The broken 0.13 release for GuixSD

The broken 0.13 release for GuixSD

[ng0]

Going from within GuixSD 0.12 to 0.13 (multiple guix pull issued to get
all the dependencies you might lack) works.

GuixSD 0.13 medium on new installations, with no system present previously,
so far failed on all types of hardware from the last 1 - 15 years for me including
bios and uefi systems.

All get stuck on the same test.
This should not happen at all with a release.


Last I will try today is create an install target from this 0.13 resembling
machine I'm writing on right now, to see if this magically fixes the tests.

Re: The broken 0.13 release for GuixSD

[ng0]

On Wed, 24 May 2017 09:59:17 +0200 (CEST), <ng0@pragmatique.xyz> wrote:

> Going from within GuixSD 0.12 to 0.13 (multiple guix pull issued to get
> all the dependencies you might lack) works.
> 
> GuixSD 0.13 medium on new installations, with no system present previously,
> so far failed on all types of hardware from the last 1 - 15 years for me including
> bios and uefi systems.
> 
> All get stuck on the same test.
> This should not happen at all with a release.
> 
> 
> Last I will try today is create an install target from this 0.13 resembling
> machine I'm writing on right now, to see if this magically fixes the tests.


Wow, this is fckd up. You know how I made the install pass (or at least go
beyond the point of the test failing)?

guix pull; guix environment --ad-hoc guix
in environment: guix system init /mnt/etc/config.scm /mnt

what?

Re: The broken 0.13 release for GuixSD

[Fox]

never mind the wrong disk order in the part about VM install (qemu)

Re: The broken 0.13 release for GuixSD

[ngillmann]

On Wed, 24 May 2017 11:17:11 +0200, Fox <firefox@firemail.cc> wrote:

> never mind the wrong disk order in the part about VM install (qemu)


What's up with these random messages without context?
As you might know or not know I know my way around in GuixSD and
am not just trying to get started with it.

disk order

[Fox]

referring to
https://www.gnu.org/software/guix/manual/guix.html#Installing-GuixSD-in-a-VM


qemu-system-x86_64 -m 1024 -smp 1 \
   -net user -net nic,model=virtio -boot menu=on \
   -drive file=guixsd.img \
   -drive file=guixsd-usb-install-0.13.0.system




will not boot in qemu unless you change it to

qemu-system-x86_64 -m 1024 -smp 1 \
   -net user -net nic,model=virtio -boot menu=on \
   -drive file=guixsd-usb-install-0.13.0.system
   -drive file=guixsd.img \

Re: disk order

[ng0]

How is that  even related to what I wrote.
That's what I meant with "your message is random and unrelated".

On Wed, 24 May 2017 11:24:42 +0200, Fox <firefox@firemail.cc> wrote:

> referring to
> https://www.gnu.org/software/guix/manual/guix.html#Installing-GuixSD-in-a-VM
> 
> 
> qemu-system-x86_64 -m 1024 -smp 1 \
>    -net user -net nic,model=virtio -boot menu=on \
>    -drive file=guixsd.img \
>    -drive file=guixsd-usb-install-0.13.0.system
> 
> 
> 
> 
> will not boot in qemu unless you change it to
> 
> qemu-system-x86_64 -m 1024 -smp 1 \
>    -net user -net nic,model=virtio -boot menu=on \
>    -drive file=guixsd-usb-install-0.13.0.system
>    -drive file=guixsd.img \

Re: disk order

[Vincent Legoll]

Hello,

On Wed, May 24, 2017 at 11:24 AM, Fox <firefox@firemail.cc> wrote:
> referring to
> https://www.gnu.org/software/guix/manual/guix.html#Installing-GuixSD-in-a-VM
>
> qemu-system-x86_64 -m 1024 -smp 1 \
>   -net user -net nic,model=virtio -boot menu=on \
>   -drive file=guixsd.img \
>   -drive file=guixsd-usb-install-0.13.0.system
>
> will not boot in qemu unless you change it to
>
> qemu-system-x86_64 -m 1024 -smp 1 \
>   -net user -net nic,model=virtio -boot menu=on \
>   -drive file=guixsd-usb-install-0.13.0.system
>   -drive file=guixsd.img \

Careful with that backslash, Eugene (*)...

-- 
Vincent Legoll

disk order

[Fox]

> will not boot in qemu unless you change it to
>
> qemu-system-x86_64 -m 1024 -smp 1 \
>   -net user -net nic,model=virtio -boot menu=on \
>   -drive file=guixsd-usb-install-0.13.0.system \
>   -drive file=guixsd.img

slash typo

Re: disk order

[ng0]

On Wed, 24 May 2017 11:30:03 +0200 (CEST), <ng0@pragmatique.xyz> wrote:

> How is that  even related to what I wrote.
> That's what I meant with "your message is random and unrelated".
> 
> On Wed, 24 May 2017 11:24:42 +0200, Fox <firefox@firemail.cc> wrote:
> 
> > referring to
> > https://www.gnu.org/software/guix/manual/guix.html#Installing-GuixSD-in-a-VM
> > 
> > 
…

This is a prime example of "people will never learn not to do things like
thread hijacking, cross-posting, etc"...

I'm open for constructive discussion on the original thread (which
for the record was "why do I have to run guix environment --ad-hoc guix"
in the SD installer on baremetal not in a VM"), I will ignore
this hijacked thread now. Please don't include me anymore in replies.
Thanks.

bug#27052: [PATCH version-0.13.0] doc: Fix qemu-system-x86_64 example.

[Danny Milosavljevic]

* doc/guix.texi (Running GuixSD in a VM): Fix qemu-system-x86_64 example.
---
 doc/guix.texi | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/doc/guix.texi b/doc/guix.texi
index aa8b705be..a662eb236 100644
--- a/doc/guix.texi
+++ b/doc/guix.texi
@@ -7687,8 +7687,8 @@ Boot the USB installation image in an VM:
 @example
 qemu-system-x86_64 -m 1024 -smp 1 \
   -net user -net nic,model=virtio -boot menu=on \
-  -drive file=guixsd.img \
-  -drive file=guixsd-usb-install-@value{VERSION}.@var{system}
+  -drive file=guixsd-usb-install-@value{VERSION}.@var{system} \
+  -drive file=guixsd.img
 @end example
 
 In the VM console, quickly press the @kbd{F12} key to enter the boot

thread hijacking

[Fox]

dude, I did not read past your headline "broken release ..." and I am 
not obliged to.



> This is a prime example of "people will never learn not to do things like
> thread hijacking, cross-posting, etc"...
>
> I'm open for constructive discussion on the original thread (which
> for the record was "why do I have to run guix environment --ad-hoc guix"
> in the SD installer on baremetal not in a VM"), I will ignore
> this hijacked thread now. Please don't include me anymore in replies.
> Thanks.

Re: thread hijacking

[ng0]

On Wed, 24 May 2017 11:58:36 +0200, Fox <firefox@firemail.cc> wrote:

> dude, I did not read past your headline "broken release ..." and I am 
> not obliged to.
> 
> 
> 
> > This is a prime example of "people will never learn not to do things like
> > thread hijacking, cross-posting, etc"...
> >
> > I'm open for constructive discussion on the original thread (which
> > for the record was "why do I have to run guix environment --ad-hoc guix"
> > in the SD installer on baremetal not in a VM"), I will ignore
> > this hijacked thread now. Please don't include me anymore in replies.
> > Thanks.


https://www.ietf.org/rfc/rfc1855.txt and common sense still applies in 2017.
Either make use of the mailinglist the correct way or don't use it at all.

Re: The broken 0.13 release for GuixSD

[Ricardo Wurmus]

ng0@pragmatique.xyz writes:

> Going from within GuixSD 0.12 to 0.13 (multiple guix pull issued to get
> all the dependencies you might lack) works.
>
> GuixSD 0.13 medium on new installations, with no system present previously,
> so far failed on all types of hardware from the last 1 - 15 years for me including
> bios and uefi systems.
>
> All get stuck on the same test.
> This should not happen at all with a release.

Please, the least you could do when using words like “broken” and
“should not happen at all” is to be precise about *what* fails.

This is not constructive.

-- 
Ricardo

GPG: BCA6 89B6 3655 3801 C3C6  2150 197A 5888 235F ACAC
https://elephly.net

Re: The broken 0.13 release for GuixSD

[ng0]

On Wed, 24 May 2017 12:26:34 +0200, Ricardo Wurmus <rekado@elephly.net> wrote:

> 
> ng0@pragmatique.xyz writes:
> 
> > Going from within GuixSD 0.12 to 0.13 (multiple guix pull issued to get
> > all the dependencies you might lack) works.
> >
> > GuixSD 0.13 medium on new installations, with no system present previously,
> > so far failed on all types of hardware from the last 1 - 15 years for me including
> > bios and uefi systems.
> >
> > All get stuck on the same test.
> > This should not happen at all with a release.
> 
> Please, the least you could do when using words like “broken” and
> “should not happen at all” is to be precise about *what* fails.
> 
> This is not constructive.
> 
> -- 
> Ricardo
> 
> GPG: BCA6 89B6 3655 3801 C3C6  2150 197A 5888 235F ACAC
> https://elephly.net


This https://debbugs.gnu.org/cgi/bugreport.cgi?bug=27034 and https://debbugs.gnu.org/cgi/bugreport.cgi?bug=27032 this.

But I can add more duplication and describe what happened in my case:

Situation and Expectation:

uefi and bios (separate "and", not "bios and uefi") computers
No system installed.
GuixSD 0.13 installer should simply work as per instructions and the
way I know it.

Problem:

When you reach the point where you run "guix system init /path/to/config.scm /mnt"
and guix gets build, reaching its 'make check', 2 tests (as linked above) fail.

Solution:

See previous message I wrote here about guix environment --ad-hoc guix.

Re: The broken 0.13 release for GuixSD

[Ricardo Wurmus]

>> Please, the least you could do when using words like “broken” and
>> “should not happen at all” is to be precise about *what* fails.

[…]

> This https://debbugs.gnu.org/cgi/bugreport.cgi?bug=27034 and
> https://debbugs.gnu.org/cgi/bugreport.cgi?bug=27032 this.

Then it’s better to add additional information to these bug reports
instead of diluting the discussion by starting a new thread on
guix-devel.

--
Ricardo

GPG: BCA6 89B6 3655 3801 C3C6  2150 197A 5888 235F ACAC
https://elephly.net

0.13 release for GuixSD

[Fox]

as I typed in IRC, I encountered some test fail too earlier. Sorry I 
forgot how I got over it, maybe it was that install where I missed to 
submit one BASH LINE from the install manual (might have been fixed by 
now anyway).
Just to say that test failures do occur.

Re: 0.13 release for GuixSD

[Ludovic Courtès]

Fox <firefox@firemail.cc> skribis:

> as I typed in IRC, I encountered some test fail too earlier. Sorry I
> forgot how I got over it, maybe it was that install where I missed to
> submit one BASH LINE from the install manual (might have been fixed by
> now anyway).
> Just to say that test failures do occur.

This is being worked on, please see <https://bugs.gnu.org/27032>.  You
can contribute to the discussion there by emailing
27032@debbugs.gnu.org.

Thanks in advance!

Ludo’.

bug#27052: [PATCH version-0.13.0] doc: Fix qemu-system-x86_64 example.

[Ludovic Courtès]

Danny Milosavljevic <dannym@scratchpost.org> skribis:

> * doc/guix.texi (Running GuixSD in a VM): Fix qemu-system-x86_64 example.
> ---
>  doc/guix.texi | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/doc/guix.texi b/doc/guix.texi
> index aa8b705be..a662eb236 100644
> --- a/doc/guix.texi
> +++ b/doc/guix.texi
> @@ -7687,8 +7687,8 @@ Boot the USB installation image in an VM:
>  @example
>  qemu-system-x86_64 -m 1024 -smp 1 \
>    -net user -net nic,model=virtio -boot menu=on \
> -  -drive file=guixsd.img \
> -  -drive file=guixsd-usb-install-@value{VERSION}.@var{system}
> +  -drive file=guixsd-usb-install-@value{VERSION}.@var{system} \
> +  -drive file=guixsd.img
>  @end example
>  
>  In the VM console, quickly press the @kbd{F12} key to enter the boot

Should we add a sentence saying that ordering matters?

Otherwise LGTM, thanks!

Ludo’.

bug#27052: [PATCH version-0.13.0] doc: Fix qemu-system-x86_64 example.

[Danny Milosavljevic]

Hi Ludo,

should I commit this to master?  Will version-0.13.1 or later merge from master?

bug#27052: [PATCH version-0.13.0] doc: Fix qemu-system-x86_64 example.

[Ludovic Courtès]

Danny Milosavljevic <dannym@scratchpost.org> skribis:

> should I commit this to master?  Will version-0.13.1 or later merge from master?

Please commit to version-0.13.0; we’ll merge it later in master.

Thanks,
Ludo’.