refactor!: modify C implementation to accept float value instead of int32 in math/base/special/negafibonaccif

This commit updates the signature of the C API to accept a float rather than int32. The rationale was so that users get the same results/behavior in both JavaScript and C. This arose in the context of ufuncs, where the diverging type signatures meant differences in what dtypes would be permissible. Instead, we decided to unify and ensure the behavior is consistent.

BREAKING CHANGE: update signature to accept floats

User code should behave similarly in the primary case of providing integer-valued input values. However, no longer will real-values truncate. Now, real-valued inputs will result in `NaN`, which is, arguably, better behavior, as real-to-integer truncation can be a source of silent bugs.

PR-URL: #7954
Reviewed-by: Athan Reines <[email protected]>

Author: gunjjoshi

lib/node_modules/@stdlib/math/base/special/negafibonaccif/README.md

@@ -174,19 +174,19 @@ logEachMap( 'negafibonaccif(%d) = %0.4f', v, negafibonaccif );
 Computes the nth [negaFibonacci number][fibonacci-number] as a [single-precision floating-point number][ieee754].
 
 ```c
-float out = stdlib_base_negafibonaccif( 0 );
+float out = stdlib_base_negafibonaccif( 0.0f );
 // returns 0.0f
 
-out = stdlib_base_negafibonaccif( -1 );
+out = stdlib_base_negafibonaccif( -1.0f );
 // returns 1.0f
 ```
 
 The function accepts the following arguments:
 
--   **n**: `[in] int32_t` input value.
+-   **n**: `[in] float` input value.
 
 ```c
-float stdlib_base_negafibonaccif( const int32_t n );
+float stdlib_base_negafibonaccif( const float n );
 ```
 
 </section>
@@ -210,15 +210,14 @@ float stdlib_base_negafibonaccif( const int32_t n );
 ```c
 #include "stdlib/math/base/special/negafibonaccif.h"
 #include <stdio.h>
-#include <stdint.h>
 
 int main( void ) {
-    int32_t i;
+    float i;
     float v;
 
-    for ( i = 0; i > -37; i-- ) {
+    for ( i = 0.0f; i > -37.0f; i-- ) {
         v = stdlib_base_negafibonaccif( i );
-        printf( "negafibonaccif(%d) = %f\n", i, v );
+        printf( "negafibonaccif(%f) = %f\n", i, v );
     }
 }
 ```

lib/node_modules/@stdlib/math/base/special/negafibonaccif/benchmark/c/native/benchmark.c

@@ -17,7 +17,6 @@
 */
 
 #include "stdlib/math/base/special/negafibonaccif.h"
-#include <stdint.h>
 #include <stdlib.h>
 #include <stdio.h>
 #include <math.h>
@@ -91,14 +90,14 @@ static float rand_float( void ) {
 * @return elapsed time in seconds
 */
 static double benchmark( void ) {
-	int32_t x[ 100 ];
+	float x[ 100 ];
 	double elapsed;
 	double t;
 	float y;
 	int i;
 
 	for ( i = 0; i < 100; i++ ) {
-		x[ i ] = (int32_t)floorf( 36.0f*rand_float());
+		x[ i ] = floorf( 36.0f*rand_float());
 	}
 
 	t = tic();

lib/node_modules/@stdlib/math/base/special/negafibonaccif/examples/c/example.c

@@ -18,14 +18,13 @@
 
 #include "stdlib/math/base/special/negafibonaccif.h"
 #include <stdio.h>
-#include <stdint.h>
 
 int main( void ) {
-	int32_t i;
+	float i;
 	float v;
 
-	for ( i = 0; i > -37; i-- ) {
+	for ( i = 0.0f; i > -37.0f; i-- ) {
 		v = stdlib_base_negafibonaccif( i );
-		printf( "negafibonaccif(%d) = %f\n", i, v );
+		printf( "negafibonaccif(%f) = %f\n", i, v );
 	}
 }

lib/node_modules/@stdlib/math/base/special/negafibonaccif/include/stdlib/math/base/special/negafibonaccif.h

@@ -19,8 +19,6 @@
 #ifndef STDLIB_MATH_BASE_SPECIAL_NEGAFIBONACCIF_H
 #define STDLIB_MATH_BASE_SPECIAL_NEGAFIBONACCIF_H
 
-#include <stdint.h>
-
 /*
 * If C++, prevent name mangling so that the compiler emits a binary file having undecorated names, thus mirroring the behavior of a C compiler.
 */
@@ -31,7 +29,7 @@ extern "C" {
 /**
 * Computes the nth negaFibonacci number as a single-precision floating-point number.
 */
-float stdlib_base_negafibonaccif( const int32_t n );
+float stdlib_base_negafibonaccif( const float n );
 
 #ifdef __cplusplus
 }

lib/node_modules/@stdlib/math/base/special/negafibonaccif/manifest.json

@@ -38,7 +38,8 @@
       "dependencies": [
         "@stdlib/math/base/napi/unary",
         "@stdlib/constants/float32/max-safe-nth-fibonacci",
-        "@stdlib/math/base/special/absf"
+        "@stdlib/math/base/special/absf",
+        "@stdlib/math/base/assert/is-integerf"
       ]
     },
     {
@@ -53,7 +54,8 @@
       "libpath": [],
       "dependencies": [
         "@stdlib/constants/float32/max-safe-nth-fibonacci",
-        "@stdlib/math/base/special/absf"
+        "@stdlib/math/base/special/absf",
+        "@stdlib/math/base/assert/is-integerf"
       ]
     },
     {
@@ -68,7 +70,8 @@
       "libpath": [],
       "dependencies": [
         "@stdlib/constants/float32/max-safe-nth-fibonacci",
-        "@stdlib/math/base/special/absf"
+        "@stdlib/math/base/special/absf",
+        "@stdlib/math/base/assert/is-integerf"
       ]
     }
   ]

lib/node_modules/@stdlib/math/base/special/negafibonaccif/src/addon.c

@@ -19,4 +19,4 @@
 #include "stdlib/math/base/special/negafibonaccif.h"
 #include "stdlib/math/base/napi/unary.h"
 
-STDLIB_MATH_BASE_NAPI_MODULE_I_F( stdlib_base_negafibonaccif )
+STDLIB_MATH_BASE_NAPI_MODULE_F_F( stdlib_base_negafibonaccif )

lib/node_modules/@stdlib/math/base/special/negafibonaccif/src/main.c

@@ -17,9 +17,10 @@
 */
 
 #include "stdlib/math/base/special/negafibonaccif.h"
+#include "stdlib/math/base/assert/is_integerf.h"
 #include "stdlib/math/base/special/absf.h"
 #include "stdlib/constants/float32/max_safe_nth_fibonacci.h"
-#include <stdint.h>
+#include <stdlib.h>
 
 static const float negafibonaccif_value[ 37 ] = {
 	0.0f,
@@ -68,17 +69,17 @@ static const float negafibonaccif_value[ 37 ] = {
 * @return     output value
 *
 * @example
-* float out = stdlib_base_negafibonaccif( -1 );
+* float out = stdlib_base_negafibonaccif( -1.0f );
 * // returns 1.0f
 */
-float stdlib_base_negafibonaccif( const int32_t n ) {
-	int32_t an;
-	if ( n > 0 ) {
+float stdlib_base_negafibonaccif( const float n ) {
+	float an;
+	if ( !stdlib_base_is_integerf( n ) || n > 0.0f ) {
 		return 0.0f / 0.0f; // NaN
 	}
 	an = stdlib_base_absf( n );
 	if ( an > STDLIB_CONSTANT_FLOAT32_MAX_SAFE_NTH_FIBONACCI ) {
 		return 0.0f / 0.0f; // NaN
 	}
-	return negafibonaccif_value[ an ];
+	return negafibonaccif_value[ (size_t)an ];
 }