MFC r317034:

Fix strcoll_l disagreeing with strxfrm by reworking the forward order case in
wcscoll_l().

Illumos fixed this while grabbing back our patches:
https://www.illumos.org/rb/r/402/

This does not 100% fix what postgresql folks reported as there is still a
remaining issue: https://www.illumos.org/issues/7962, it improves the situation

The initial issue was reported in postgresql mailing lists:
https://www.postgresql.org/message-id/flat/111D0E27-A8F3-4A84-A4E0-B0FB703863DF@s24.com#111D0E27-A8F3-4A84-A4E0-B0FB703863DF@s24.com

Submitted by:	Yuri Pankov <yuri.pankov@nexenta.com>
Obtained from:	Illumos

1 files changed, 93 insertions, 0 deletions

diff --git a/lib/libc/tests/string/wcscoll_test.c b/lib/libc/tests/string/wcscoll_test.c
index 2f2472b..022ad8e 100644
--- a/lib/libc/tests/string/wcscoll_test.c
+++ b/lib/libc/tests/string/wcscoll_test.c
@@ -1,5 +1,7 @@
 /*-
  * Copyright (c) 2016 Baptiste Daroussin <bapt@FreeBSD.org>
+ * Copyright 2016 Tom Lane <tgl@sss.pgh.pa.us>
+ * Copyright 2017 Nexenta Systems, Inc.
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
@@ -30,6 +32,8 @@ __FBSDID("$FreeBSD$");
 #include <wchar.h>
 #include <locale.h>
 #include <stdlib.h>
+#include <time.h>
+#include <errno.h>
 
 #include <atf-c.h>
 
@@ -55,9 +59,98 @@ ATF_TC_BODY(russian_collation, tc)
 	    "Bad collation, expected: '%ls' got '%ls'", res, c);
 }
 
+#define	NSTRINGS 2000
+#define	MAXSTRLEN 20
+#define	MAXXFRMLEN (MAXSTRLEN * 20)
+
+typedef struct {
+	char	sval[MAXSTRLEN];
+	char	xval[MAXXFRMLEN];
+} cstr;
+
+ATF_TC_WITHOUT_HEAD(strcoll_vs_strxfrm);
+ATF_TC_BODY(strcoll_vs_strxfrm, tc)
+{
+	cstr	data[NSTRINGS];
+	char	*curloc;
+	int	i, j;
+
+	curloc = setlocale(LC_ALL, "en_US.UTF-8");
+	ATF_CHECK_MSG(curloc != NULL, "Fail to set locale");
+
+	/* Ensure new random() values on every run */
+	srandom((unsigned int) time(NULL));
+
+	/* Generate random UTF8 strings of length less than MAXSTRLEN bytes */
+	for (i = 0; i < NSTRINGS; i++) {
+		char	*p;
+		int	len;
+
+again:
+		p = data[i].sval;
+		len = 1 + (random() % (MAXSTRLEN - 1));
+		while (len > 0) {
+			int c;
+			/*
+			 * Generate random printable char in ISO8859-1 range.
+			 * Bias towards producing a lot of spaces.
+			 */
+
+			if ((random() % 16) < 3) {
+				c = ' ';
+			} else {
+				do {
+					c = random() & 0xFF;
+				} while (!((c >= ' ' && c <= 127) ||
+				    (c >= 0xA0 && c <= 0xFF)));
+			}
+
+			if (c <= 127) {
+				*p++ = c;
+				len--;
+			} else {
+				if (len < 2)
+					break;
+				/* Poor man's utf8-ification */
+				*p++ = 0xC0 + (c >> 6);
+				len--;
+				*p++ = 0x80 + (c & 0x3F);
+				len--;
+			}
+		}
+		*p = '\0';
+		/* strxfrm() each string as we produce it */
+		errno = 0;
+		ATF_CHECK_MSG(strxfrm(data[i].xval, data[i].sval,
+		    MAXXFRMLEN) < MAXXFRMLEN, "strxfrm() result for %d-length "
+		    " string exceeded %d bytes", (int)strlen(data[i].sval),
+		    MAXXFRMLEN);
+
+		/*
+		 * Amend strxfrm() failing on certain characters to be fixed and
+		 * test later
+		 */
+		if (errno != 0)
+			goto again;
+	}
+
+	for (i = 0; i < NSTRINGS; i++) {
+		for (j = 0; j < NSTRINGS; j++) {
+			int sr = strcoll(data[i].sval, data[j].sval);
+			int sx = strcmp(data[i].xval, data[j].xval);
+
+			ATF_CHECK_MSG(!((sr * sx < 0) ||
+			    (sr * sx == 0 && sr + sx != 0)),
+			    "%s: diff for \"%s\" and \"%s\"",
+			    curloc, data[i].sval, data[j].sval);
+		}
+	}
+}
+
 ATF_TP_ADD_TCS(tp)
 {
 	ATF_TP_ADD_TC(tp, russian_collation);
+	ATF_TP_ADD_TC(tp, strcoll_vs_strxfrm);
 
 	return (atf_no_error());
 }