Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

On 05/13/2016 05:58 PM, Mike wrote: > Wow! I must have been off in la la land when I wrote that... nonsense. > > I just found that in packaging/[deb|rpm] there are basic default/gpsd > and sysconfig/gpsd. exactly - they are supposed to work with the init script and the systemd service. Not a

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

On 05/10/2016 09:06 PM, Mike wrote: On 05/10/2016 08:46 PM, Gary E. Miller wrote: Yo Mike! On Tue, 10 May 2016 20:34:35 -0400 Mike wrote: Then the systemd distros have something to start with. I will preface this with I am not using a hat. My Rpi is a model B. The gps module is attached t

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

On 05/10/2016 08:46 PM, Gary E. Miller wrote: Yo Mike! On Tue, 10 May 2016 20:34:35 -0400 Mike wrote: Then the systemd distros have something to start with. I will preface this with I am not using a hat. My Rpi is a model B. The gps module is attached to tx, rtx, and gpio18. Almost there.

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Mike! On Tue, 10 May 2016 20:34:35 -0400 Mike wrote: > > Then the systemd distros have something to start with. > I will preface this with I am not using a hat. My Rpi is a model B. > The gps module is attached to tx, rtx, and gpio18. Almost there. Hmm, I see there is already a system

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

On 05/10/2016 03:09 PM, Gary E. Miller wrote: Yo Mike! On Tue, 10 May 2016 14:40:56 -0400 Mike wrote: Currently I'm running gpsd from my home directory as I've been experimenting with some build options as a learning adventure. cat /usr/lib/systemd/system/gpsd.service Cool. If you can figu

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Mike! On Tue, 10 May 2016 14:40:56 -0400 Mike wrote: > Currently I'm running gpsd from my home directory as I've been > experimenting with some build options as a learning adventure. > > cat /usr/lib/systemd/system/gpsd.service Cool. If you can figure it out, and document it, send it to g

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

On 05/10/2016 10:42 AM, Eric S. Raymond wrote: Bernd Zeimetz : The sane way would be to use systemd to ensure that they keep running. If systemd eats ressources on your machine, maybe deactivating the journal will help. Do you mean that systemd does automatic restart of service daemons? If so,

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Eric S. Raymond : > I'm dealing with a lot of movomh parts here. Tweaking /etc/rc.local is a > solution I know will work. That's "moving" parts. I need some capsaicin... -- http://www.catb.org/~esr/";>Eric S. Raymond ___ devel mailing

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Bernd Zeimetz : > The sane way would be to use systemd to ensure that they keep running. > If systemd eats ressources on your machine, maybe deactivating the > journal will help. Do you mean that systemd does automatic restart of service daemons? If so, I agree that might be worthwhile. I would

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Gary E. Miller : > But, you still need to tell the user how to remove the distros gpsd and > ntpd service files. Already handled. -- http://www.catb.org/~esr/";>Eric S. Raymond signature.asc Description: Digital signature ___ devel mai

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

> On May 9, 2016, at 7:23 PM, Oliver Jowett wrote: > > init.d scripts are usually conffiles so only get removed on purge, not on > simple package removal. Traditionally the first thing they do is test that > the corresponding binary is actually Really? Init.d scripts are configured files and

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Frank! On Mon, 9 May 2016 19:22:23 -0400 Frank Nicholas wrote: > > On May 9, 2016, at 7:18 PM, Gary E. Miller wrote: > > > > Almost good enough. A lot of people might install gpsd, then > > forget they did, or their lab partner might have installed it. Or > > a new update of their distro

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

> On May 9, 2016, at 7:18 PM, Gary E. Miller wrote: > > Almost good enough. A lot of people might install gpsd, then forget they > did, or their lab partner might have installed it. Or a new update of > their distro may skip it in as a hidden dependency. Understood. Then a troubleshooting se

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Frank! On Mon, 9 May 2016 19:11:56 -0400 Frank Nicholas wrote: > The reason I suggested it was part of the package & removed by > removing the package is because I Eric’s How To has steps for > removing Debian NTP. (GPSd is not installed by default.) Fair enough, but Eric wants the Howto to

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

> On May 9, 2016, at 7:03 PM, Gary E. Miller wrote: > > We see a lot of users that do not remove the stock gpsd and ntpd when they > install gpsd and ntpd from source. That way they can compage the old and > the new before going all in to source compiled versions. Or they are > just too lazy..

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Frank! On Mon, 9 May 2016 18:58:45 -0400 Frank Nicholas wrote: > > On May 9, 2016, at 6:56 PM, Gary E. Miller wrote: > > > > But, you still need to tell the user how to remove the distros gpsd > > and ntpd service files. > > Aren’t they part of the Debian package (I hope so - you wouldn’

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

> On May 9, 2016, at 6:56 PM, Gary E. Miller wrote: > > But, you still need to tell the user how to remove the distros gpsd and > ntpd service files. Aren’t they part of the Debian package (I hope so - you wouldn’t want them there if there were no application to service requests)? If so, they

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Eric! On Mon, 9 May 2016 18:30:03 -0400 "Eric S. Raymond" wrote: > Bernd Zeimetz : > > http://unix.stackexchange.com/questions/233468/how-does-systemd-use-etc-init-d-scripts > > > > the first reply is a nice summary of the init script/systemd > > interaction. links to the real docs are inclu

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Bernd Zeimetz : > http://unix.stackexchange.com/questions/233468/how-does-systemd-use-etc-init-d-scripts > > the first reply is a nice summary of the init script/systemd interaction. > links to the real docs are included... Thanks, that was useful. I think I'm going to end-run potential problems

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Frank! On Fri, 6 May 2016 17:52:22 -0400 Frank Nicholas wrote: > dev2-rpi ~ # diff gpsd.sve/ppsthread.c gpsd/ppsthread.c I'm not gonna waste cycles on strange ways to git pull. > What’s the best/correct way to update the local ~/gpsd, for this > testing? One time only: ~ # rm gpsd -r gpsd

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

> On May 6, 2016, at 1:38 PM, Gary E. Miller wrote: > So if you are trully on git head you should see this in your source: > > dagwood gpsd # fgrep 'accuracy will' *c > ppsthread.c: "KPPS:%s kernel PPS unavailable, PPS accuracy will > suffer\n", > ppsthread.c: "KPPS:%s no

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

fr...@nicholasfamilycentral.com said: > Would refclockd be a part of NTP? If so, I’d be ok with that. I just have > no need for GPSd on a pure time server. I use "refclockd" as a tag to describe a great cleanup. I haven't seen a serious proposal so there aren't any details to discuss. Mov

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Frank! On Fri, 6 May 2016 13:22:36 -0400 Frank wrote: > > On May 6, 2016, at 12:01 PM, Eric S. Raymond > > wrote: > > > > Do a "git pull". If it tells you its getting commits rather than > > barfing, we're good. (I just modified tod/NEWS.) > > This morning I blew away “~/gpsd" and did a

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Eric! On Thu, 5 May 2016 23:05:28 -0400 "Eric S. Raymond" wrote: > Gary E. Miller : > > The problem is 98% of UNIX users are afraid to compile anything. > > They want to know how to 'apt-get' a binary. How do we get distros > > to do a gpsd-max and a gpsd-lite, and how does the user know the

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Gary E. Miller : > The problem is 98% of UNIX users are afraid to compile anything. They > want to know how to 'apt-get' a binary. How do we get distros to do > a gpsd-max and a gpsd-lite, and how does the user know the difference? This seems like a non-problem to me. The people picky enough to

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Eric! On Thu, 5 May 2016 22:25:58 -0400 "Eric S. Raymond" wrote: > Gary E. Miller : > > To avoid confusion we > > should refer to this as gpsd-lite, or, I'm open to a better name. > > People hear just 'gpsd' and they think of the gpsd-max. > >

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Gary E. Miller : > To avoid confusion we > should refer to this as gpsd-lite, or, I'm open to a better name. People > hear just 'gpsd' and they think of the gpsd-max. The Linux kernel doesn't get a different name when they leave out the modular driver

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Norton! On Thu, 5 May 2016 21:24:34 -0400 Norton Allen wrote: > I know SHM is not portable to QNX, which has shm_open() POSIX 1003.1. A cut/paste of the SHM refclock would be easy. If NTPsec did it I would add it to gpsd. > I am still hoping to work on getting gpsd+PPS-SHM working, but hav

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

On Thu, May 5, 2016 at 9:24 PM, Norton Allen wrote: > On 5/5/2016 5:17 PM, Gary E. Miller wrote: > >> On Thu, 5 May 2016 16:48:26 -0400 >> Frank Nicholas wrote: >> >> > >communicates with ntpd via SHM. It's not clear why refclock 20 > >should survive thaat transition when gpsd is already b

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Frank! On Thu, 5 May 2016 20:13:53 -0400 Frank Nicholas wrote: > > NMEA sentences are vaiable length. So you want to use the first one > > to cut jitter on the NMEA time. In practice you have the PPS, > > 10,000 times more accurate, so never a problem unless you lose > > PPS. > > And on

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

> On May 5, 2016, at 8:07 PM, Gary E. Miller wrote: > > Yo Frank! > > On Thu, 5 May 2016 19:49:28 -0400 > Frank Nicholas wrote: > >> We want to use PGGA sentence because it appears first, closest to >> the start of the second in each cycle. >> >> You have mode set to 17, using message GPRMC.

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Frank! On Thu, 5 May 2016 19:49:28 -0400 Frank Nicholas wrote: > We want to use PGGA sentence because it appears first, closest to > the start of the second in each cycle. > > You have mode set to 17, using message GPRMC. Can someone who knows > more than me, explain to me if it really matt

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Eric! On Thu, 5 May 2016 18:30:25 -0400 "Eric S. Raymond" wrote: > Gary E. Miller : > > Or, maybe looking at in the inverse manner, have a gpsd that can be > > stripped down to the bare minimum for ntpd use. > > We can already do that. Try building GPSD with just the NMEA0183 > driver and

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Gary E. Miller : > Or, maybe looking at in the inverse manner, have a gpsd that can be > stripped down to the bare minimum for ntpd use. We can already do that. Try building GPSD with just the NMEA0183 driver and SHM and a fixed baud rate sometime - socket_export turned off so it doesn't even spe

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

> On May 5, 2016, at 5:59 PM, Frank Nicholas > wrote: > > >> On May 5, 2016, at 5:54 PM, Gary E. Miller > > wrote: >> >> Yo Frank! >> >> On Thu, 5 May 2016 17:41:51 -0400 >> Frank Nicholas > > wrote: >> >>> The above does not s

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Frank Nicholas : > > The hardware management we keep should go into a separate refclockd that > > Would refclockd be a part of NTP? Yes. >If so, I’d be ok with that. I just have no need for GPSd on a pure time >server. That implies a separate question, which is: should both refclockd and gps

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

> On May 5, 2016, at 5:54 PM, Gary E. Miller wrote: > > Yo Frank! > > On Thu, 5 May 2016 17:41:51 -0400 > Frank Nicholas wrote: > >> The above does not show PPS. I wouldn’t expect it to - how does >> `gpsmon` know which device to use for PPS? > > very recent gpsd is special cased to add /de

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Frank! On Thu, 5 May 2016 17:41:51 -0400 Frank Nicholas wrote: > The above does not show PPS. I wouldn’t expect it to - how does > `gpsmon` know which device to use for PPS? very recent gpsd is special cased to add /dev/pps0 to its device when it sees /dev/ttyAMA0. So consider it lightly t

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Frank! On Thu, 5 May 2016 16:48:26 -0400 Frank Nicholas wrote: > > communicates with ntpd via SHM. It's not clear why refclock 20 > > should survive thaat transition when gpsd is already better at > > adapting to weird sentence inventories than it is. > > How portable is “SHM”? Mostly po

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

> On May 5, 2016, at 4:04 PM, Eric S. Raymond wrote: > > The hardware management we keep should go into a separate refclockd that Would refclockd be a part of NTP? If so, I’d be ok with that. I just have no need for GPSd on a pure time server. > communicates with ntpd via SHM. It's not cle

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Eric! On Thu, 5 May 2016 16:05:54 -0400 "Eric S. Raymond" wrote: > > It appends this at the bottom: > > > > # Generated by dhcpcd from lxcbr0.dhcp > > server 204.17.205.8 > > server 204.17.205.1 > > server 204.17.205.7 > > # End of dhcpcd from lxcbr0.dhcp > > Bletch. That is totally inap

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Eric! On Thu, 5 May 2016 16:04:26 -0400 "Eric S. Raymond" wrote: > I'm familiar with all these arguments. But sometimes the right > short-term solution is opposite to the best thing in the long term. I think we are mostly in agreement. Short term we need #20, long term it is cruft. > It's

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

fr...@nicholasfamilycentral.com said: > dhcpcd will merge DHCP supplied NTP servers at the > bottom of a ntp.conf. This has not caused me any issues (on Gentoo). NTP > generally does the right things and uses all the NTP > servers in the config & will decide the local server, with PPS is the > b

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Gary E. Miller : > Yo Eric! > > On Thu, 5 May 2016 15:45:09 -0400 > "Eric S. Raymond" wrote: > > > Frank : > > > dhcpcd will “merge” DHCP supplied NTP servers at the bottom of a > > > ntp.conf. This has not caused me any issues (on Gentoo). NTP > > > generally does the “right” things and uses

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Frank : > I agree 100% - I’d like to see clock type 20 stay in NTPSec. > K.I.S.S when a non-moving or low power or SoC system needs accurate > time and has no desire to care about location or motion. I'm familiar with all these arguments. But sometimes the right short-term solution is opposite to

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Eric! On Thu, 5 May 2016 15:45:09 -0400 "Eric S. Raymond" wrote: > Frank : > > dhcpcd will “merge” DHCP supplied NTP servers at the bottom of a > > ntp.conf. This has not caused me any issues (on Gentoo). NTP > > generally does the “right” things and uses all the NTP servers in > > the conf

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Frank : > Eric’s already fixed this. The section: > > # Disable Bluetooth so serial-tty speed is no longer tied to CPU speed > dtoverlay=pi3-miniuart-bt-overlay > force_turbo=1 > --

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Frank! On Thu, 5 May 2016 15:23:15 -0400 Frank wrote: > dhcpcd will “merge” DHCP supplied NTP servers at the bottom of a > ntp.conf. Easy to fix: dhcpcd -C ntp.conf This is in my Gentoo /etc/conf.d/net: config_wlan1="dhcp" dhcpcd_wlan1="-h dagwood1 -C resolv.conf -C ntp.conf" RGDS

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

> On May 5, 2016, at 1:36 PM, Gary E. Miller wrote: > > As much as I prefer, and defend, the gpsd/ntpd combo, I also understand > there is resistance from a lot of people that want to keep things > really simple. Strong resistance. Based on good reasons. > > And I have to say the current gpsd

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Frank! On Thu, 5 May 2016 13:14:34 -0400 Frank wrote: > See this page: https://www.ntpsec.org/removal-plan.html > > 2. Duplicates capabilities of GPSD, which specializes in GPSes. > Types 20, 29, 30, 31. . > . > Thus, we plan to remove the followi

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Frank! On Thu, 5 May 2016 13:17:57 -0400 Frank wrote: > > If you can explain how to do that to Eric, so he can explain it to > > us, that would be great. > > > > Saving power is a big deal in mobile applications. > > > > > The “force_turbo=1” was part of the solution, if you wanted to *

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

> On May 5, 2016, at 1:14 PM, Gary E. Miller wrote: > > Yo Frank! >> Nope - that only applies on the Raspberry Pi 3, with the BT using the >> ttyAM0 lines. If you remap the serial back, how the previous PI’s >> are, then the serial speed is not tied to the CPU clock. > > If you can explain how

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

> On May 5, 2016, at 1:08 PM, Gary E. Miller wrote: > > As of last night, here are the NTPsec refclocks: > > IDDescription > ~~~~~ > 1 Undisciplined Local Clock > 4 Spectracom WWVB/GPS Receivers > 5 TrueTime GPS/GOES/OMEGA Receivers > 6 IRIG Audio Decoder > 7

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Frank! On Thu, 5 May 2016 13:11:05 -0400 Frank wrote: > > On May 5, 2016, at 1:08 PM, Gary E. Miller wrote: > > > >> “force_turbo=1” forces the Pi to run at maximum clock speed all the > >> time. This generates heat & increased power consumption. If > >> constant/max clock speed is desi

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

> On May 5, 2016, at 1:08 PM, Gary E. Miller wrote: > >> “force_turbo=1” forces the Pi to run at maximum clock speed all the >> time. This generates heat & increased power consumption. If >> constant/max clock speed is desired (for NTP constancy?), keep this. >> If power/heat are a concern, re

Re: [gpsd-dev] New 0.5 draft of the SemPiTernal HOWTO

Yo Frank! On Thu, 5 May 2016 12:51:04 -0400 Frank wrote: > For root commands, instead of sudo bash, maybe just preface sudo > before each command - this is how most Pi/Distro’s I’ve seen do it Yeah, and it is a PITA. Much more to type, lulls the user into thinking he is improving ecurity when