Re: [gpsd-dev] Warnings from 32 bit FreeBSD

[Gary E. Miller]

Yo Fred!

On Thu, 13 Jun 2019 17:16:33 -0700 (PDT)
Fred Wright <address@hidden> wrote:

> On Thu, 13 Jun 2019, Gary E. Miller wrote:
> 
> > I think I just fixed about half your warnings.  Please test.
> >
> > Very odd that your osX behaves so different than mine...  
> 
> Well, for any OS version, there may be more than one Xcode version
> that can be installed, and a given Xcode version sometimes includes
> more than one toolchain version.  In general, the way I set up the
> VMs is to have the latest Xcode (with command-line tools) for the
> given OS, and whatever toolchain it gets by default via command-line
> invocation.

And yet, my XCode is much newer than yours...

And your Ubuntu out of date:

MacLinux: Welcome to Ubuntu 14.04.6 LTS (GNU/Linux 4.4.0-143-generic x86_64)
MacLinux: 79 packages can be updated.
MacLinux: 64 updates are security updates.

And even expired:

MacLinux: Your Hardware Enablement Stack (HWE) is supported until April 2019.

> This time I just attached the whole output of the run, instead of 
> selectively copying and pasting.

Good plan.  Best, and easiest.

But the config output is missing.  That is usually pretty helpful.

Please include that next time.

Where is the compiler line?  Most of the remeing problems seem to be
when pkg_config puts things on the cc command line.

So much is missing that I can not see what is happening...

> 1) The script supplies the -s option, to exclude most of the
> uninteresting chatter while still showing warnings and errors.
> That's why I tweaked SConstruct to avoid suppressing the "This system
> is..." message in -s mode.

One mans dreck is another's gold.  I prefer my Matrix unfiltered...

Can you rerun without that?  W/o the config.log, gcc command line,
and the complete warings I am lost...

> 2) Some compilers emit a huge pile of warnings over some structure 
> initialization issue.  That's been a known problem for ages, but
> there are so many of these messages that I included a grep filter to
> get rid of them.

Which is why I never knew about it, so it does not get fixed.

AFAIK, these do not happen on my osX and Gentoo.

>  But it doesn't filter out messages on preceding
> lines specifying the error location, so you may sometimes see stuff
> like:
> 
>       MacSLS: rtcm3_json.c: In function 'json_rtcm3_read':
>       MacSLS: rtcm2_json.c: In function 'json_rtcm2_read':

Thus a new regression...

> Those should just be ignored.

May I suggest fixing instead of sweeping under the rug?

> These runs are build-only, since the regression tests aren't needed
> to chase these sorts of problems.

For now.

Several of these are not gpsd related:

This is a problem with your Python conflicting with pkg_config:

MacFed: /usr/include/python2.7/pyconfig-64.h:1221:0: warning: "_XOPEN_SOURCE" 
redefined
MacFed:  #define _XOPEN_SOURCE 600

This is because you did not build all the drivers:

MacML: /opt/local/bin/ranlib: file: libgpsd.a(driver_nmea2000.o) has no symbols

These I already fixed:

MacNet: driver_nmea0183.c: In function 'merge_ddmmyy':
MacNet: driver_nmea0183.c:151:9: warning: array subscript has type 'char'

The only problems I see:

Openbsd and your old Darwin, not my new Darwin:

MacO32: gps2udp.c:142: warning: implicit declaration of function 'strsep'

This in your CentOS and Hal reports in NetBSD:

MacCent: driver_nmea2000.c:1549:18: error: storage size of 'ifr' isn't known
MacCent:      struct ifreq ifr;

So, good progress.  But still missing the info to fix these last ones.

RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
        address@hidden  Tel:+1 541 382 8588

            Veritas liberabit vos. -- Quid est veritas?
    "If you can’t measure it, you can’t improve it." - Lord Kelvin

pgpzHWYyvfOzL.pgp
Description: OpenPGP digital signature