From f1eb9cf6c302a10a939b76252e9704da176167e5 Mon Sep 17 00:00:00 2001
From: Samantha McVey <samantham@posteo.net>
Date: Fri, 7 Apr 2017 22:19:22 -0700
Subject: [PATCH] Use FreeBSD memmem function on MacOS and improve comments

Use FreeBSD memmem on MacOS, because although MacOS comes with a memmem,
it is much slower than the FBSD one which implements
Crochemore+Perrin two-way string matching.

Also improve the comments to be more descriptive.
---
 src/platform/memmem.h | 19 ++++++++++++++-----
 src/strings/ops.c     |  6 ++++--
 2 files changed, 18 insertions(+), 7 deletions(-)

diff --git a/src/platform/memmem.h b/src/platform/memmem.h
index 1e6c2dee37..b193e51528 100644
--- a/src/platform/memmem.h
+++ b/src/platform/memmem.h
@@ -1,13 +1,22 @@
-#if defined _WIN32
+/* On Linux we use glibc's memmem which uses the Knuth-Morris-Pratt algorithm.
+ * We use FreeBSD's libc memmem on Windows and MacOS, which uses
+ * Crochemore-Perrin two-way string matching.
+ * Reasoning:
+ * Windows, does not include any native memmem
+ * MacOS has a memmem but is slower and originates from FreeBSD dated to 2005 */
+
+#if defined(_WIN32) || defined(__APPLE__) || defined(__Darwin__)
 #include "../3rdparty/freebsd/memmem.c"
 #else
-/* On systems that use Glibc, you must defined _GNU_SOURCE before including string.h
- * to get access to memmem. On BSD and MacOS this is not needed, though if they
- * happen to be using glibc instead of libc, it shouldn't hurt have defined _GNU_SOURCE */
+/* On systems that use glibc, you must define _GNU_SOURCE before including string.h
+ * to get access to memmem. */
 #define _GNU_SOURCE
-#endif
 #include <string.h>
+#endif
 
 void* MVM_memmem(const void *haystack, size_t haystacklen, const void *needle, size_t needlelen) {
     return memmem(haystack, haystacklen, needle, needlelen);
 }
+
+/* Extended info:
+ * In glibc, the Knuth-Morris-Pratt algorithm was added as of git tag glibc-2.8-44-g0caca71ac9 */
diff --git a/src/strings/ops.c b/src/strings/ops.c
index 7a12d34441..fb47d9a896 100644
--- a/src/strings/ops.c
+++ b/src/strings/ops.c
@@ -216,7 +216,8 @@ MVMint64 MVM_string_index(MVMThreadContext *tc, MVMString *haystack, MVMString *
         return -1;
 
     /* Fast paths when storage types are identical. Uses memmem function, which
-     * in glibc uses Knuth-Morris-Pratt algorithm as of glibc-2.8-44-g0caca71ac9 */
+     * uses Knuth-Morris-Pratt algorithm on Linux and on others
+     * Crochemore+Perrin two-way string matching */
     switch (haystack->body.storage_type) {
         case MVM_STRING_GRAPHEME_32:
             if (needle->body.storage_type == MVM_STRING_GRAPHEME_32) {
@@ -245,7 +246,8 @@ MVMint64 MVM_string_index(MVMThreadContext *tc, MVMString *haystack, MVMString *
                     haystack->body.storage.blob_8 + start, /* start position */
                     (hgraphs - start) * sizeof(MVMGrapheme8), /* length of haystack from start position to end */
                     needle->body.storage.blob_8, /* needle start */
-                    ngraphs * sizeof(MVMGrapheme8)); /* needle length */
+                    ngraphs * sizeof(MVMGrapheme8) /* needle length */
+                );
                 if (mm_return_8 == NULL)
                     return -1;
                 else