1 |
> > Oh, nice. What do you think of moving those helper implementations to |
2 |
> > tmpfiles-d.eclass and make |
3 |
> > |
4 |
> > systemd_dotmpfilesd() { dotmpfiles_d "$@"; } |
5 |
> > |
6 |
> > as William suggested? That way we would not impose blocker DEPENDs |
7 |
> > from systemd.eclass to users and might help adding {do,new}tmpfiles_d |
8 |
> > to later EAPIs |
9 |
> |
10 |
> These blockers do not hurt. They're against ancient systemd versions, |
11 |
> and they're going to be removed soon as well. |
12 |
|
13 |
Sounds good. |
14 |
|
15 |
> > > Also, maybe for some devs a easy make_tmpfiles_d() function will be |
16 |
> > > useful too... |
17 |
> > |
18 |
> > Yeah, it would be handier to have more freedom in file naming in |
19 |
> > ${FILESDIR} like 'proftpd/files/tmpfiles.d-r1.conf' and not just |
20 |
> > proftpd.conf (which is the name of server config file) |
21 |
> |
22 |
> I mean function where you wouldn't have to learn the syntax, like we |
23 |
> have make_desktop_entry(). |
24 |
|
25 |
Ah, I see. There is not much to abstract over, but tiny wrapper for |
26 |
|
27 |
echo "${entry}" > "${T}/${1}".conf || die; dotmpfiles_d "${T}/${1}".conf |
28 |
|
29 |
would make sense. |
30 |
|
31 |
-- |
32 |
|
33 |
Sergei |