Re: [gpsd-dev] More on Fedora timezone(?) glitch

[Eric S. Raymond]

Hal Murray <address@hidden>:
> 
> It works on Fedora 21 x86_64
> It fails on Fedora 21 i686
> It fails on Fedora 19 i686
> It works on Fedora 18 i686
> 
> It works for me on Debian wheezy and Ubuntu 14.04.1 LTS.  Both are amd64.
> Does anybody have a 32 bit Debian or Ubuntu setup?
> 
> Works on FreeBSD, both 9.1/i386 and 10.1/amd64
> 10.1 is using clang rather than gcc.

For the rest of the list: It looks like the Fedora 21 failure is due to stale 
shared libraries.  There's some build chain issue with setting
LD_LIBRARY_PATH not quite doing the right thing.

> On NetBSD, I get a blizzard of errors.  It's not consistent from run to run.  
> I get similar errors from the 3.11 tar file.  I'll put this on the back 
> burner for now.  I'm not sure it ever worked, but it seems to be trying, and 
> some runs get only a few errors.

Yeah, this is the pty timing problem I've been trying to track down
for like *four years* now.  I believe it's not a GPSD bug, but
something deep in Python or the kernel.  The regression tests beat the
crap out of the pty layer and sometimes it revolts.

Hal, you were able to work around it it under FreeBSD like this:

# From Hal Murray on NetBSD 6.1.2 on an Intel(R) Celeron(R) CPU 2.80GHz
#  WRITE_PAD = 0.0 / CLOSE_DELAY = 0.4    Works, takes 688.69s real
#  WRITE_PAD = 0.0 / CLOSE_DELAY = 0.3    Fails tcp-torture.log, 677.53s real

You might want to try increasing the values under freebsd in gps/fake.py

> The 3.11 tar file gets:
>   Processing test/daemon/GPSmap-76S.log
>   gpsd:ERROR: shmget(1196446532, 31532, 0666) failed: Invalid argument
>   gpsd:ERROR: shared-segment creation failed,
> That continues, one per test.
> 
> That's on both NetBSD and Fedora.  

I don't know what's up with that last error.  Test with ntpshm off?
-- 
                <a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>