d7482997 |
1 | /* |
2 | * ustring.c: Unicode string routines |
3 | */ |
4 | |
5 | #include <wchar.h> |
6 | #include <time.h> |
7 | #include "halibut.h" |
8 | |
9 | wchar_t *ustrdup(wchar_t *s) { |
10 | wchar_t *r; |
11 | if (s) { |
12 | r = mknewa(wchar_t, 1+ustrlen(s)); |
13 | ustrcpy(r, s); |
14 | } else { |
15 | r = mknew(wchar_t); |
16 | *r = 0; |
17 | } |
18 | return r; |
19 | } |
20 | |
21 | char *ustrtoa(wchar_t *s, char *outbuf, int size) { |
22 | char *p; |
23 | if (!s) { |
24 | *outbuf = '\0'; |
25 | return outbuf; |
26 | } |
27 | for (p = outbuf; *s && p < outbuf+size; p++,s++) |
28 | *p = *s; |
29 | if (p < outbuf+size) |
30 | *p = '\0'; |
31 | else |
32 | outbuf[size-1] = '\0'; |
33 | return outbuf; |
34 | } |
35 | |
50d6b4bd |
36 | char *utoa_dup(wchar_t *s) { |
37 | int len; |
38 | char *buf = NULL; |
39 | |
40 | len = ustrlen(s) + 1; |
41 | do { |
42 | buf = resize(buf, len); |
43 | ustrtoa(s, buf, len); |
44 | len = (3 * len) / 2 + 1; /* this guarantees a strict increase */ |
45 | } while ((int)strlen(buf) >= len-1); |
46 | |
47 | buf = resize(buf, strlen(buf)+1); |
48 | return buf; |
49 | } |
50 | |
d7482997 |
51 | int ustrlen(wchar_t *s) { |
52 | int len = 0; |
53 | while (*s++) len++; |
54 | return len; |
55 | } |
56 | |
57 | wchar_t *uadv(wchar_t *s) { |
58 | return s + 1 + ustrlen(s); |
59 | } |
60 | |
61 | wchar_t *ustrcpy(wchar_t *dest, wchar_t *source) { |
62 | wchar_t *ret = dest; |
63 | do { |
64 | *dest++ = *source; |
65 | } while (*source++); |
66 | return ret; |
67 | } |
68 | |
69 | int ustrcmp(wchar_t *lhs, wchar_t *rhs) { |
70 | if (!lhs && !rhs) return 0; |
71 | if (!lhs) return -1; |
72 | if (!rhs) return +1; |
73 | while (*lhs && *rhs && *lhs==*rhs) |
74 | lhs++, rhs++; |
75 | if (*lhs < *rhs) |
76 | return -1; |
77 | else if (*lhs > *rhs) |
78 | return 1; |
79 | return 0; |
80 | } |
81 | |
82 | wchar_t utolower(wchar_t c) { |
83 | if (c == L'\0') |
84 | return c; /* this property needed by ustricmp */ |
85 | /* FIXME: this doesn't even come close */ |
86 | if (c >= 'A' && c <= 'Z') |
87 | c += 'a'-'A'; |
88 | return c; |
89 | } |
90 | |
831da32e |
91 | int uisalpha(wchar_t c) { |
92 | /* FIXME: this doesn't even come close */ |
93 | return (c >= 'A' && c <= 'Z') || (c >= 'a' && c <= 'z'); |
94 | } |
95 | |
d7482997 |
96 | int ustricmp(wchar_t *lhs, wchar_t *rhs) { |
97 | wchar_t lc, rc; |
98 | while ((lc = utolower(*lhs)) == (rc = utolower(*rhs)) && lc && rc) |
99 | lhs++, rhs++; |
100 | if (!lc && !rc) |
101 | return 0; |
102 | if (lc < rc) |
103 | return -1; |
104 | else |
105 | return 1; |
106 | } |
107 | |
108 | wchar_t *ustrlow(wchar_t *s) { |
109 | wchar_t *p = s; |
110 | while (*p) { |
111 | *p = utolower(*p); |
112 | p++; |
113 | } |
114 | return s; |
115 | } |
116 | |
117 | int utoi(wchar_t *s) { |
118 | int sign = +1; |
119 | int n; |
120 | |
121 | if (*s == L'-') { |
122 | s++; |
123 | sign = -1; |
124 | } |
125 | |
126 | n = 0; |
127 | while (*s && *s >= L'0' && *s <= L'9') { |
128 | n *= 10; |
129 | n += (*s - '0'); |
130 | s++; |
131 | } |
132 | |
133 | return n; |
134 | } |
135 | |
136 | int utob(wchar_t *s) { |
137 | if (!ustricmp(s, L"yes") || !ustricmp(s, L"y") || |
138 | !ustricmp(s, L"true") || !ustricmp(s, L"t")) |
139 | return TRUE; |
140 | return FALSE; |
141 | } |
142 | |
143 | int uisdigit(wchar_t c) { |
144 | return c >= L'0' && c <= L'9'; |
145 | } |
146 | |
147 | #define USTRFTIME_DELTA 128 |
148 | wchar_t *ustrftime(wchar_t *wfmt, struct tm *timespec) { |
149 | void *blk = NULL; |
150 | wchar_t *wblk, *wp; |
151 | char *fmt, *text, *p; |
152 | size_t size = 0; |
153 | size_t len; |
154 | |
155 | /* |
156 | * strftime has the entertaining property that it returns 0 |
157 | * _either_ on out-of-space _or_ on successful generation of |
158 | * the empty string. Hence we must ensure our format can never |
159 | * generate the empty string. Somebody throw a custard pie at |
160 | * whoever was responsible for that. Please? |
161 | */ |
162 | if (wfmt) { |
163 | len = ustrlen(wfmt); |
164 | fmt = mknewa(char, 2+len); |
165 | ustrtoa(wfmt, fmt+1, len+1); |
166 | fmt[0] = ' '; |
167 | } else |
168 | fmt = " %c"; |
169 | |
170 | while (1) { |
171 | size += USTRFTIME_DELTA; |
172 | blk = resize((char *)blk, size); |
173 | len = strftime((char *)blk, size-1, fmt, timespec); |
174 | if (len > 0) |
175 | break; |
176 | } |
177 | |
178 | /* Note: +1 for the terminating 0, -1 for the initial space in fmt */ |
179 | wblk = resize((wchar_t *)blk, len); |
180 | text = mknewa(char, len); |
181 | strftime(text, len, fmt+1, timespec); |
182 | /* |
183 | * We operate in the C locale, so this all ought to be kosher |
184 | * ASCII. If we ever move outside ASCII machines, we may need |
185 | * to make this more portable... |
186 | */ |
187 | for (wp = wblk, p = text; *p; p++, wp++) |
188 | *wp = *p; |
189 | *wp = 0; |
190 | if (wfmt) |
191 | sfree(fmt); |
192 | sfree(text); |
193 | return wblk; |
194 | } |