Call for testing

Call for testing

[Ludovic Courtès]

[-- Attachment #1: Type: text/plain, Size: 1080 bytes --]

Hello!

Please consider the following GuixSD installation image for testing, and
report any issues or successes!

  https://www.fdn.fr/~lcourtes/software/guix/guixsd-usb-install-0.10.0pre.x86_64-linux.xz
  SHA1: 85296c9352740fbcc2b83211a1af39302d9d5f41

  https://www.fdn.fr/~lcourtes/software/guix/guixsd-usb-install-0.10.0pre.x86_64-linux.xz.sig

Likewise, a binary installation tarball:

  https://www.fdn.fr/~lcourtes/software/guix/guix-binary-0.10.0pre.x86_64-linux.tar.xz
  SHA1: aa2772dccc342a49ef6757c3ca9a831210dbdb92

  https://www.fdn.fr/~lcourtes/software/guix/guix-binary-0.10.0pre.x86_64-linux.tar.xz.sig

Please follow the instructions of the manual that’s in Git master; the
copy that’s on www.gnu.org is a bit outdated (the GuixSD installation
image above includes the latest manual on tty2.)

As you can see, due to popular demand ;-), the minor version number has
been bumped.  I think it makes sense.

I’m planning to release sometime between Sunday and Tuesday (long
week-end plus one day off for me.)

TIA.  :-)

Ludo’.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 818 bytes --]

Re: Call for testing

[Tobias Geerinckx-Rice]

Woo!

On 25/03/2016, Ludovic Courtès <ludo@gnu.org> wrote:
> As you can see, due to popular demand ;-), the minor version number has
> been bumped.  I think it makes sense.

I'll take an extra laptop along for Easter dinner... :-)

Thanks to all contributors,

T G-R

Re: Call for testing

[Jan Nieuwenhuizen]

Ludovic Courtès writes:

Hi,

> Please consider the following GuixSD installation image for testing, and
> report any issues or successes!
>
>   https://www.fdn.fr/~lcourtes/software/guix/guixsd-usb-install-0.10.0pre.x86_64-linux.xz

I tried this.

  + Booting and instructions are great
  - Please provide and suggest tested desktop.scm and server.scm
  - It seems that network is mandatory.  If that is so, you may want
    to stress that.  guix system init failed on me several times with

   guix system: error: build failed: while setting up the build environment: getting attributes of path `/etc/resolv.conf': No such file or directory.

    It would be nice if an install was possible without network, e.g.
    server.scm/minimal.scm.

  - It seems that after this error, grub still got (half-) installed,
    because i could not boot anymore :-( Fell into grub rescue> which
    used to be just great once, but now I could not load a kernel there,
    because of "error: symbol 'grub_term_highlight_color" not found" :-(

Greetings,
Jan

-- 
Jan Nieuwenhuizen <janneke@gnu.org> | GNU LilyPond http://lilypond.org
Freelance IT http://JoyofSource.com | Avatar®  http://AvatarAcademy.nl  

Re: Call for testing

[Leo Famulari]

On Fri, Mar 25, 2016 at 11:12:35AM +0100, Ludovic Courtès wrote:
> Hello!
> 
> Please consider the following GuixSD installation image for testing, and
> report any issues or successes!

I built an i686 installer against 36c2c8c8 (gnu: beets: Run tests with
python-nose.) and used that.

Do you think the motd should suggest that the user runs `guix pull`?
It's not necessary now, but in a couple months it would be helpful.

I didn't experience any problems with nscd (<https://bugs.gnu.org/22209>).

I first tried to install on a btrfs filesystem but the partition could
not be found by its label upon rebooting. Well, the manual did warn me
about this ;)

Re: Call for testing

[Ludovic Courtès]

Leo Famulari <leo@famulari.name> skribis:

> I built an i686 installer against 36c2c8c8 (gnu: beets: Run tests with
> python-nose.) and used that.
>
> Do you think the motd should suggest that the user runs `guix pull`?
> It's not necessary now, but in a couple months it would be helpful.

I thought we’d need a better fix rather than warning users about the
same thing in several different places but… I haven’t come up with such
a fix yet.  :-/

> I didn't experience any problems with nscd (<https://bugs.gnu.org/22209>).

Good.

> I first tried to install on a btrfs filesystem but the partition could
> not be found by its label upon rebooting. Well, the manual did warn me
> about this ;)

Heh.  :-)

Thanks for testing!

Ludo’.

Re: Call for testing

[Ludovic Courtès]

Jan Nieuwenhuizen <janneke@gnu.org> skribis:

> I tried this.
>
>   + Booting and instructions are great
>   - Please provide and suggest tested desktop.scm and server.scm

Yeah, though each person has their own idea of what “server” or
“desktop” means (to me “desktop” means ratpoison, but I’ve heard about
that GNOME thing that many like to use ;-)).  I guess we could add more
examples, probably not for this time, though.

>   - It seems that network is mandatory.  If that is so, you may want
>     to stress that.  guix system init failed on me several times with
>
>    guix system: error: build failed: while setting up the build environment: getting attributes of path `/etc/resolv.conf': No such file or directory.

OK, I’ve added a note to emphasize that.

>     It would be nice if an install was possible without network, e.g.
>     server.scm/minimal.scm.

Yes, though again it’s difficult because any change in the config may
lead to additional downloads.

>   - It seems that after this error, grub still got (half-) installed,
>     because i could not boot anymore :-( Fell into grub rescue> which
>     used to be just great once, but now I could not load a kernel there,
>     because of "error: symbol 'grub_term_highlight_color" not found" :-(

So you did a ‘guix system init’ that failed with the error above, and
then you rebooted and got a broken GRUB?

Unless I’m mistaken, a “build failed” error prevents the rest of the
installation, including GRUB install, from happening.  Not sure what
happened.

Thanks for your feedback!

Ludo’.

Re: Call for testing

[Jan Nieuwenhuizen]

Ludovic Courtès writes:

>>   - Please provide and suggest tested desktop.scm and server.scm
>
> Yeah, though each person has their own idea of what “server” or
> “desktop” means

Sure, it would help with bug reports like mine... Can you try the
plain xxx.scm?

>>     It would be nice if an install was possible without network, e.g.
>>     server.scm/minimal.scm.
>
> Yes, though again it’s difficult because any change in the config may
> lead to additional downloads.

Exactly, even more difficult for a user to determine.  Just a
suggestion.

> So you did a ‘guix system init’ that failed with the error above, and
> then you rebooted and got a broken GRUB?

Yes, that has been my perception.

> Unless I’m mistaken, a “build failed” error prevents the rest of the
> installation, including GRUB install, from happening.  Not sure what
> happened.

Yes, that was what I was hoping; quite possibly I goofed up somewhere.
I'm quite terrible at following instructions manually; scripts and
second-tries work magic for me :-)

Greetings,
Jan

-- 
Jan Nieuwenhuizen <janneke@gnu.org> | GNU LilyPond http://lilypond.org
Freelance IT http://JoyofSource.com | Avatar®  http://AvatarAcademy.nl  

Re: Call for testing

[Tobias Geerinckx-Rice]

On 28/03/2016, Leo Famulari <leo@famulari.name> wrote:
> I first tried to install on a btrfs filesystem but the partition could
> not be found by its label upon rebooting. Well, the manual did warn me
> about this ;)

I hope to send a rough series of patches for btrfs root support soon,
for unknown values of soon.

I spent most of my time so far reverse-engineering my machine's
firmware to re-enable the well-hidden CPU virtualisation support
required to even build the installer.

Such is life.

Kind regards,

T G-R

Re: Call for testing

[Ludovic Courtès]

Tobias Geerinckx-Rice <tobias.geerinckx.rice@gmail.com> skribis:

> On 28/03/2016, Leo Famulari <leo@famulari.name> wrote:
>> I first tried to install on a btrfs filesystem but the partition could
>> not be found by its label upon rebooting. Well, the manual did warn me
>> about this ;)
>
> I hope to send a rough series of patches for btrfs root support soon,
> for unknown values of soon.

Cool!  :-)

> I spent most of my time so far reverse-engineering my machine's
> firmware to re-enable the well-hidden CPU virtualisation support
> required to even build the installer.

Uh.  What machine is this?  (So we don’t buy the same… ;-))

Ludo’.

Re: Call for testing

[Daniel Pimentel]

I did upgrade on my computers (x86 and i386(thinkpad)) to new GuixSD 
version with GNOME and it works!

Thanks for all :)

-- 
Daniel Pimentel (d4n1)

Re: Call for testing

[Leo Famulari]

On Tue, Mar 29, 2016 at 08:30:22AM +0200, Tobias Geerinckx-Rice wrote:
> On 28/03/2016, Leo Famulari <leo@famulari.name> wrote:
> > I first tried to install on a btrfs filesystem but the partition could
> > not be found by its label upon rebooting. Well, the manual did warn me
> > about this ;)
> 
> I hope to send a rough series of patches for btrfs root support soon,
> for unknown values of soon.

Cool, I'm looking forward to them! I'm happy to help test or debug!

> 
> I spent most of my time so far reverse-engineering my machine's
> firmware to re-enable the well-hidden CPU virtualisation support
> required to even build the installer.
> 
> Such is life.
> 
> Kind regards,
> 
> T G-R

Re: Call for testing

[Tobias Geerinckx-Rice]

Hullo Ludo',

On 29/03/2016, Ludovic Courtès <ludo@gnu.org> wrote:
> Tobias Geerinckx-Rice <tobias.geerinckx.rice@gmail.com> skribis:
>> I spent most of my time so far reverse-engineering my machine's
>> firmware to re-enable the well-hidden CPU virtualisation support
>> required to even build the installer.
>
> Uh.  What machine is this?  (So we don’t buy the same… ;-))

Unlikely. It's a donated Acer Aspire V5-121 from 2013 which, perhaps
due to it being Swiss, has weird UEFI firmware that's different from
what it should be according to, well, everybody.

Kind regards,

T G-R