libgrapheme

commit f128a915ba3140dc0b755d2f77d599cdbc2df7be
parent 92a5a67ed17573d9ad7d0a85ae602cb17d154f82
Author: Laslo Hunhold <dev@frign.de>
Date:   Sun, 12 Dec 2021 16:03:29 +0100

Change lg_grapheme_isbreak() return type from int to bool

Just as a disclaimer, I still really like int as a return type for
functions to indicate errors (return 0 means success and anything >0
means failure). For a library though one would most certainly use
an enum to name those errors, too, which would then turn the function
signature to "enum foo_errors foo(...)".

lg_grapheme_isbreak() cannot fail, however, and as in the previous
commit on reducing implicit assumptions on char, reducing assumptions
and making clearer what a return value can entail is good. This is
why the return type is changed to bool.

The bool type is C99 and even the more "independent" compilers like scc
support it. If it is internally expanded to an int, so be it, but it
helps readability.

Signed-off-by: Laslo Hunhold <dev@frign.de>

Diffstat:
M
grapheme.h
 | 
3
++-
M
src/grapheme.c
 | 
12
+++++++-----

2 files changed, 9 insertions(+), 6 deletions(-)
diff --git a/grapheme.h b/grapheme.h
@@ -2,6 +2,7 @@
 #ifndef GRAPHEME_H
 #define GRAPHEME_H
 
+#include <stdbool.h>
 #include <stddef.h>
 #include <stdint.h>
 
@@ -20,7 +21,7 @@ typedef struct lg_internal_segmentation_state {
 
 size_t lg_grapheme_nextbreak(const uint8_t *);
 
-int lg_grapheme_isbreak(uint_least32_t, uint_least32_t, LG_SEGMENTATION_STATE *);
+bool lg_grapheme_isbreak(uint_least32_t, uint_least32_t, LG_SEGMENTATION_STATE *);
 
 size_t lg_utf8_decode(const uint8_t *, size_t, uint_least32_t *);
 size_t lg_utf8_encode(uint_least32_t, uint8_t *, size_t);
diff --git a/src/grapheme.c b/src/grapheme.c
@@ -1,4 +1,5 @@
 /* See LICENSE file for copyright and license details. */
+#include <stdbool.h>
 #include <stddef.h>
 #include <stdlib.h>
 #include <string.h>
@@ -12,11 +13,12 @@ enum {
 	GRAPHEME_FLAG_EMOJI  = 1 << 1, /* within emoji modifier or zwj sequence */
 };
 
-int
+bool
 lg_grapheme_isbreak(uint_least32_t a, uint_least32_t b, LG_SEGMENTATION_STATE *state)
 {
 	struct lg_internal_heisenstate *p[2] = { 0 };
-	int ret = 1, flags = 0;
+	int flags = 0;
+	bool isbreak = true;
 
 	/* set state depending on state pointer */
 	if (state != NULL) {
@@ -160,7 +162,7 @@ lg_grapheme_isbreak(uint_least32_t a, uint_least32_t b, LG_SEGMENTATION_STATE *s
 	/* GB999 */
 	goto hasbreak;
 nobreak:
-	ret = 0;
+	isbreak = false;
 hasbreak:
 	if (state != NULL) {
 		/* move b-state to a-state, discard b-state */
@@ -168,12 +170,12 @@ hasbreak:
 		memset(&(state->b), 0, sizeof(state->b));
 
 		/* reset flags */
-		if (ret == 1) {
+		if (isbreak) {
 			state->flags = 0;
 		}
 	}
 
-	return ret;
+	return isbreak;
 }
 
 size_t