Re: [Chicken-users] C_truep( C_fixnump( w ))
On Feb 25, 2008, at 5:55 AM, felix winkelmann wrote:
On Mon, Feb 25, 2008 at 1:28 PM, Heinrich Taube <[EMAIL PROTECTED]>
wrote:
thank you both very much. just to be sure:
(1) should i be using a C_truep() around C_immediatep() like the
other
predicate tests or is this ok:
if ( C_immediatep(w) ) {
if ( C_truep(C_fixnump(w)) ) {
...
}
else if ( C_truep(C_blockp(w)) && C_truep( C_flonump
( w ) )) {
Yes, this is correct. It is obviously inconsistent, but
C_immediatep was
intended to be used from C and C_fixnump from compiled code.
Changing this will break old (compiled) code.
I didn't bother to check for immediate or block since
1) We assume it is a Chicken numeric value
2) An immediate value is an invalid address of a block/special item
3) So checking 1st for a fixnum & then for flonum is enough
but real argument checking should be more stringent than my example.
(2) how can i test for strictly #f #t boolean value inside the
immedatep clause?
(x == C_SCHEME_FALSE || x == C_SCHEME_TRUE)
#define C_truep(x) ((x) != C_SCHEME_FALSE)
Most of the C_*p are wrapped in a C_mk_bool which is why I used
C_truep for the test.
cheers,
felix
___
Chicken-users mailing list
Chicken-users@nongnu.org
http://lists.nongnu.org/mailman/listinfo/chicken-users
Best Wishes,
Kon
___
Chicken-users mailing list
Chicken-users@nongnu.org
http://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] C_truep( C_fixnump( w ))
On Mon, Feb 25, 2008 at 1:28 PM, Heinrich Taube <[EMAIL PROTECTED]> wrote:
> thank you both very much. just to be sure:
>
> (1) should i be using a C_truep() around C_immediatep() like the other
> predicate tests or is this ok:
>
> if ( C_immediatep(w) ) {
>if ( C_truep(C_fixnump(w)) ) {
>...
>}
>
> else if ( C_truep(C_blockp(w)) && C_truep( C_flonump( w ) )) {
>
Yes, this is correct. It is obviously inconsistent, but C_immediatep was
intended to be used from C and C_fixnump from compiled code.
Changing this will break old (compiled) code.
>
> (2) how can i test for strictly #f #t boolean value inside the
> immedatep clause?
>
(x == C_SCHEME_FALSE || x == C_SCHEME_TRUE)
cheers,
felix
___
Chicken-users mailing list
Chicken-users@nongnu.org
http://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] C_truep( C_fixnump( w ))
thank you both very much. just to be sure:
(1) should i be using a C_truep() around C_immediatep() like the other
predicate tests or is this ok:
if ( C_immediatep(w) ) {
if ( C_truep(C_fixnump(w)) ) {
...
}
else if ( C_truep(C_blockp(w)) && C_truep( C_flonump( w ) )) {
(2) how can i test for strictly #f #t boolean value inside the
immedatep clause?
___
Chicken-users mailing list
Chicken-users@nongnu.org
http://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] C_truep( C_fixnump( w ))
On Sun, Feb 24, 2008 at 8:15 PM, Jim Ursetto <[EMAIL PROTECTED]> wrote:
> On 2/24/08, Heinrich Taube <[EMAIL PROTECTED]> wrote:
> > im sure this must be something stupid but i dont understand why it
> > crashes if its not float data
>
>
> > C_word w = C_u_i_car( lyst );
> > if (C_truep( C_flonump( w ) )) {
>
>
> Heinrich,
>
> Try if (C_truep(C_blockp(w)) && C_truep(C_flonump(w))) {...}
> and see if it helps.
>
> Since you're doing this completely in C, you don't really
> have to convert to a scheme boolean either:
>
> if (!C_immediatep(w) && C_block_header(x) == C_FLONUM_TAG) { ... }
>
>
Right.
C_flonump does not check whether the argument is non-immediate
and fetches the header of the object pointed to. If the argument is
immediate it naturally will not point to a valid block.
This may appear as some kind of premature optimization, but doing
the check in the C_flonump macro would evaluate the argument
twice, and this might result in invalid behaviour in case the argument
is a ("direct") call to an optimized function.
cheers,
felix
___
Chicken-users mailing list
Chicken-users@nongnu.org
http://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] C_truep( C_fixnump( w ))
On 2/24/08, Heinrich Taube <[EMAIL PROTECTED]> wrote:
> im sure this must be something stupid but i dont understand why it
> crashes if its not float data
> C_word w = C_u_i_car( lyst );
> if (C_truep( C_flonump( w ) )) {
Heinrich,
Try if (C_truep(C_blockp(w)) && C_truep(C_flonump(w))) {...}
and see if it helps.
Since you're doing this completely in C, you don't really
have to convert to a scheme boolean either:
if (!C_immediatep(w) && C_block_header(x) == C_FLONUM_TAG) { ... }
___
Chicken-users mailing list
Chicken-users@nongnu.org
http://lists.nongnu.org/mailman/listinfo/chicken-users
Re: [Chicken-users] C_truep( C_fixnump( w ))
hmmm it has nothing to do with the C_truep( C_fixnump( w )) test --
even if the else clause is empty it still crashes if the data are NOT
floats.
im sure this must be something stupid but i dont understand why it
crashes if its not float data
#>
#include "Csound.h"
void cs_event (int type, int len, C_word lyst) {
// MYFLT buf[len];
printf("IN CS_EVENT, BEFORE LOOP LEN=%d\n", len);
int i=0;
for (; C_SCHEME_END_OF_LIST != lyst; lyst = C_u_i_cdr( lyst ) ) {
if (i==len) break;
C_word w = C_u_i_car( lyst );
if (C_truep( C_flonump( w ) )) {
printf("FLO: buf[%d]=%f\n", i++, (float)C_flonum_magnitude( w ));
}
else {
}
}
// ((GraceApp *)GraceApp::getInstance())->getCsoundPort()-
>sendEvent(typ, len, buf);
}
<#
(define (cs:i . args)
(if (not (null? args))
((foreign-lambda void "cs_event" int int scheme-object)
1
(length args)
args))
(values))
---
(cs:i 1.0 0.0 1.0 60.0 1000.0)
IN CS_EVENT, BEFORE LOOP LEN=5
FLO: buf[0]=1.00
FLO: buf[1]=0.00
FLO: buf[2]=1.00
FLO: buf[3]=60.00
FLO: buf[4]=1000.00
(cs:i 1 0 1 60 1000)
<<>>
___
Chicken-users mailing list
Chicken-users@nongnu.org
http://lists.nongnu.org/mailman/listinfo/chicken-users
[Chicken-users] C_truep( C_fixnump( w ))
hi im using chicken 3.0.0 im trying to parse a list of floats and/or
ints that come from scheme.
in the example below the floating point test:
if (C_truep( C_flonump( w ) ))
works fine but the integer test
else if (C_truep( C_fixnump( w ) ))
crashes my app. im not sure what is wrong and ive spent about an hour
trying to figure this out. aplogies if its somethig stupid that my
eyes are not seeing!
note that its the actuall test not the printf thats crashing, (if i
comment out the printf it still crashes).
int i=0;
for (; C_SCHEME_END_OF_LIST != lyst; lyst = C_u_i_cdr( lyst ) ) {
C_word w = C_u_i_car( lyst );
if (C_truep( C_flonump( w ) )) {
printf("FLO: buf[%d]=%f\n", i++, (float)C_flonum_magnitude( w ));
}
else if (C_truep( C_fixnump( w ) )) {
printf("FIX: buf[%d]=%d\n", i++, C_unfix( w ));
}
else {
}
}
___
Chicken-users mailing list
Chicken-users@nongnu.org
http://lists.nongnu.org/mailman/listinfo/chicken-users
