| 1 |
On Wed, 2019-06-12 at 21:11 -0400, Michael Orlitzky wrote: |
| 2 |
> On 6/11/19 12:23 PM, Michał Górny wrote: |
| 3 |
> > |
| 4 |
> > +# @FUNCTION: user_get_nologin |
| 5 |
> > +# @INTERNAL |
| 6 |
> > +# @DESCRIPTION: |
| 7 |
> > +# Find an appropriate 'nologin' shell for the platform, and output |
| 8 |
> > +# its path. |
| 9 |
> > +user_get_nologin() { |
| 10 |
> |
| 11 |
> This isn't a great name for this function, because it doesn't have |
| 12 |
> anything to do with the user being added. How about |
| 13 |
> userland_get_nologin? Then it could take the userland as an argument... |
| 14 |
|
| 15 |
user_ is eclass name prefix. |
| 16 |
|
| 17 |
> |
| 18 |
> |
| 19 |
> > + case ${USERLAND} in |
| 20 |
> > + GNU) eshell="/bin/false" ;; |
| 21 |
> |
| 22 |
> to eliminate that dependency on the global variable. In other words, the |
| 23 |
> above could become "case $1..." to make the function self-contained. |
| 24 |
> |
| 25 |
> |
| 26 |
> |
| 27 |
> > + eshell=$(user_get_nologin) |
| 28 |
> |
| 29 |
> Then this would have to become |
| 30 |
> |
| 31 |
> eshell=$(userland_get_nologin "${USERLAND}") |
| 32 |
|
| 33 |
Do you have any real use for that? |
| 34 |
|
| 35 |
-- |
| 36 |
Best regards, |
| 37 |
Michał Górny |
Archives