From: Chris Hanson <org/chris-hanson/cph>
Date: Wed, 29 Mar 2017 04:52:44 +0000 (-0700)
Subject: Normalize strings prior to hashing so equivalent sequences hash the same.
X-Git-Tag: mit-scheme-pucked-9.2.12~158^2~65
X-Git-Url: https://birchwood-abbey.net/git?a=commitdiff_plain;h=dc80c76d803b1e729aa58a7f802a1bc4d35106b4;p=mit-scheme.git

Normalize strings prior to hashing so equivalent sequences hash the same.

I've arbitrarily chosen NFD because its faster than NFC, but a case could be
made that NFC is preferable.
---

diff --git a/src/runtime/ustring.scm b/src/runtime/ustring.scm
index 073380e10..af79646ce 100644
--- a/src/runtime/ustring.scm
+++ b/src/runtime/ustring.scm
@@ -1598,7 +1598,7 @@ USA.
 	      string))
 
 (define (string-hash string #!optional modulus)
-  (let ((string* (string-for-primitive string)))
+  (let ((string* (string-for-primitive (string->nfd string))))
     (if (default-object? modulus)
 	((ucode-primitive string-hash) string*)
 	((ucode-primitive string-hash-mod) string* modulus))))