From d54f66d1b4ce32d78b526b1ea9e3f213a763d07c Mon Sep 17 00:00:00 2001
From: Matt Valentine-House <matt@eightbitraptor.com>
Date: Thu, 8 Jun 2023 21:12:14 +0100
Subject: [PATCH] Assign into optimal size pools using String#split("")

When String#split is used with an empty string as the field seperator it
effectively splits the original string into chars, and there is a
pre-existing fast path for this using SPLIT_TYPE_CHARS.

However this path creates an empty array in the smallest size pool and
grows from there, despite already knowing the size of the desired array.

This commit pre-allocates the correct size array in this case in order
to allow the arrays to be embedded and avoid being allocated in the
transient heap
---
 string.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/string.c b/string.c
index a7d05693f1..ad502a1920 100644
--- a/string.c
+++ b/string.c
@@ -8662,7 +8662,6 @@ rb_str_split_m(int argc, VALUE *argv, VALUE str)
 
 #define SPLIT_STR(beg, len) (empty_count = split_string(result, str, beg, len, empty_count))
 
-    if (result) result = rb_ary_new();
     beg = 0;
     char *ptr = RSTRING_PTR(str);
     char *eptr = RSTRING_END(str);
@@ -8671,6 +8670,7 @@ rb_str_split_m(int argc, VALUE *argv, VALUE str)
         int skip = 1;
         unsigned int c;
 
+        if (result) result = rb_ary_new();
         end = beg;
         if (is_ascii_string(str)) {
             while (ptr < eptr) {
@@ -8730,6 +8730,7 @@ rb_str_split_m(int argc, VALUE *argv, VALUE str)
         char *sptr = RSTRING_PTR(spat);
         long slen = RSTRING_LEN(spat);
 
+        if (result) result = rb_ary_new();
         mustnot_broken(str);
         enc = rb_enc_check(str, spat);
         while (ptr < eptr &&
@@ -8751,6 +8752,7 @@ rb_str_split_m(int argc, VALUE *argv, VALUE str)
         char *str_start = ptr;
         int n;
 
+        if (result) result = rb_ary_new_capa(RSTRING_LEN(str));
         mustnot_broken(str);
         enc = rb_enc_get(str);
         while (ptr < eptr &&
@@ -8762,6 +8764,7 @@ rb_str_split_m(int argc, VALUE *argv, VALUE str)
         beg = ptr - str_start;
     }
     else {
+        if (result) result = rb_ary_new();
         long len = RSTRING_LEN(str);
         long start = beg;
         long idx;