Re: [gpsd-users] Building the master branch and the split24 option

[Roy Barkas]

The error is the same whether I build with pps=no or not.

The error in both cases first appears in testregress here:  

Processing test/daemon/GPSmap-76S.log...gpsd:ERROR: shmget(1196446532, 31616, 0666) failed: Invalid argument

gpsd:ERROR: shared-segment creation failed,

...(1.54 sec) done.

As i mentioned in the previous message - this only happens when building master, not 3.9.

I can't imagine what might be in my local environment that would cause the two versions to behave differently.

I'm building both versions as the same (linux) user, just from different sub-directories.

On Mon, Oct 21, 2013 at 10:10 PM, Eric S. Raymond <address@hidden> wrote:

Roy Barkas <address@hidden>:
> There was a commit on Sept 15th that added a "split24" option to the ?WATCH
> object and added a -s option to gpsdecode.
>
> I cloned the *master* branch, and then verified that the split 24 option
> was included in the source.
>
> When I run  scons && scons testregress I get many errors like:
>
> gpsd:ERROR: shared-segment creation failed,...(1.68 sec) done.
> Processing test/daemon/polarx2.log...gpsd:ERROR: shmget(1196446532, 31616,
> 0666) failed:
>
> If I run gpsd despite the build errors, an error message is displayed that
> says:
>
> gpsd:ERROR: shmget(1196446532, 31616, 0666) failed: Invalid argument
> gpsd:ERROR: shared-segment creation failed,
>
> I have seen some comment on these errors occurring during the build process
> but I haven't seen any mention of the run-time error.

Hm.  I think this is something in your local environment causing the
attachment to the NTPD shared-memory segment to fail. Try building with pps=no;
that should banish these errors.  And tell me if it does - I may want to
add a note to the test instructions.

I'll test and get back to you about the split24 issue.
--
                <a href="" href="http://www.catb.org/~esr/" target="_blank">http://www.catb.org/~esr/">Eric S. Raymond</a>