GNU bug report logs -
#52539
Fwd: Comments in /etc/passwd don't get updated
Previous Next
To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 52539 in the body.
You can then email your comments to 52539 AT debbugs.gnu.org in the normal way.
Toggle the display of automated, internal messages from the tracker.
Report forwarded
to
bug-guix <at> gnu.org
:
bug#52539
; Package
guix
.
(Thu, 16 Dec 2021 07:02:02 GMT)
Full text and
rfc822 format available.
Acknowledgement sent
to
Jacob First <jacob.first <at> member.fsf.org>
:
New bug report received and forwarded. Copy sent to
bug-guix <at> gnu.org
.
(Thu, 16 Dec 2021 07:02:02 GMT)
Full text and
rfc822 format available.
Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):
In my Guix system's /etc/passwd file, my user named "abc" has a
comment attached to it. The relevant line is:
abc:x:1000:998:Old
Comment:/home/jkf:/gnu/store/71yp1p06jy2j96bfdz43f4p6ncdym5a1-zsh-5.8/bin/zsh
Today the users section of my current config.scm looks like this:
(users (cons* (user-account
(name "abc")
(group "users")
(comment "New Comment")
(supplementary-groups '("wheel"
"netdev"
"audio"
"video"
"disk"
"cdrom"
"docker"
"libvirt"
"kvm"))
(shell #~(string-append #$zsh "/bin/zsh")))
%base-user-accounts))
After I apply this configuration with `guix system reconfigure', I
expect /etc/passwd to have been updated with "New Comment" in place of
"Old Comment". However, "Old Comment" remains.
Similarly, if I omit the `comment' field entirely, I expect my user
comment to be removed from /etc/passwd, since the default value of the
`comment' field is documented to be an empty string (manual 10.6).
Again, the old comment remains.
I am reporting this on a recent Guix version cev9c6c5, but have
noticed this issue for a year at least.
Information forwarded
to
bug-guix <at> gnu.org
:
bug#52539
; Package
guix
.
(Fri, 17 Dec 2021 04:03:01 GMT)
Full text and
rfc822 format available.
Message #8 received at 52539 <at> debbugs.gnu.org (full text, mbox):
Hi,
Am Donnerstag, dem 16.12.2021 um 07:00 +0000 schrieb Jacob First:
> In my Guix system's /etc/passwd file, my user named "abc" has a
> comment attached to it. The relevant line is:
>
> abc:x:1000:998:Old
> Comment:/home/jkf:/gnu/store/71yp1p06jy2j96bfdz43f4p6ncdym5a1-zsh-
> 5.8/bin/zsh
>
> Today the users section of my current config.scm looks like this:
>
> (users (cons* (user-account
> (name "abc")
> (group "users")
> (comment "New Comment")
> (supplementary-groups '("wheel"
> "netdev"
> "audio"
> "video"
> "disk"
> "cdrom"
> "docker"
> "libvirt"
> "kvm"))
> (shell #~(string-append #$zsh "/bin/zsh")))
> %base-user-accounts))
>
> After I apply this configuration with `guix system reconfigure', I
> expect /etc/passwd to have been updated with "New Comment" in place
> of "Old Comment". However, "Old Comment" remains.
>
> Similarly, if I omit the `comment' field entirely, I expect my user
> comment to be removed from /etc/passwd, since the default value of
> the `comment' field is documented to be an empty string (manual
> 10.6). Again, the old comment remains.
>
> I am reporting this on a recent Guix version cev9c6c5, but have
> noticed this issue for a year at least.
What if you were to temporarily change your login shell to let's say
bash? IIRC, Guix is quite lazy when it comes to updating these values,
but a change in the shell ought to get them revised. I think the
reason behind it is that it doesn't want to lock you out by messing
with the password field, but that's a little unrelated here.
Cheers
Information forwarded
to
bug-guix <at> gnu.org
:
bug#52539
; Package
guix
.
(Mon, 20 Dec 2021 23:47:01 GMT)
Full text and
rfc822 format available.
Message #11 received at 52539 <at> debbugs.gnu.org (full text, mbox):
Changing the shell indeed causes the comment to be updated.
If lazy update is the correct behavior, then the docs about user
accounts are a bit misleading:
"When booting or upon completion of guix system reconfigure, the
system ensures that only the user accounts and groups specified in the
operating-system declaration exist, and with the specified properties.
Thus, account or group creations or modifications made by directly
invoking commands such as useradd are lost upon reconfiguration or
reboot. This ensures that the system remains exactly as declared."
As a user it would be helpful to know from the docs that some of the
fields actually persist across reboots/reconfigurations.
Thanks for the workaround in any case!
On Fri, Dec 17, 2021 at 4:02 AM Liliana Marie Prikler
<liliana.prikler <at> gmail.com> wrote:
>
> Hi,
>
> Am Donnerstag, dem 16.12.2021 um 07:00 +0000 schrieb Jacob First:
> > In my Guix system's /etc/passwd file, my user named "abc" has a
> > comment attached to it. The relevant line is:
> >
> > abc:x:1000:998:Old
> > Comment:/home/jkf:/gnu/store/71yp1p06jy2j96bfdz43f4p6ncdym5a1-zsh-
> > 5.8/bin/zsh
> >
> > Today the users section of my current config.scm looks like this:
> >
> > (users (cons* (user-account
> > (name "abc")
> > (group "users")
> > (comment "New Comment")
> > (supplementary-groups '("wheel"
> > "netdev"
> > "audio"
> > "video"
> > "disk"
> > "cdrom"
> > "docker"
> > "libvirt"
> > "kvm"))
> > (shell #~(string-append #$zsh "/bin/zsh")))
> > %base-user-accounts))
> >
> > After I apply this configuration with `guix system reconfigure', I
> > expect /etc/passwd to have been updated with "New Comment" in place
> > of "Old Comment". However, "Old Comment" remains.
> >
> > Similarly, if I omit the `comment' field entirely, I expect my user
> > comment to be removed from /etc/passwd, since the default value of
> > the `comment' field is documented to be an empty string (manual
> > 10.6). Again, the old comment remains.
> >
> > I am reporting this on a recent Guix version cev9c6c5, but have
> > noticed this issue for a year at least.
> What if you were to temporarily change your login shell to let's say
> bash? IIRC, Guix is quite lazy when it comes to updating these values,
> but a change in the shell ought to get them revised. I think the
> reason behind it is that it doesn't want to lock you out by messing
> with the password field, but that's a little unrelated here.
>
> Cheers
>
Information forwarded
to
bug-guix <at> gnu.org
:
bug#52539
; Package
guix
.
(Wed, 22 Dec 2021 22:37:02 GMT)
Full text and
rfc822 format available.
Message #14 received at 52539 <at> debbugs.gnu.org (full text, mbox):
Hi,
Jacob First <jacob.first <at> member.fsf.org> skribis:
> After I apply this configuration with `guix system reconfigure', I
> expect /etc/passwd to have been updated with "New Comment" in place of
> "Old Comment". However, "Old Comment" remains.
This is on purpose, per these lines in ‘allocate-passwd’ in (gnu build
accounts):
(real-name (if previous
(password-entry-real-name previous)
real-name))
This wasn’t a firm decision, but I think the rationale here could be
that we’d rather allow users to change their name with ‘chfn’ to
something other than what the sysadmin chose, than enforce the
sysadmin’s choice. (Except that ‘chfn’ is currently not setuid by
default on Guix System, so the scenario above is unlikely.)
Thoughts?
Thanks,
Ludo’.
Information forwarded
to
bug-guix <at> gnu.org
:
bug#52539
; Package
guix
.
(Sun, 26 Dec 2021 21:57:02 GMT)
Full text and
rfc822 format available.
Message #17 received at 52539 <at> debbugs.gnu.org (full text, mbox):
Thanks for the explanation. This rationale sounds reasonable to me. If
someone with proper permission has modified a user account's comment,
it could be important to preserve that customization.
As I pointed out in an earlier message, I believe this behavior
contradicts the documentation of the `user-account' data type pretty
strongly. So it appears there is at least a "doc bug" here.
FWIW, I actually reported the present behavior as a bug not because of
the doc issue, but based on an assumption that "purely declarative"
implied the result of applying a given configuration should not depend
on any preexisting system state like a user comment. Sorry for not
making that clear: it seemed obvious, but I think it's actually a
misconception (even if user accounts are one of the few areas, AFAICT,
where the property doesn't hold true).
On Wed, Dec 22, 2021 at 10:36 PM Ludovic Courtès <ludo <at> gnu.org> wrote:
>
> Hi,
>
> Jacob First <jacob.first <at> member.fsf.org> skribis:
>
> > After I apply this configuration with `guix system reconfigure', I
> > expect /etc/passwd to have been updated with "New Comment" in place of
> > "Old Comment". However, "Old Comment" remains.
>
> This is on purpose, per these lines in ‘allocate-passwd’ in (gnu build
> accounts):
>
> (real-name (if previous
> (password-entry-real-name previous)
> real-name))
>
> This wasn’t a firm decision, but I think the rationale here could be
> that we’d rather allow users to change their name with ‘chfn’ to
> something other than what the sysadmin chose, than enforce the
> sysadmin’s choice. (Except that ‘chfn’ is currently not setuid by
> default on Guix System, so the scenario above is unlikely.)
>
> Thoughts?
>
> Thanks,
> Ludo’.
Reply sent
to
Ludovic Courtès <ludo <at> gnu.org>
:
You have taken responsibility.
(Fri, 31 Dec 2021 17:33:01 GMT)
Full text and
rfc822 format available.
Notification sent
to
Jacob First <jacob.first <at> member.fsf.org>
:
bug acknowledged by developer.
(Fri, 31 Dec 2021 17:33:01 GMT)
Full text and
rfc822 format available.
Message #22 received at 52539-done <at> debbugs.gnu.org (full text, mbox):
Hi Jacob,
Jacob First <jacob.first <at> member.fsf.org> skribis:
> Thanks for the explanation. This rationale sounds reasonable to me. If
> someone with proper permission has modified a user account's comment,
> it could be important to preserve that customization.
>
> As I pointed out in an earlier message, I believe this behavior
> contradicts the documentation of the `user-account' data type pretty
> strongly. So it appears there is at least a "doc bug" here.
>
> FWIW, I actually reported the present behavior as a bug not because of
> the doc issue, but based on an assumption that "purely declarative"
> implied the result of applying a given configuration should not depend
> on any preexisting system state like a user comment. Sorry for not
> making that clear: it seemed obvious, but I think it's actually a
> misconception (even if user accounts are one of the few areas, AFAICT,
> where the property doesn't hold true).
I clarified this and ensures ‘chfn’ is actually usable in commit
c76775263e56a10cc1b84d03a5827f42436afe40.
Thanks!
Ludo’.
Information forwarded
to
bug-guix <at> gnu.org
:
bug#52539
; Package
guix
.
(Sat, 01 Jan 2022 14:23:02 GMT)
Full text and
rfc822 format available.
Message #25 received at 52539 <at> debbugs.gnu.org (full text, mbox):
> I clarified this and ensures ‘chfn’ is actually usable in commit
> c76775263e56a10cc1b84d03a5827f42436afe40.
Oops, I hadn’t pushed; the right commit ID is
671e6a81804f264ddcdd6fe7579644404da079b8.
Ludo’.
bug archived.
Request was from
Debbugs Internal Request <help-debbugs <at> gnu.org>
to
internal_control <at> debbugs.gnu.org
.
(Sun, 30 Jan 2022 12:24:07 GMT)
Full text and
rfc822 format available.
This bug report was last modified 2 years and 86 days ago.
Previous Next
GNU bug tracking system
Copyright (C) 1999 Darren O. Benham,
1997,2003 nCipher Corporation Ltd,
1994-97 Ian Jackson.