gpsd-dev
[Top][All Lists]
Advanced

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

Re: [gpsd-dev] Baud barf problem on Pi3 SOLVED


From: Gary E. Miller
Subject: Re: [gpsd-dev] Baud barf problem on Pi3 SOLVED
Date: Thu, 5 May 2016 11:01:54 -0700

Yo Bernd!

On Thu, 5 May 2016 10:18:46 +0200
Bernd Zeimetz <address@hidden> wrote:

> Hi Gary!
> 
> On 05/04/2016 05:12 AM, Gary E. Miller wrote:
> >> oh, I did not say its not perfectly backward compatible, actually
> >> it is. But it is much more annoying to fight with shell snippets
> >> and files which are sourced from somewhere than to configure a
> >> systemd service. Also systemd comes with a lot of features one
> >> wants to have, which are not supported by these programs yet.  
> > 
> > Maybe so, but I know a lot of people that tried and failed at using
> > existing scripts or writing new service descriptions.  
> 
> Hmm, that brings me back to: read the documentation first. In Debian
> the sytemd guys are happy to help.

Except those that forked to Devuan.

> > Since it is easy, maybe you can fix, and document the gpsd
> > service?  
> 
> I still do not know what the bug is that needs fixing.
> So far I did not yet receive a single bug report in Debian.

Since Debian does not support any recent gpsd, we find no use in sending
bug reports.  Recent gpsd should speak for itself.

> And to be
> hones, I'm not keen on figuring out what the differences between
> Raspbian and Debian are to see what needs to be changed in Debian to
> get the issues fixed in Raspbian.

Getting Raspbian help is also a problem for us, like the known broken
Raspbian gpsd startup stripts in Wheezy.

> Afaik it works just fine for the
> Fedora people and at least in testing/unstable (that is what I test
> and use - and what should be stable at some point) I did not run into
> problems.

Absence of evidence is not evidence of absence.

> > And my personal pet peave; the binary syslog.  That is INSANE!  
> 
> Oh, you can still read the normal syslog. But the binary one gives you
> in-line forward secure sealing, so you are able to detect if somebody
> tried to modify the log files - which is the reason why I actually
> like this insanity. And the binary logs are much easier to
> seek/filter than the normal syslog.

I'm all for improving the syslogs, but the world is going from binary to 
printable formats, this is retrograde.  Harder to debug, harder for crash
recovery, and demands task specific tools, none of that is the UNIX way.

I have found many a bug by grabbing ascii fragments off a
crashing/crashed host/disk.  I can't imagine how much harder to grep the
raw disk sectors for binary log data...

> > Oh, I was expecting a web page, that is just code search results.
> > You can get a zillion hits on that by one commit.  Then your name
> > appears in all further revisions for all time.  
> 
> Oh, the Debian one is a different kind of code search, it takes the
> source from packages in Debian - you could click on the
> group-by-package button... Also we hopefully do not have code copies
> in Debian, if so, it would be a policy violation....

Either way, useless data for this purpose.  The problem is systemd itself.

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

Attachment: pgpRR_BhmnRf9.pgp
Description: OpenPGP digital signature


reply via email to

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