Package: apt
Version: 1.6~beta1
Severity: minor
Please stop suggesting powermgmt-base: this is an obsolete, orphaned and
unmantained package which I last NMU'ed myself in 2014.
It provides the on_ac_power command which is only used on non-systemd
systems anyway.
--
ciao,
Marco
Acknowledgement sent
to Adam Borowski <[email protected]>:
Extra info received and forwarded to list. Copy sent to APT Development Team <[email protected]>.
(Wed, 28 Feb 2018 02:57:06 GMT) (full text, mbox, link).
> Please stop suggesting powermgmt-base: this is an obsolete, orphaned and
> unmantained package which I last NMU'ed myself in 2014.
No maintenance is needed there -- kernel interface hasn't changed in a long
long time. I just took a look at the package, and didn't find anything
worth the effort to improve: you can slightly optimize tests ([ -d dir ] &&
[ -d dir/subdir ] is redundant), bump standards version/etc, and that's it.
At the first glance, it is tempting to drop support for ancient interfaces
like APM or PMU, but RTFKing, I see that the only PMU driver
(drivers/macintosh/via-pmu.c) doesn't register itself in sysfs. That driver
is used only for Apple Powermacs, which are a platform not exactly likely to
see lively development, but are still supported by current unstable.
Thus, 99.99% machines will use the modern sysfs interface, while stragglers
keep working. Which can't be said about systemd, which doesn't support PMU.
> It provides the on_ac_power command which is only used on non-systemd
> systems anyway.
Newsflash: not everyone suffers from systemd -- either by choice or by
systemd not supporting people's hardware or use cases. One example are the
aforemented Apple Powermacs (ancient), another are both laptops I own, the
newer of which was released in April 2017.
powermgmt-base's on_ac_power knows about all interfaces provided by Linux,
systemd's equivalent does not. Thus, it's clearly better to use the former.
Meow!
--
⢀⣴⠾⠻⢶⣦⠀
⣾⠁⢠⠒⠀⣿⡁ A dumb species has no way to open a tuna can.
⢿⡄⠘⠷⠚⠋⠀ A smart species invents a can opener.
⠈⠳⣄⠀⠀⠀⠀ A master species delegates.
Acknowledgement sent
to Adam Borowski <[email protected]>:
Extra info received and forwarded to list. Copy sent to APT Development Team <[email protected]>.
(Fri, 16 Mar 2018 00:27:03 GMT) (full text, mbox, link).
Subject: consider using on_ac_power over systemd's handling
Date: Fri, 16 Mar 2018 01:23:18 +0100
Marco d'Itri wrote:
> Please stop suggesting powermgmt-base: this is an obsolete, orphaned and
> unmantained package which I last NMU'ed myself in 2014.
> It provides the on_ac_power command which is only used on non-systemd
> systems anyway.
I've adopted powermgmt-base and fixed some issues, with more work left to
do, but even the old version was strictly better than what systemd does.
And, it turns out, the differences are not just truly ancient machines
(Powerbooks and previous millenium's i386), but also modern tablets and
phones. On any of those that are charged via USB (that identifies as USB)
but the SoC could possibly support a non-smart connector ("ac"), even if
it's not physically wired on the machine, ConditionACPower will always think
you're on battery.
This particular problem can be fixed in systemd (simplified results of my
research so far: )
https://salsa.debian.org/debian/powermgmt-base/raw/master/power_supply.txt
as it's a straightforward extension of existing sysfs-reading code, but
looking at other scenarios that we'd also want fixed (whenever me or someone
more competent learns how), that becomes way more complex than anyone would
reasonably put into a low-level tool running as PID 1.
Thus, I'd suggest preferring on_ac_power even on systemd.
Meow!
--
⢀⣴⠾⠻⢶⣦⠀
⣾⠁⢠⠒⠀⣿⡁ A dumb species has no way to open a tuna can.
⢿⡄⠘⠷⠚⠋⠀ A smart species invents a can opener.
⠈⠳⣄⠀⠀⠀⠀ A master species delegates.