/usr/include/glib-2.0/glib/gutils.h in 2.15.3 has:

#if defined (G_HAVE_INLINE) && defined (__GNUC__) && defined (__STRICT_ANSI__)
#  undef inline
#  define inline __inline__
#elif !defined (G_HAVE_INLINE)
#  undef inline
#  if defined (G_HAVE___INLINE__)
#    define inline __inline__
#  elif defined (G_HAVE___INLINE)
#    define inline __inline
#  else /* !inline && !__inline__ && !__inline */
#    define inline  /* don't inline, then */
#  endif
#endif
#ifdef G_IMPLEMENT_INLINES
#  define G_INLINE_FUNC
#  undef  G_CAN_INLINE
#elif defined (__GNUC__) 
#  define G_INLINE_FUNC static __inline __attribute__ ((unused))
#elif defined (G_CAN_INLINE) 
#  define G_INLINE_FUNC static inline
#else /* can't inline */
#  define G_INLINE_FUNC
#endif /* !G_INLINE_FUNC */

gnucash-2.2.3/src/engine/SplitP.h has:

G_INLINE_FUNC void mark_split (Split *s);

This evaluates (when building with gcc 4.1.2 on x86_64) to:

static __inline __attribute__ ((unused)) void mark_split (Split *s);

which causes Transaction.c, when compiled, to not actually have a prototype for mark_split. This generates a PC-relative relocation against the undefined symbol, causing linking to fail.

That being said.... why exactly is it declaring mark_split as inline in the header when it's being called from Transaction.c, which will have no idea what mark_split is to inline?