Introduce `at_char_boundary` function

4 files changed, 12 insertions, 8 deletions

diff --git a/file.c b/file.c
index e753e116de..30dd1cc73b 100644
--- a/file.c
+++ b/file.c
@@ -4610,7 +4610,7 @@ rmext(const char *p, long l0, long l1, const char *e, long l2, rb_encoding *enc)
     if (l1 < l2) return l1;
 
     s = p+l1-l2;
-    if (rb_enc_left_char_head(p, s, p+l1, enc) != s) return 0;
+    if (!at_char_boundary(p, s, p+l1, enc)) return 0;
 #if CASEFOLD_FILESYSTEM
 #define fncomp strncasecmp
 #else
diff --git a/internal/string.h b/internal/string.h
index 5f59d9621b..cfaf628e02 100644
--- a/internal/string.h
+++ b/internal/string.h
@@ -119,6 +119,12 @@ is_broken_string(VALUE str)
     return rb_enc_str_coderange(str) == ENC_CODERANGE_BROKEN;
 }
 
+static inline bool
+at_char_boundary(const char *s, const char *p, const char *e, rb_encoding *enc)
+{
+    return rb_enc_left_char_head(s, p, e, enc) == p;
+}
+
 /* expect tail call optimization */
 // YJIT needs this function to never allocate and never raise
 static inline VALUE
diff --git a/io.c b/io.c
index 6e27ed6273..bd1db9aa5d 100644
--- a/io.c
+++ b/io.c
@@ -4144,8 +4144,7 @@ rb_io_getline_0(VALUE rs, long limit, int chomp, rb_io_t *fptr)
                 s = RSTRING_PTR(str);
                 e = RSTRING_END(str);
                 p = e - rslen;
-                pp = rb_enc_left_char_head(s, p, e, enc);
-                if (pp != p) continue;
+                if (!at_char_boundary(s, p, e, enc)) continue;
                 if (!rspara) rscheck(rsptr, rslen, rs);
                 if (memcmp(p, rsptr, rslen) == 0) {
                     if (chomp) {
diff --git a/string.c b/string.c
index 0b9ede4e2c..b65868a226 100644
--- a/string.c
+++ b/string.c
@@ -3930,8 +3930,7 @@ str_ensure_byte_pos(VALUE str, long pos)
     const char *s = RSTRING_PTR(str);
     const char *e = RSTRING_END(str);
     const char *p = s + pos;
-    const char *pp = rb_enc_left_char_head(s, p, e, rb_enc_get(str));
-    if (p != pp) {
+    if (!at_char_boundary(s, p, e, rb_enc_get(str))) {
         rb_raise(rb_eIndexError,
                  "offset %ld does not land on character boundary", pos);
     }
@@ -9521,7 +9520,7 @@ chompped_length(VALUE str, VALUE rs)
     if (p[len-1] == newline &&
         (rslen <= 1 ||
          memcmp(rsptr, pp, rslen) == 0)) {
-        if (rb_enc_left_char_head(p, pp, e, enc) == pp)
+        if (at_char_boundary(p, pp, e, enc))
             return len - rslen;
         RB_GC_GUARD(rs);
     }
@@ -10497,7 +10496,7 @@ rb_str_end_with(int argc, VALUE *argv, VALUE str)
         p = RSTRING_PTR(str);
         e = p + slen;
         s = e - tlen;
-        if (rb_enc_left_char_head(p, s, e, enc) != s)
+        if (!at_char_boundary(p, s, e, enc))
             continue;
         if (memcmp(s, RSTRING_PTR(tmp), RSTRING_LEN(tmp)) == 0)
             return Qtrue;
@@ -10605,7 +10604,7 @@ deleted_suffix_length(VALUE str, VALUE suffix)
     suffixptr = RSTRING_PTR(suffix);
     s = strptr + olen - suffixlen;
     if (memcmp(s, suffixptr, suffixlen) != 0) return 0;
-    if (rb_enc_left_char_head(strptr, s, strptr + olen, enc) != s) return 0;
+    if (!at_char_boundary(strptr, s, strptr + olen, enc)) return 0;
 
     return suffixlen;
 }