glib: Problems with g_fprintf (fout, "\n");

[Stefan Andersen <zingo zingo org>]

Hello

I have some trouble with something simple as g_fprintf (fout, "\n");
the result is that on the fout-stream on my platform I get 
(in hex):"0x0a 0x00" If I patch _g_gnulib_vfprintf() to not fwrite
length+1 (se below) I dont get this, but I think there has to something
fishy somewere else, and I don't thing that my patch is the solution and
if I sometimes lose some characters, 

If anyone has any clue In what could cause this... 
 
Background:
I am in the middle of a port to a new target (Tao's Elate platform)
and I was living on a pre 2.2 branch until now were I updated to HEAD.
Doing this I got a bunch of new function like g_fprintf(), and a lot of
internal _g_gnulib_*() functions. 
When using this functions for the first time I got into trouble, the
first place were they were used was when compiling glib/gobjects and
executing the glib-genmarshal that was just compiled. 
The resulting output had a few 0x00 in it. 
Looking at the code this seems like impossible, hacking around a bit it
seams that if I patch _g_gnulib_vfprintf() my 0x00 goes away, but it
seems very strange that I should be the only one needing to to this so
my problem is probably somthing else since this would have been noticed
on other platform. But currently I have no clue what it could be, here
is my patched function:

int _g_gnulib_vfprintf (FILE *file, char const *format, va_list args)
{
  char *result;
  size_t length;

  result = vasnprintf (NULL, &length, format, args);
  if (result == NULL) 
    return -1;

	length-=1;  //<<-- Here is my patch 
	
  fwrite (result, 1, length + 1, file);
  free (result);
  
  return length;
}

-- 
/Stefan "Zingo" Andersen (www.zingo.org & www.vectrace.com)