[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [gpsd-dev] Refactor the way NTP shared memory segments are addressed
From: |
Eric S. Raymond |
Subject: |
Re: [gpsd-dev] Refactor the way NTP shared memory segments are addressed. |
Date: |
Wed, 14 Jan 2015 17:15:46 -0500 |
User-agent: |
Mutt/1.5.23 (2014-03-12) |
Hal Murray <address@hidden>:
>
> address@hidden said:
> > That's very odd. Are my eyes deceiving me or are those sentences identical
> > by pairs except for checksums?
>
> No, there is a difference. For example:
> > -$GPZDA,090614.00,11,06,2005,00,00*6D
> > +$GPZDA,080614.00,11,06,2005,00,00*6C
>
> It's a 9 vs 8 in 090614.00 vs 080614.00
>
> >> Compiles and checks for me on Debian and Ubuntu.
> > So the test failure is specific to Fedora?
>
> I haven't seen that error on a non-Fedora box.
> I get the same error on multiple runs.
>
> I don't get that error on a really old Fedora box.
Oh, shit on a stick. I know what *that* means.
It means a bug in recent Fedora's setting of either timezone or the DST flag
has been unmasked by my recent attempt to replace the use of timegm(3), which
is a glibc extension.
I have reverted that attempt. If you pull head and rebuild the code should
scons check of Fedora 21 correctly.
This reversion also solves a BSD port problem (some BSDs don't have the timezone
global) but it means we're stuck to glibc. Dammit.
--
<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>
- Re: [gpsd-dev] Refactor the way NTP shared memory segments are addressed., (continued)
Re: [gpsd-dev] Refactor the way NTP shared memory segments are addressed., Eric S. Raymond, 2015/01/13
Re: [gpsd-dev] Refactor the way NTP shared memory segments are addressed., Hal Murray, 2015/01/14