Re: [Chicken-users] chicken-wrap segfault; GPS egg

[Ivan Raikov]

Hello,

  Where is the source for the gps egg? I only see an empty directory
in the SVN repository.

   -Ivan
 
"Shawn Rutledge" <address@hidden> writes:

> I'm trying to get started with libgps/gpsd again, since the DBus
> support that gpsd provides is quite minimal (satellite "fixes" are
> broadcast, and that's it... nothing about which satellites are in
> view, and no methods to query anything.  And to me the "fix" messages
> look incomplete, too.)  So I'm attempting to use chicken-wrap to
> generate the wrapping for the big structs that libgps provides.  I had
> to simplify them to get that far... there is a big union/struct
> monstrosity right in the middle of one struct, which the easyffi
> parser can't handle, so I wrote a C program to print out the sizeof
> that union.  In the simplified version of gps.h that I wrote for
> easyffi, it's now just an array of 900 bytes instead, and fortunately
> I might not need any of the data from there anyway: it's GPS almanac
> data, apparently.
>
>  [proton][12:52:07 AM] make
>  csc -dynamic -X easyffi  -o gpsd.so gps.scm -lgps
>  *** Shell command terminated with exit status 1: /usr/bin/chicken
>  gps.scm -output-file gpsd.c -dynamic -feature chicken-compile-shared
>  -quiet -extend easyffi
>  make: *** [gpsd.so] Error 1
>
>  OK try it separately:
>
>  [proton][01:03:29 AM] /usr/bin/chicken gps.scm -output-file gpsd.c
>  -dynamic -feature chicken-compile-shared -extend easyffi
>  compiling `gps.scm' ...
>  Segmentation fault
>
> It's happening with Chicken 3.1.0 and 3.1.5, and I tried on two
> different machines.  There is output anyway (it created gpsd.c) but
> segfaulting is not good, and prevents me from generating and compiling
> the code in one csc call.  If I compile the output gpsd.c, I can
> generate an .so file, but it didn't load properly with (use gpsd).
>
> If I'm able to log in to svn again tonight, I will check in the code
> so far for the gps egg, but this is the current state of it.