gpsd-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [gpsd-dev] ✘GPSD_API


From: Fred Wright
Subject: Re: [gpsd-dev] ✘GPSD_API
Date: Thu, 14 Apr 2016 15:52:27 -0700 (PDT)

On Thu, 14 Apr 2016, Gary E. Miller wrote:
> On Thu, 14 Apr 2016 12:20:20 -0700 (PDT)
> Fred Wright <address@hidden> wrote:
>
> > On Thu, 14 Apr 2016, Gary E. Miller wrote:
> >
> > > Thanks for the confirmation.  Now I am hitting my head on a wall as
> > > when I checn gpsdata_t my scons check fails.
> > >
> > > Processing test/daemon/passthrough.log
> > > gpsd:ERROR: shmget(0x47505345, 9512, 0666) for SHM export f ailed:
> > > Invalid argument
> > >
> > > Which usually means mismatched shmget() sizes, but I can't figure
> > > out why the test harness is using the old sizeof(shmexport_t) and
> > > not the new one.  Deleting all system gpsd libs and starting from a
> > > fresh git clone, plus the one cchange, is no help.
> >
> > Do you have an ntpd running with a SHM source enabled?  BTW, changing
> > the size without breaking NTP might be an issue.
>
> My use case is 'scons check' which uses shmget().  The SHM that
> scons check uses is not the same SHM that NTP uses.

In that case you might have to resort to strace. :-)

> > Also, if by "fresh git clone" you mean re-cloning the repo, that's
> > almost certainly unnecessary since "git clean -dxf" gives you
> > essentially the same result.
>
> Not really, I also delete all system gpsd components.  That has
> bitten me in the past.  So I go really clean.

Well, neither git clean nor git clone is going to remove anything from the
system areas.  You either have to do that manually or run "scons
uninstall" (which you should do *before* any other cleaning, in case
options affect what to remove).  But most testing doesn't even require
installing anything in the first place, so just not installing is the
easiest way to avoid that sort of trouble.

I think the test code *tries* to avoid being affected by system libraries,
but it's an easy thing to get wrong.

> And the git clean man page imprlies you need to double up on the -f,
> like this:  git clean -dxff.  To hard for me to remember, I just delete
> everything and I know I got it right.

It looks like that's only if you have subdirectories with their own .git
subdirectories, i.e., which look like repos of their own.  That shouldn't
happen here unless you go out of your way to set that up.

Fred Wright



reply via email to

[Prev in Thread] Current Thread [Next in Thread]