compile
> >   meta/recipes-extended/sysklogd/sysklogd_1.5.bb, do_compile
> >   meta/recipes-extended/ethtool/ethtool_3.14.bb, do_compile
> >   meta/recipes-extended/psmisc/psmisc_22.21.bb, do_compile
> >   meta/recipes-extended/pax/pax_3.4.bb, do_compile
> >   meta/recipes-extended/procps/procps_3.2.8.bb, do_compile
> >   meta/recipes-extended/net-tools/net-tools_1.60-25.bb, do_compile
> >   meta/recipes-extended/findutils/findutils_4.4.2.bb, do_compile
> >   meta/recipes-extended/logrotate/logrotate_3.8.7.bb, do_compile
> >   meta/recipes-extended/cronie/cronie_1.4.11.bb, do_compile
> >   meta/recipes-devtools/elfutils/elfutils_0.148.bb, do_compile
> >   meta/recipes-extended/at/at_3.1.14.bb, do_compile
> >   meta/recipes-extended/iputils/iputils_s20121221.bb, do_compile
> >   meta/recipes-extended/gzip/gzip_1.6.bb, do_compile
> >   meta/recipes-extended/sudo/sudo_1.8.10p3.bb, do_compile
> >   meta/recipes-extended/mc/mc_4.8.12.bb, do_compile
> >
> > I can share detailed logs if anyone wants them. I'll have a look and see if I
> > can find some low hanging fruit amongst these which are easy to fix.
> >
> 
> 
> Thanks cool. Can you open a ticket on github with this information ?

Will do. Do you want just want one ticket or do you want to separate these into
one ticket per failing recipe or similar?

I've put the failing "log.do_compile" files into one directory which is 1.3MB
uncompressed or just 41kB compressed. I've posted them to:
  http://www.paulbarker.me.uk/musl-error-logs_2014-08-18/
  http://www.paulbarker.me.uk/musl-error-logs_2014-08-18/musl-error-logs_2014-08-18.tar.xz

Cheers,

-- 
Paul Barker

Email: paul@paulbarker.me.uk
http://www.paulbarker.me.uk

[-- Attachment #2: Type: application/pgp-signature, Size: 484 bytes --]

Re: Announcing meta-musl OE/Yocto layer supporting musl C library

[Khem Raj]

On Mon, Aug 18, 2014 at 11:50 AM, Paul Barker <paul@paulbarker.me.uk> wrote:
> On Mon, Aug 18, 2014 at 11:26:26AM -0700, Khem Raj wrote:
>> On Mon, Aug 18, 2014 at 11:14 AM, Paul Barker <paul@paulbarker.me.uk> wrote:
>> > On Sat, Aug 16, 2014 at 05:50:38PM -0700, Khem Raj wrote:
>> >> Hello All
>> >>
>> >> musl <http://www.musl-libc.org/> is a newer implementation licensed
>> >> under MIT licence
>> >> A layer to support musl to supply your system C library as an
>> >> alternative to uclibc and eglibc/glibc  is now available at
>> >>
>> >> https://github.com/kraj/meta-musl
>> >>
>> >> README describes way to get going with it.
>> >>
>> >> Few things to note
>> >>
>> >> * Supports gcc 4.9 in OE not 4.8
>> >> * Current musl recipe is 1.1.4 based
>> >> * You need latest master of OE-Core
>> >> (106305227003761c3fc562c21bb859a5256f2b36) or          newer and
>> >> bitbake
>> >> * Its only able to build core-image-minimal
>> >> * core-image-minimal boots on qemuarm, qemuppc, qemumips, qemux86, qemux86-64
>> >>
>> >> I will send a request to add it to layerindex soon and then you should
>> >> be able to find it using
>> >> layerindex.opernembedded.org as well
>> >>
>> >> If you try it out on your hardware and/or have feedback I am
>> >> interested. Fork it from github, file bugs
>> >> send patches ....
>> >>
>> >> Happy hacking
>> >>
>> >
>> > core-image-minimal works with my configuration on qemuarm. It also successfully
>> > builds what will soon be released as opkg-0.2.3 and the recipe for toybox-0.4.9
>> > which I've recently created.
>> >
>> > I tried building core-image-full-cmdline as I think that would be good next step
>> > from core-image-minimal. I'm seeing failures in:
>> >
>> >   meta/recipes-devtools/e2fsprogs/e2fsprogs_1.42.9.bb, do_compile
>> >   meta/recipes-extended/tcp-wrappers/tcp-wrappers_7.6.bb, do_compile
>> >   meta/recipes-extended/libtirpc/libtirpc_0.2.4.bb, do_compile
>> >   meta/recipes-extended/iptables/iptables_1.4.21.bb, do_compile
>> >   meta/recipes-connectivity/nfs-utils/libnfsidmap_0.25.bb, do_compile
>> >   meta/recipes-devtools/perl/perl_5.20.0.bb, do_compile
>> >   meta/recipes-extended/sysklogd/sysklogd_1.5.bb, do_compile
>> >   meta/recipes-extended/ethtool/ethtool_3.14.bb, do_compile
>> >   meta/recipes-extended/psmisc/psmisc_22.21.bb, do_compile
>> >   meta/recipes-extended/pax/pax_3.4.bb, do_compile
>> >   meta/recipes-extended/procps/procps_3.2.8.bb, do_compile
>> >   meta/recipes-extended/net-tools/net-tools_1.60-25.bb, do_compile
>> >   meta/recipes-extended/findutils/findutils_4.4.2.bb, do_compile
>> >   meta/recipes-extended/logrotate/logrotate_3.8.7.bb, do_compile
>> >   meta/recipes-extended/cronie/cronie_1.4.11.bb, do_compile
>> >   meta/recipes-devtools/elfutils/elfutils_0.148.bb, do_compile
>> >   meta/recipes-extended/at/at_3.1.14.bb, do_compile
>> >   meta/recipes-extended/iputils/iputils_s20121221.bb, do_compile
>> >   meta/recipes-extended/gzip/gzip_1.6.bb, do_compile
>> >   meta/recipes-extended/sudo/sudo_1.8.10p3.bb, do_compile
>> >   meta/recipes-extended/mc/mc_4.8.12.bb, do_compile
>> >
>> > I can share detailed logs if anyone wants them. I'll have a look and see if I
>> > can find some low hanging fruit amongst these which are easy to fix.
>> >
>>
>>
>> Thanks cool. Can you open a ticket on github with this information ?
>
> Will do. Do you want just want one ticket or do you want to separate these into
> one ticket per failing recipe or similar?
>

That would be handy too. But I wanted was to keep this info in project
for folks to try it out.

> I've put the failing "log.do_compile" files into one directory which is 1.3MB
> uncompressed or just 41kB compressed. I've posted them to:
>   http://www.paulbarker.me.uk/musl-error-logs_2014-08-18/
>   http://www.paulbarker.me.uk/musl-error-logs_2014-08-18/musl-error-logs_2014-08-18.tar.xz
>
> Cheers,
>
> --
> Paul Barker
>
> Email: paul@paulbarker.me.uk
> http://www.paulbarker.me.uk

Re: Announcing meta-musl OE/Yocto layer supporting musl C library

[Paul Barker]

[-- Attachment #1: Type: text/plain, Size: 714 bytes --]

On Mon, Aug 18, 2014 at 12:04:54PM -0700, Khem Raj wrote:
> On Mon, Aug 18, 2014 at 11:50 AM, Paul Barker <paul@paulbarker.me.uk> wrote:
> > On Mon, Aug 18, 2014 at 11:26:26AM -0700, Khem Raj wrote:
> >>
> >> Thanks cool. Can you open a ticket on github with this information ?
> >
> > Will do. Do you want just want one ticket or do you want to separate these into
> > one ticket per failing recipe or similar?
> >
> 
> That would be handy too. But I wanted was to keep this info in project
> for folks to try it out.
> 

I'll just post one ticket for now.

I think I've got a fix for e2fsprogs as well.

Cheers,

-- 
Paul Barker

Email: paul@paulbarker.me.uk
http://www.paulbarker.me.uk

[-- Attachment #2: Type: application/pgp-signature, Size: 484 bytes --]

Re: Announcing meta-musl OE/Yocto layer supporting musl C library

[Khem Raj]

On Mon, Aug 18, 2014 at 12:08 PM, Paul Barker <paul@paulbarker.me.uk> wrote:
>> That would be handy too. But I wanted was to keep this info in project
>> for folks to try it out.
>>
>
> I'll just post one ticket for now.
>

Thats fine too.

> I think I've got a fix for e2fsprogs as well.

Great. when we write patches that arise out of musl systems we should
make sure that

1. If patch fixes a genuine issue surfaced with musl, fight it out ar
respective package upstream and in OE universe add it to the original
layer where recipe primarily resides
but we can keep the patches in meta-musl as a last resort.

2. If its something musl specific then lets keep it in meta-musl and
see how musl can be changed to fix it.

Thanks
-Khem

Re: Announcing meta-musl OE/Yocto layer supporting musl C library

[Paul Barker]

[-- Attachment #1: Type: text/plain, Size: 1308 bytes --]

On Mon, Aug 18, 2014 at 12:14:08PM -0700, Khem Raj wrote:
> On Mon, Aug 18, 2014 at 12:08 PM, Paul Barker <paul@paulbarker.me.uk> wrote:
> >> That would be handy too. But I wanted was to keep this info in project
> >> for folks to try it out.
> >>
> >
> > I'll just post one ticket for now.
> >
> 
> Thats fine too.
> 
> > I think I've got a fix for e2fsprogs as well.
> 
> Great. when we write patches that arise out of musl systems we should
> make sure that
> 
> 1. If patch fixes a genuine issue surfaced with musl, fight it out ar
> respective package upstream and in OE universe add it to the original
> layer where recipe primarily resides
> but we can keep the patches in meta-musl as a last resort.
> 
> 2. If its something musl specific then lets keep it in meta-musl and
> see how musl can be changed to fix it.
> 
> Thanks
> -Khem

That makes sense. I'd suggest adding that to the README file in meta-musl so
other would-be-contributors notice it.

In this case, it's ensuring that 'uint64_t' is used instead of '__uint64_t' and
that <limits.h> is included where needed for 'PATH_MAX'. So I'd say those fixes
belong in oe-core and are candidates for submitting upstream.

Thanks,

-- 
Paul Barker

Email: paul@paulbarker.me.uk
http://www.paulbarker.me.uk

[-- Attachment #2: Type: application/pgp-signature, Size: 484 bytes --]

Re: Announcing meta-musl OE/Yocto layer supporting musl C library

[Paul Barker]

[-- Attachment #1: Type: text/plain, Size: 1904 bytes --]

On Mon, Aug 18, 2014 at 07:31:02PM +0000, Paul Barker wrote:
> On Mon, Aug 18, 2014 at 12:14:08PM -0700, Khem Raj wrote:
> > On Mon, Aug 18, 2014 at 12:08 PM, Paul Barker <paul@paulbarker.me.uk> wrote:
> > >> That would be handy too. But I wanted was to keep this info in project
> > >> for folks to try it out.
> > >>
> > >
> > > I'll just post one ticket for now.
> > >
> > 
> > Thats fine too.
> > 
> > > I think I've got a fix for e2fsprogs as well.
> > 
> > Great. when we write patches that arise out of musl systems we should
> > make sure that
> > 
> > 1. If patch fixes a genuine issue surfaced with musl, fight it out ar
> > respective package upstream and in OE universe add it to the original
> > layer where recipe primarily resides
> > but we can keep the patches in meta-musl as a last resort.
> > 
> > 2. If its something musl specific then lets keep it in meta-musl and
> > see how musl can be changed to fix it.
> > 
> > Thanks
> > -Khem
> 
> That makes sense. I'd suggest adding that to the README file in meta-musl so
> other would-be-contributors notice it.
> 
> In this case, it's ensuring that 'uint64_t' is used instead of '__uint64_t' and
> that <limits.h> is included where needed for 'PATH_MAX'. So I'd say those fixes
> belong in oe-core and are candidates for submitting upstream.
> 

Actually, both of these are fixed upstream already, but not as isolated commits
I can backport. There is a newer upstream release (v1.42.11) but it doesn't
contain the necessary fixes. It appears that the missing <limits.h> inclusion is
in a file added by patches in openembedded-core.

I shan't attempt the upgrade as there are a lot of patches in openembedded-core
to consider. I will try to make an isolated fix for the build issues with musl
though.

Cheers,

-- 
Paul Barker

Email: paul@paulbarker.me.uk
http://www.paulbarker.me.uk

[-- Attachment #2: Type: application/pgp-signature, Size: 484 bytes --]