Re: Shouldn't we be developing with TYPING_STRICTNESS=2?

[Dirk Herrmann]

On Thu, 13 Sep 2001, Rob Browning wrote:

> As a temporary fix in the tree I'm hacking on, I did this, but I
> wasn't sure if it was at all correct...
> 
>   if (SCM_EQ_P (SCM_PACK(SCM_SHOW_FILE_NAME), sym_base))

Well, it's not too bad :-)  It might be better, though, to have the
SCM_PACK be part of the SCM_SHOW_FILE_NAME macro as in the cases
below.  But, this solution is only preferable if SCM_SHOW_FILE_NAME is
really always treated as a SCM value.  Otherwise putting the SCM_PACK into
the macro would require to use SCM_UNPACK at other uses of
SCM_SHOW_FILE_NAME.

> I did this in eval.h (which was also a guess):
> 
>   -#define SCM_ENTER_FRAME_HDLR   (SCM)(scm_evaluator_trap_table[4].val)
>   -#define SCM_APPLY_FRAME_HDLR   (SCM)(scm_evaluator_trap_table[5].val)
>   -#define SCM_EXIT_FRAME_HDLR    (SCM)(scm_evaluator_trap_table[6].val)
>   +#define SCM_ENTER_FRAME_HDLR   (SCM_PACK(scm_evaluator_trap_table[4].val))
>   +#define SCM_APPLY_FRAME_HDLR   (SCM_PACK(scm_evaluator_trap_table[5].val))
>   +#define SCM_EXIT_FRAME_HDLR    (SCM_PACK(scm_evaluator_trap_table[6].val))

As described above, this is a nice fix if the macros are treated as SCM
values everywhere.  Still, it should not prevent us from looking into the
options implementation some time...

Best regards,
Dirk Herrmann