| 1 |
On 10/18/20 2:29 AM, Joonas Niilola wrote: |
| 2 |
> |
| 3 |
> |
| 4 |
> On 10/18/20 8:48 AM, Michał Górny wrote: |
| 5 |
>> On Sat, 2020-10-17 at 18:05 -0400, Aisha Tammy wrote: |
| 6 |
>>> This package provides the 'display-manager' startup script for |
| 7 |
>>> handling your chosen display manager, without being dependent on |
| 8 |
>>> Xorg server. |
| 9 |
> being dependent -> depending |
| 10 |
> |
| 11 |
>> |
| 12 |
>> Do you really think a rename for the sake of renaming justifies |
| 13 |
>> requiring all users to rewrite their configuration? While we're |
| 14 |
>> requiring the users to do that, wouldn't it be better to stop using |
| 15 |
>> the awful layout of 'one script to run them all', and switch to separate |
| 16 |
>> scripts for every DM? |
| 17 |
>> |
| 18 |
> This is exactly what I proposed in the previous RFC, systemd already works this way and is IMHO a lot clearer to use. |
| 19 |
> |
| 20 |
> -- juippis |
| 21 |
> |
| 22 |
|
| 23 |
This would need some more tinkering as OpenRC doesn't have a dedicated |
| 24 |
mechanism to control vt's while systemd controls the vts. |
| 25 |
|
| 26 |
IMO systemd is also very centralized, except that it hides it in its internals, |
| 27 |
as every services that wants to use a tty has to register with it. |
| 28 |
Nothing wrong with that, its a perfectly good method. |
| 29 |
|
| 30 |
I'm curious about why people are averse to a centralized script, it seems to |
| 31 |
be the simpler and cleaner option, atleast to me, presumably as different |
| 32 |
scripts for each DM *can* have different working mechanisms, and any fix applied |
| 33 |
to one of the DMs might have to be applied to others. |
| 34 |
|
| 35 |
Also the fact that the packages don't supply an OpenRC script on their own so |
| 36 |
every package would need an additional file from our end. |
| 37 |
|
| 38 |
Additionally we would need to able to say that each of them is going to conflict |
| 39 |
with the other. |
| 40 |
|
| 41 |
Aisha |
Archives