Bug#679751: Lintian now detect package pointing to /home

[Michael Biebl]

Am 18.09.2017 um 02:18 schrieb Michael Biebl:
> It's not as simple as that. You might still have running processes with
> that uid in which case usermod complains and exists.
> So to successfully run usermod you'd have to kill a processes running
> under that uid.

Fwiw, I ran into this issue when an upstream of mine switched runtimedir
from /var/run to /run and my postinst so far has created a system user
with --home /var/run/foo which I now wanted to change to --home /run/foo.
As the package starts a (socket activated) daemon, this needs some
dedicated code to disable socket activation, then stop the daemon, then
change the home directory via usermod, then re-enable socket activation.

Or I just don't try to anticipate all possible corner cases and simply
use the new home directory only on new installations.

Not sure what's better.


-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?



signature.asc
Description: OpenPGP digital signature

Bug#679751: Lintian now detect package pointing to /home

[Michael Biebl]

On Fri, 31 Jul 2015 13:16:24 +0200 Guillem Jover  wrote:
> Hi!
> 
> On Fri, 2015-07-31 at 11:34:20 +0200, Bastien ROUCARIES wrote:
> > On Fri, Jul 31, 2015 at 11:04 AM, Bastien ROUCARIES wrote:
> > > Lintian now detect script creating user pointing to /home.
> 
> > After a chat under #debian-qa it appear that canonical path for non
> > existant home dir is /nonexistant could be documented ?
> > 
> > Moreover how can a developper could correct a mistake in previous
> > package ? Rip the /home directory correct hte passd file ?
> 
> To fix a wrong home path you can do something like:
> 
> ,---
> # Fix the home directory name created by an old package.
> u_home=$(getent passwd USERNAME | cut -d: -f6)
> if [ "$u_home" = "OLDHOME" ] ; then
>   usermod --home NEWHOME USERNAME >/dev/null || true
> fi
> `---
> 

It's not as simple as that. You might still have running processes with
that uid in which case usermod complains and exists.
So to successfully run usermod you'd have to kill a processes running
under that uid.

-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?



signature.asc
Description: OpenPGP digital signature

Bug#679751: Lintian now detect package pointing to /home

[Colin Watson]

On Fri, Jul 31, 2015 at 11:34:20AM +0200, Bastien ROUCARIES wrote:
 After a chat under #debian-qa it appear that canonical path for non
 existant home dir is /nonexistant could be documented ?

If we do (and I'm not expressing any opinion on that here), we should
spell it correctly: /nonexistent.

-- 
Colin Watson   [cjwat...@debian.org]


-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Bug#679751: Lintian now detect package pointing to /home

[Julien Cristau]

On Fri, Jul 31, 2015 at 13:16:24 +0200, Guillem Jover wrote:

 Hi!
 
 On Fri, 2015-07-31 at 11:34:20 +0200, Bastien ROUCARIES wrote:
  On Fri, Jul 31, 2015 at 11:04 AM, Bastien ROUCARIES wrote:
   Lintian now detect script creating user pointing to /home.
 
  After a chat under #debian-qa it appear that canonical path for non
  existant home dir is /nonexistant could be documented ?
  
  Moreover how can a developper could correct a mistake in previous
  package ? Rip the /home directory correct hte passd file ?
 
 To fix a wrong home path you can do something like:
 
 ,---
 # Fix the home directory name created by an old package.
 u_home=$(getent passwd USERNAME | cut -d: -f6)
 if [ $u_home = OLDHOME ] ; then
   usermod --home NEWHOME USERNAME /dev/null || true
 fi
 `---
 
I think fixing that on upgrades is more dangerous than anything.  If the
dir exists, you'd need to move it to the new location, adding a bunch of
failure modes for what seems to me a rather minor benefit.

Cheers,
Julien


signature.asc
Description: Digital signature

Bug#679751: Lintian now detect package pointing to /home

[Bastien ROUCARIES]

On Fri, Jul 31, 2015 at 11:34 AM, Bastien ROUCARIES
roucaries.bast...@gmail.com wrote:
 On Fri, Jul 31, 2015 at 11:04 AM, Bastien ROUCARIES
 roucaries.bast...@gmail.com wrote:
 control: block -1 by 793613


 Hi,

 Lintian now detect script creating user pointing to /home.


 Could you also document that new added user should have an /home
 directory or /dev/null


 Ntp pass wrongfully also the --no-create-home flag, this should be
 documented as a not to do.

 With my lintian maint hat, I will like this issue get more guidance in policy

 Bastien
 After a chat under #debian-qa it appear that canonical path for non
 existant home dir is /nonexistant could be documented ?

 Moreover how can a developper could correct a mistake in previous
 package ? Rip the /home directory correct hte passd file ?

 Thanks
quoting debian-qa:
11:32:23) jcristau: not with --no-create-home it won't
(11:32:54) jcristau: but if it does exist, then not messing with it
further is probably safest anyway
(11:33:00) Myon: unless the crap was in stable, you don't need to
clean up everything
(11:33:30) Myon: an unused directory somewhere doesn't hurt, but any
postinst magic will probably at some point


-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Bug#679751: Lintian now detect package pointing to /home

[Bastien ROUCARIES]

control: block -1 by 793613


Hi,

Lintian now detect script creating user pointing to /home.


Could you also document that new added user should have an /home
directory or /dev/null


Ntp pass wrongfully also the --no-create-home flag, this should be
documented as a not to do.

With my lintian maint hat, I will like this issue get more guidance in policy

Bastien


-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Bug#679751: Lintian now detect package pointing to /home

[Bastien ROUCARIES]

On Fri, Jul 31, 2015 at 11:04 AM, Bastien ROUCARIES
roucaries.bast...@gmail.com wrote:
 control: block -1 by 793613


 Hi,

 Lintian now detect script creating user pointing to /home.


 Could you also document that new added user should have an /home
 directory or /dev/null


 Ntp pass wrongfully also the --no-create-home flag, this should be
 documented as a not to do.

 With my lintian maint hat, I will like this issue get more guidance in policy

 Bastien
After a chat under #debian-qa it appear that canonical path for non
existant home dir is /nonexistant could be documented ?

Moreover how can a developper could correct a mistake in previous
package ? Rip the /home directory correct hte passd file ?

Thanks


-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org

Bug#679751: Lintian now detect package pointing to /home

[Guillem Jover]

Hi!

On Fri, 2015-07-31 at 11:34:20 +0200, Bastien ROUCARIES wrote:
 On Fri, Jul 31, 2015 at 11:04 AM, Bastien ROUCARIES wrote:
  Lintian now detect script creating user pointing to /home.

 After a chat under #debian-qa it appear that canonical path for non
 existant home dir is /nonexistant could be documented ?
 
 Moreover how can a developper could correct a mistake in previous
 package ? Rip the /home directory correct hte passd file ?

To fix a wrong home path you can do something like:

,---
# Fix the home directory name created by an old package.
u_home=$(getent passwd USERNAME | cut -d: -f6)
if [ $u_home = OLDHOME ] ; then
  usermod --home NEWHOME USERNAME /dev/null || true
fi
`---

Thanks,
Guillem


-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org