ntpd never achieves synchronized, even with PPS source; ntpdig KOD

2018-12-17 Thread Robin H. Johnson via devel
TL;DR: -- 1. ntpdig localhost always gets KOD 2. ntpsec never gets syncronized state: ntpwait never gives success 3. local clock ends up drifting :-( Background: --- I previously had another GPS puck, without PPS, and used it with classic ntpd; the puck eventually stopped working, and

Re: puthon curses on NetBSD

2018-12-17 Thread Hal Murray via devel
> And maybe "argparse" since it's not part of python2.6? I can add non-fatal > checks for these, similar to what I did for "curses". Seems like a good idea to me. If you don't find it, the warning message should point to INSTALL -- These are my opinions. I hate spam. _

Re: puthon curses on NetBSD

2018-12-17 Thread Hal Murray via devel
matthew.sel...@twosigma.com said: > I'd like to see the software list (not the package list) added to .../INSTALL > so that users on platforms not yet supported by .../buildprep can get a hint > on what software is needed. There is quite a bit along those lines in devel/hacking.txt. Maybe we s

Re: puthon curses on NetBSD

2018-12-17 Thread Matthew Selsky via devel
On Tue, Dec 11, 2018 at 03:01:06PM -0800, Gary E. Miller via devel wrote: > > I am unaware of any other Python libraries not in the base Python that > NTPsec uses. For some reason *BSD don't like curses. > > You know any other used Python packages not in the base Python? What about the "gps" mo

Re: puthon curses on NetBSD

2018-12-17 Thread Matthew Selsky via devel
On Mon, Dec 17, 2018 at 01:55:16PM -0800, Hal Murray via devel wrote: > > James Browning said: > > IIRC OpenSUSE does the same thing. I don't recall seeing a streak of red > > build fail indicators though. Possibly because I seem to remember adding it > > to the CI file. https://gitlab.com/NTPse

Re: puthon curses on NetBSD

2018-12-17 Thread Hal Murray via devel
James Browning said: > IIRC OpenSUSE does the same thing. I don't recall seeing a streak of red > build fail indicators though. Possibly because I seem to remember adding it > to the CI file. Thanks. Are you in a position to test buildprep on OpenSUSE and/or friends? It looks to me like it n

Re: Leftover junk

2018-12-17 Thread Hal Murray via devel
> That's because the version information actually comes from the ntp python > module. This is not specific to ntpmon. To fix this, we could interpolate the > version string into each script on install. That's easy, but there was some > resistance to that (interpolation on each script) in the past

Re: Leftover junk

2018-12-17 Thread James Browning via devel
On Mon, Dec 17, 2018, 12:48 PM Hal Murray via devel > I noticed another example. > > When we fixed install to skip ntpmon when it couldn't find curses, the > previous ntpmon was left installed. > > After installing curses, the old ntpmon --version would happily print out > the > current version te

Re: Leftover junk

2018-12-17 Thread Richard Laager via devel
On 12/17/18 2:48 PM, Hal Murray via devel wrote: > After installing curses, the old ntpmon --version would happily print out the > current version text. ;) That's because the version information actually comes from the ntp python module. This is not specific to ntpmon. To fix this, we could inte

Re: Leftover junk

2018-12-17 Thread Hal Murray via devel
I noticed another example. When we fixed install to skip ntpmon when it couldn't find curses, the previous ntpmon was left installed. After installing curses, the old ntpmon --version would happily print out the current version text. ;) -- These are my opinions. I hate spam. _