Re: Waypoint name with non-ascii incorrecto on Garmin

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Waypoint name with non-ascii incorrecto on Garmin

Göran Uddeborg
Switching to the development list since the topic becomes more
developer related.  I've taken a first look now.

Robert Lipe:
> It's kind of an ugly
> collision

I've seen worse.  Maybe I'm even responsible for some code that could
be considered worse. :-)

I started to try to run configure, and to add whatever it wanted that
I didn't already have.  I came upon an error message you might be
interested in:

    You must build your code with position independent code if Qt was built with -reduce-relocations.
    Compile your code with -fPIC (-fPIE is not enough).

This is from the qt5-qtbase-devel-5.5.0-15.fc23 version of
/usr/include/qt5/QtCore/qglobal.h.

Configure uses -fPIE, why this failed.  I have of course made the
obvious change in configure.in to fix this, but I thought I'd mention
it here as you might want to fix in the upstreams sources too.

And of course I had to look what "random stuff" you were checking!
:-)

(With that fix my compilation went through.  Next I'll see the binary
runs, but that's for another day.  It's getting too late.)

------------------------------------------------------------------------------
_______________________________________________
Gpsbabel-code mailing list  http://www.gpsbabel.org
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gpsbabel-code
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Waypoint name with non-ascii incorrecto on Garmin

Robert Lipe-4

    You must build your code with position independent code if Qt was built with -reduce-relocations.
    Compile your code with -fPIC (-fPIE is not enough).

This is from the qt5-qtbase-devel-5.5.0-15.fc23 version of
/usr/include/qt5/QtCore/qglobal.h.

Configure uses -fPIE, why this failed.  I have of course made the

The Linux distro situation is such a mess that I leave it to individuals or companies to try to work that all out.  
 
obvious change in configure.in to fix this, but I thought I'd mention
it here as you might want to fix in the upstreams sources too.

In GPSBabel-land, I *am* the upstream, IMO.  If Fedora wants PIE, then Fedora's tools should default to PIE or their configure should provide PIE or whatever.  You or Fedora can send me patches if they don't hose Ubuntu or FreeBSD or anyone else.

 
And of course I had to look what "random stuff" you were checking!
:-)

Funny.  That's been there nearly ten years.  The number of of people that have mentioned it is probably under five, but two have been this month.

My point, though it's a little better these days, is that autoconf runs so many completely ridiculous tests that approximately nobody (though on the rise!) reads or understands all that.  "Checking for 6 bit system with bzero, but not memset, but that has the arguments inexplicably reversed....(OK)."  Yay?   "Checking for <stdio.h>"  That header is part of SVID, is part of C89, and certainly part of C++ - and has been for over 20 years - why are you telling me this?  Parts of it I get - for example, we use it to check for libusb - but most of autoconf really is terrible.  See "autoconf insanity" for an example of an article I don't totally disagree with.

I could almost remove it from our project.  "Hey, that compile error you're getting not having libusb?  Well, install libusb!".  It's worth noting that we have three *.cc files that test HAVE_foo.  One of them should be moved to deprecated. (Delorme doesn't support Delorme protocol devices...) One of them could be deprecated, but could be replaced by a better Qt abstraction.  THe remaining one could be special cased for the only two remaining classes of operating systems: Windows and "UNIX enough".

I have no love of autoconf.  I have active contempt for m4....

RJL

------------------------------------------------------------------------------

_______________________________________________
Gpsbabel-code mailing list  http://www.gpsbabel.org
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gpsbabel-code
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Waypoint name with non-ascii incorrecto on Garmin

tsteven4-2
In reply to this post by Göran Uddeborg
Göran,

Thanks for the heads up on -fPIC.  I committed the obvious change here
https://github.com/gpsbabel/gpsbabel/commit/fdaf94fae19e1fa3d90286dff00fe583897e345a

Steve

On 8/27/2015 3:08 PM, Göran Uddeborg wrote:

> Switching to the development list since the topic becomes more
> developer related.  I've taken a first look now.
>
> Robert Lipe:
>> It's kind of an ugly
>> collision
> I've seen worse.  Maybe I'm even responsible for some code that could
> be considered worse. :-)
>
> I started to try to run configure, and to add whatever it wanted that
> I didn't already have.  I came upon an error message you might be
> interested in:
>
>      You must build your code with position independent code if Qt was built with -reduce-relocations.
>      Compile your code with -fPIC (-fPIE is not enough).
>
> This is from the qt5-qtbase-devel-5.5.0-15.fc23 version of
> /usr/include/qt5/QtCore/qglobal.h.
>
> Configure uses -fPIE, why this failed.  I have of course made the
> obvious change in configure.in to fix this, but I thought I'd mention
> it here as you might want to fix in the upstreams sources too.
>
> And of course I had to look what "random stuff" you were checking!
> :-)
>
> (With that fix my compilation went through.  Next I'll see the binary
> runs, but that's for another day.  It's getting too late.)
>
> ------------------------------------------------------------------------------
> _______________________________________________
> Gpsbabel-code mailing list  http://www.gpsbabel.org
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gpsbabel-code


------------------------------------------------------------------------------
_______________________________________________
Gpsbabel-code mailing list  http://www.gpsbabel.org
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gpsbabel-code
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Waypoint name with non-ascii incorrecto on Garmin

Göran Uddeborg
In reply to this post by Robert Lipe-4
Robert Lipe:
> In GPSBabel-land, I *am* the upstream, IMO.  If Fedora wants PIE,

Sorry if I was unclear, and maybe confused things by mentioning
Fedora.

I meant to say that I was forwarding the suggestion to you exactly
SINCE you are upstream.  Upstream from my local clone, that is.  An it
isn't Fedora that wants PIE, it is what was in the configure.in I got
when I cloned from GitHub.

>     And of course I had to look what "random stuff" you were checking!
>     :-)

> My point, though it's a little better these days, is that autoconf runs so many
> completely ridiculous tests that approximately nobody (though on the rise!)
> reads or understands all that.

When I saw the code, I figured that was the point. :-)  It reminded me
of

http://www.geek.com/games/gamestation-eula-collects-7500-souls-from-unsuspecting-customers-1194091/

------------------------------------------------------------------------------
_______________________________________________
Gpsbabel-code mailing list  http://www.gpsbabel.org
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gpsbabel-code
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Waypoint name with non-ascii incorrecto on Garmin

Göran Uddeborg
In reply to this post by tsteven4-2
tsteven4:
> I committed the obvious change here
> https://github.com/gpsbabel/gpsbabel/commit/fdaf94fae19e1fa3d90286dff00fe583897e345a
>
> Steve

Ah, I should have read all my mails before replying.  Then my reply to
Robert's mail is a bit superfluous.  Sorry!

------------------------------------------------------------------------------
_______________________________________________
Gpsbabel-code mailing list  http://www.gpsbabel.org
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gpsbabel-code
Loading...