1 /*
2  * This file and its contents are supplied under the terms of the
3  * Common Development and Distribution License ("CDDL"), version 1.0.
4  * You may only use this file in accordance with the terms of version
5  * 1.0 of the CDDL.
6  *
7  * A full copy of the text of the CDDL should have accompanied this
8  * source.  A copy of the CDDL is also available via the Internet at
9  * http://www.illumos.org/license/CDDL.
10  */
11 
12 /*
13  * Copyright 2016 Tom Lane <tgl@sss.pgh.pa.us>
14  * Copyright 2017 Nexenta Systems, Inc.
15  */
16 
17 #include <err.h>
18 #include <errno.h>
19 #include <locale.h>
20 #include <stdio.h>
21 #include <stdlib.h>
22 #include <string.h>
23 #include <time.h>
24 
25 /*
26  * #6907: generate random UTF8 strings, strxfrm'ing them in process.
27  * Walk through comparing each string with all strings, and checking
28  * that strcoll() and strcmp() for strxfrm'ed data produce same results.
29  */
30 #define	NSTRINGS 2000
31 #define	MAXSTRLEN 20
32 #define	MAXXFRMLEN (MAXSTRLEN * 20)
33 
34 typedef struct {
35 	char	sval[MAXSTRLEN];
36 	char	xval[MAXXFRMLEN];
37 } cstr;
38 
39 int
main(void)40 main(void)
41 {
42 	cstr	data[NSTRINGS];
43 	char	*curloc;
44 	int	i, j;
45 
46 	if ((curloc = setlocale(LC_ALL, "")) == NULL)
47 		err(1, "setlocale");
48 
49 	/* Ensure new random() values on every run */
50 	srandom((unsigned int) time(NULL));
51 
52 	/* Generate random UTF8 strings of length less than MAXSTRLEN bytes */
53 	for (i = 0; i < NSTRINGS; i++) {
54 		char	*p;
55 		int	len;
56 
57 again:
58 		p = data[i].sval;
59 		len = 1 + (random() % (MAXSTRLEN - 1));
60 		while (len > 0) {
61 			int c;
62 
63 			/*
64 			 * Generate random printable char in ISO8859-1 range.
65 			 * Bias towards producing a lot of spaces.
66 			 */
67 			if ((random() % 16) < 3) {
68 				c = ' ';
69 			} else {
70 				do {
71 					c = random() & 0xFF;
72 				} while (!((c >= ' ' && c <= 127) ||
73 				    (c >= 0xA0 && c <= 0xFF)));
74 			}
75 
76 			if (c <= 127) {
77 				*p++ = c;
78 				len--;
79 			} else {
80 				if (len < 2)
81 					break;
82 				/* Poor man's utf8-ification */
83 				*p++ = 0xC0 + (c >> 6);
84 				len--;
85 				*p++ = 0x80 + (c & 0x3F);
86 				len--;
87 			}
88 		}
89 		*p = '\0';
90 
91 		/* strxfrm() each string as we produce it */
92 		errno = 0;
93 		if (strxfrm(data[i].xval, data[i].sval,
94 		    MAXXFRMLEN) >= MAXXFRMLEN) {
95 			errx(1, "strxfrm() result for %d-length string "
96 			    "exceeded %d bytes", (int)strlen(data[i].sval),
97 			    MAXXFRMLEN);
98 		}
99 		/* Amend strxfrm() failing for certain characters (#7962) */
100 		if (errno != 0)
101 			goto again;
102 	}
103 
104 	for (i = 0; i < NSTRINGS; i++) {
105 		for (j = 0; j < NSTRINGS; j++) {
106 			int sr = strcoll(data[i].sval, data[j].sval);
107 			int sx = strcmp(data[i].xval, data[j].xval);
108 
109 			if ((sr * sx < 0) || (sr * sx == 0 && sr + sx != 0)) {
110 				errx(1, "%s: diff for \"%s\" and \"%s\"",
111 				    curloc, data[i].sval, data[j].sval);
112 			}
113 		}
114 	}
115 
116 	return (0);
117 }
118