[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Installing 3.20 gpsd from source on Raspberry Pi cards old and new.

From: Paul Theodoropoulos
Subject: Re: Installing 3.20 gpsd from source on Raspberry Pi cards old and new.
Date: Tue, 9 Jun 2020 12:38:59 -0700
User-agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:68.0) Gecko/20100101 Thunderbird/68.9.0

On 6/9/2020 11:42 AM, Gary E. Miller wrote:
Yo Paul!

On Tue, 9 Jun 2020 11:25:17 -0700
Paul Theodoropoulos via <> wrote:

On 6/9/2020 10:23 AM, Gary E. Miller wrote:
I've been running gpsd/ntpd under systemd for years now. It's a
_different_ means of managing startup/running of daemons. If one
considers 'learning a new way of doing something' as equivalent to 'a
lot of complications for negative benefits', then I guess I can
understand your perspective, but it's a false equivalence.
50 lines vs 2 lines is not just different, it is over complicated and

??  50 lines vs 2 lines? What, specifically, are you talking about? You have a two line system V init file for gpsd?

Here's the systemd gpsd.service file that is distributed with gpsd:

Description=GPS (Global Positioning System) Daemon
# Needed with chrony SOCK refclock

ExecStart=/usr/local/sbin/gpsd $GPSD_OPTIONS $OPTIONS $DEVICES


That's longer than what I use, since I don't need the chronyd or gpsd.socket stuff. But ignoring that, it's only thirteen lines. The basic system V init.d startup file included (in packaging/rpm/gpsd.init), even after stripping comments and blank lines, is 55 lines. The example Debian init.d startup file (in packaging/deb/etc_init.d_gpsd) is 110 lines after stripping extraneous.

So, I think you have it backwards.

By default, the distros gpsd scripts only start gpsd after the first
client connects.  Then wonder why their ntpd gets no time from

I could could go on, but that is a show stopper, so why bother.
As Mr. Taylor clearly stated, "I've now managed to get gpsd 3.20 working and installed as a service
using only the files from the current GIT on a Raspberry Pi."

Distros were not in discussion. Therefore, that's not evidence of a shortcoming of systemd, notwithstanding that distros vs source aren't evidence of a shortcoming of systemd. You're describing an entirely different issue as if it has something inherent to do with systemd itself.

I will note that I agree that there is a lot to dislike about systemd
philosophically as well as in practice.
God, we agree it is junk.  Just Say No.

Please don't put words in my mouth, then suggest that's what I said. I have dozens and dozens of Debian servers that use systemd for daemon startup. Systemd never causes any problems. If it were junk, I'd be having daemons dying left and right, and have to put in place separate mechanisms to handle that. Your *personal dislike* of systemd is quite clear. But it's an opinion, not fact-based on measurable evidence. And that's dandy - do as you wish. But you mislead list readers by relentlessly suggesting that there is something *broken* with systemd, when the issues that users here have is - to my reading - ONLY in getting the initial installation correct.

You requested that I provide some additional documentation regarding systemd installation. I started the process, then was confronted with a non-sequitur challenge regarding the documentation's discussion of a problem with a long-past-EOL Debian version, followed by dead silence.  Too much pushback, here on the list and elsewhere, if the word 'systemd' is involved.

Assuming the packagers get the options right.  In the case of gpsd/ntpd,
they do not.
Again, not even under discussion.

"Raspberry Pi OS (previously called Raspbian) is our official
operating system for all models of the Raspberry Pi."

So the distro formerly known as Raspbian.
Nope. Please go a little deeper.
I specifically said: no mention on that page.  I have no desite to go deeper.
There's also no mention on that page of Blueberries or Sandals, similarly non-sequitur to the introduction of 'Ubuntu' into the discussion. You made a false statement.  Raspberry Pi OS is not Raspbian. It's fine if you want to be wrong. Just don't suggest that you're right, please.

Paul Theodoropoulos

reply via email to

[Prev in Thread] Current Thread [Next in Thread]