Christian Heimes | 1a6387e | 2008-03-26 12:49:49 +0000 | [diff] [blame] | 1 | /* NOTE: this API is -ONLY- for use with single byte character strings. */ |
| 2 | /* Do not use it with Unicode. */ |
| 3 | |
Christian Heimes | 1a6387e | 2008-03-26 12:49:49 +0000 | [diff] [blame] | 4 | /* the more complicated methods. parts of these should be pulled out into the |
| 5 | shared code in bytes_methods.c to cut down on duplicate code bloat. */ |
| 6 | |
| 7 | PyDoc_STRVAR(expandtabs__doc__, |
| 8 | "B.expandtabs([tabsize]) -> copy of B\n\ |
| 9 | \n\ |
| 10 | Return a copy of B where all tab characters are expanded using spaces.\n\ |
| 11 | If tabsize is not given, a tab size of 8 characters is assumed."); |
| 12 | |
| 13 | static PyObject* |
| 14 | stringlib_expandtabs(PyObject *self, PyObject *args) |
| 15 | { |
| 16 | const char *e, *p; |
| 17 | char *q; |
Benjamin Peterson | 18fc493 | 2014-03-30 19:47:57 -0400 | [diff] [blame] | 18 | Py_ssize_t i, j; |
Christian Heimes | 1a6387e | 2008-03-26 12:49:49 +0000 | [diff] [blame] | 19 | PyObject *u; |
| 20 | int tabsize = 8; |
Antoine Pitrou | c2f0221 | 2009-01-13 23:13:52 +0000 | [diff] [blame] | 21 | |
Christian Heimes | 1a6387e | 2008-03-26 12:49:49 +0000 | [diff] [blame] | 22 | if (!PyArg_ParseTuple(args, "|i:expandtabs", &tabsize)) |
Antoine Pitrou | c2f0221 | 2009-01-13 23:13:52 +0000 | [diff] [blame] | 23 | return NULL; |
| 24 | |
Christian Heimes | 1a6387e | 2008-03-26 12:49:49 +0000 | [diff] [blame] | 25 | /* First pass: determine size of output string */ |
Antoine Pitrou | c2f0221 | 2009-01-13 23:13:52 +0000 | [diff] [blame] | 26 | i = j = 0; |
Christian Heimes | 1a6387e | 2008-03-26 12:49:49 +0000 | [diff] [blame] | 27 | e = STRINGLIB_STR(self) + STRINGLIB_LEN(self); |
Benjamin Peterson | 18fc493 | 2014-03-30 19:47:57 -0400 | [diff] [blame] | 28 | for (p = STRINGLIB_STR(self); p < e; p++) { |
Christian Heimes | 1a6387e | 2008-03-26 12:49:49 +0000 | [diff] [blame] | 29 | if (*p == '\t') { |
Antoine Pitrou | c2f0221 | 2009-01-13 23:13:52 +0000 | [diff] [blame] | 30 | if (tabsize > 0) { |
Benjamin Peterson | 18fc493 | 2014-03-30 19:47:57 -0400 | [diff] [blame] | 31 | Py_ssize_t incr = tabsize - (j % tabsize); |
| 32 | if (j > PY_SSIZE_T_MAX - incr) |
| 33 | goto overflow; |
| 34 | j += incr; |
Christian Heimes | 1a6387e | 2008-03-26 12:49:49 +0000 | [diff] [blame] | 35 | } |
| 36 | } |
Antoine Pitrou | c2f0221 | 2009-01-13 23:13:52 +0000 | [diff] [blame] | 37 | else { |
Benjamin Peterson | 18fc493 | 2014-03-30 19:47:57 -0400 | [diff] [blame] | 38 | if (j > PY_SSIZE_T_MAX - 1) |
| 39 | goto overflow; |
Antoine Pitrou | c2f0221 | 2009-01-13 23:13:52 +0000 | [diff] [blame] | 40 | j++; |
| 41 | if (*p == '\n' || *p == '\r') { |
Benjamin Peterson | 18fc493 | 2014-03-30 19:47:57 -0400 | [diff] [blame] | 42 | if (i > PY_SSIZE_T_MAX - j) |
| 43 | goto overflow; |
Antoine Pitrou | c2f0221 | 2009-01-13 23:13:52 +0000 | [diff] [blame] | 44 | i += j; |
| 45 | j = 0; |
Antoine Pitrou | c2f0221 | 2009-01-13 23:13:52 +0000 | [diff] [blame] | 46 | } |
| 47 | } |
Christian Heimes | 1a6387e | 2008-03-26 12:49:49 +0000 | [diff] [blame] | 48 | } |
Antoine Pitrou | c2f0221 | 2009-01-13 23:13:52 +0000 | [diff] [blame] | 49 | |
Benjamin Peterson | 18fc493 | 2014-03-30 19:47:57 -0400 | [diff] [blame] | 50 | if (i > PY_SSIZE_T_MAX - j) |
| 51 | goto overflow; |
| 52 | |
Christian Heimes | 1a6387e | 2008-03-26 12:49:49 +0000 | [diff] [blame] | 53 | /* Second pass: create output string and fill it */ |
| 54 | u = STRINGLIB_NEW(NULL, i + j); |
| 55 | if (!u) |
| 56 | return NULL; |
Antoine Pitrou | c2f0221 | 2009-01-13 23:13:52 +0000 | [diff] [blame] | 57 | |
Christian Heimes | 1a6387e | 2008-03-26 12:49:49 +0000 | [diff] [blame] | 58 | j = 0; |
| 59 | q = STRINGLIB_STR(u); |
Antoine Pitrou | c2f0221 | 2009-01-13 23:13:52 +0000 | [diff] [blame] | 60 | |
Benjamin Peterson | 18fc493 | 2014-03-30 19:47:57 -0400 | [diff] [blame] | 61 | for (p = STRINGLIB_STR(self); p < e; p++) { |
Christian Heimes | 1a6387e | 2008-03-26 12:49:49 +0000 | [diff] [blame] | 62 | if (*p == '\t') { |
Antoine Pitrou | c2f0221 | 2009-01-13 23:13:52 +0000 | [diff] [blame] | 63 | if (tabsize > 0) { |
| 64 | i = tabsize - (j % tabsize); |
| 65 | j += i; |
| 66 | while (i--) |
| 67 | *q++ = ' '; |
| 68 | } |
| 69 | } |
| 70 | else { |
Christian Heimes | 1a6387e | 2008-03-26 12:49:49 +0000 | [diff] [blame] | 71 | j++; |
Antoine Pitrou | c2f0221 | 2009-01-13 23:13:52 +0000 | [diff] [blame] | 72 | *q++ = *p; |
Christian Heimes | 1a6387e | 2008-03-26 12:49:49 +0000 | [diff] [blame] | 73 | if (*p == '\n' || *p == '\r') |
| 74 | j = 0; |
| 75 | } |
Benjamin Peterson | 18fc493 | 2014-03-30 19:47:57 -0400 | [diff] [blame] | 76 | } |
| 77 | |
Christian Heimes | 1a6387e | 2008-03-26 12:49:49 +0000 | [diff] [blame] | 78 | return u; |
Benjamin Peterson | 18fc493 | 2014-03-30 19:47:57 -0400 | [diff] [blame] | 79 | overflow: |
| 80 | PyErr_SetString(PyExc_OverflowError, "result too long"); |
| 81 | return NULL; |
Christian Heimes | 1a6387e | 2008-03-26 12:49:49 +0000 | [diff] [blame] | 82 | } |
| 83 | |
| 84 | Py_LOCAL_INLINE(PyObject *) |
| 85 | pad(PyObject *self, Py_ssize_t left, Py_ssize_t right, char fill) |
| 86 | { |
| 87 | PyObject *u; |
| 88 | |
| 89 | if (left < 0) |
| 90 | left = 0; |
| 91 | if (right < 0) |
| 92 | right = 0; |
| 93 | |
| 94 | if (left == 0 && right == 0 && STRINGLIB_CHECK_EXACT(self)) { |
| 95 | #if STRINGLIB_MUTABLE |
| 96 | /* We're defined as returning a copy; If the object is mutable |
| 97 | * that means we must make an identical copy. */ |
| 98 | return STRINGLIB_NEW(STRINGLIB_STR(self), STRINGLIB_LEN(self)); |
| 99 | #else |
| 100 | Py_INCREF(self); |
| 101 | return (PyObject *)self; |
| 102 | #endif /* STRINGLIB_MUTABLE */ |
| 103 | } |
| 104 | |
| 105 | u = STRINGLIB_NEW(NULL, |
| 106 | left + STRINGLIB_LEN(self) + right); |
| 107 | if (u) { |
| 108 | if (left) |
| 109 | memset(STRINGLIB_STR(u), fill, left); |
| 110 | Py_MEMCPY(STRINGLIB_STR(u) + left, |
| 111 | STRINGLIB_STR(self), |
| 112 | STRINGLIB_LEN(self)); |
| 113 | if (right) |
| 114 | memset(STRINGLIB_STR(u) + left + STRINGLIB_LEN(self), |
| 115 | fill, right); |
| 116 | } |
| 117 | |
| 118 | return u; |
| 119 | } |
| 120 | |
| 121 | PyDoc_STRVAR(ljust__doc__, |
| 122 | "B.ljust(width[, fillchar]) -> copy of B\n" |
| 123 | "\n" |
| 124 | "Return B left justified in a string of length width. Padding is\n" |
| 125 | "done using the specified fill character (default is a space)."); |
| 126 | |
| 127 | static PyObject * |
| 128 | stringlib_ljust(PyObject *self, PyObject *args) |
| 129 | { |
| 130 | Py_ssize_t width; |
| 131 | char fillchar = ' '; |
| 132 | |
| 133 | if (!PyArg_ParseTuple(args, "n|c:ljust", &width, &fillchar)) |
| 134 | return NULL; |
| 135 | |
| 136 | if (STRINGLIB_LEN(self) >= width && STRINGLIB_CHECK_EXACT(self)) { |
| 137 | #if STRINGLIB_MUTABLE |
| 138 | /* We're defined as returning a copy; If the object is mutable |
| 139 | * that means we must make an identical copy. */ |
| 140 | return STRINGLIB_NEW(STRINGLIB_STR(self), STRINGLIB_LEN(self)); |
| 141 | #else |
| 142 | Py_INCREF(self); |
| 143 | return (PyObject*) self; |
| 144 | #endif |
| 145 | } |
| 146 | |
| 147 | return pad(self, 0, width - STRINGLIB_LEN(self), fillchar); |
| 148 | } |
| 149 | |
| 150 | |
| 151 | PyDoc_STRVAR(rjust__doc__, |
| 152 | "B.rjust(width[, fillchar]) -> copy of B\n" |
| 153 | "\n" |
| 154 | "Return B right justified in a string of length width. Padding is\n" |
| 155 | "done using the specified fill character (default is a space)"); |
| 156 | |
| 157 | static PyObject * |
| 158 | stringlib_rjust(PyObject *self, PyObject *args) |
| 159 | { |
| 160 | Py_ssize_t width; |
| 161 | char fillchar = ' '; |
| 162 | |
| 163 | if (!PyArg_ParseTuple(args, "n|c:rjust", &width, &fillchar)) |
| 164 | return NULL; |
| 165 | |
| 166 | if (STRINGLIB_LEN(self) >= width && STRINGLIB_CHECK_EXACT(self)) { |
| 167 | #if STRINGLIB_MUTABLE |
| 168 | /* We're defined as returning a copy; If the object is mutable |
| 169 | * that means we must make an identical copy. */ |
| 170 | return STRINGLIB_NEW(STRINGLIB_STR(self), STRINGLIB_LEN(self)); |
| 171 | #else |
| 172 | Py_INCREF(self); |
| 173 | return (PyObject*) self; |
| 174 | #endif |
| 175 | } |
| 176 | |
| 177 | return pad(self, width - STRINGLIB_LEN(self), 0, fillchar); |
| 178 | } |
| 179 | |
| 180 | |
| 181 | PyDoc_STRVAR(center__doc__, |
| 182 | "B.center(width[, fillchar]) -> copy of B\n" |
| 183 | "\n" |
| 184 | "Return B centered in a string of length width. Padding is\n" |
| 185 | "done using the specified fill character (default is a space)."); |
| 186 | |
| 187 | static PyObject * |
| 188 | stringlib_center(PyObject *self, PyObject *args) |
| 189 | { |
| 190 | Py_ssize_t marg, left; |
| 191 | Py_ssize_t width; |
| 192 | char fillchar = ' '; |
| 193 | |
| 194 | if (!PyArg_ParseTuple(args, "n|c:center", &width, &fillchar)) |
| 195 | return NULL; |
| 196 | |
| 197 | if (STRINGLIB_LEN(self) >= width && STRINGLIB_CHECK_EXACT(self)) { |
| 198 | #if STRINGLIB_MUTABLE |
| 199 | /* We're defined as returning a copy; If the object is mutable |
| 200 | * that means we must make an identical copy. */ |
| 201 | return STRINGLIB_NEW(STRINGLIB_STR(self), STRINGLIB_LEN(self)); |
| 202 | #else |
| 203 | Py_INCREF(self); |
| 204 | return (PyObject*) self; |
| 205 | #endif |
| 206 | } |
| 207 | |
| 208 | marg = width - STRINGLIB_LEN(self); |
| 209 | left = marg / 2 + (marg & width & 1); |
| 210 | |
| 211 | return pad(self, left, marg - left, fillchar); |
| 212 | } |
| 213 | |
| 214 | PyDoc_STRVAR(zfill__doc__, |
| 215 | "B.zfill(width) -> copy of B\n" |
| 216 | "\n" |
| 217 | "Pad a numeric string B with zeros on the left, to fill a field\n" |
| 218 | "of the specified width. B is never truncated."); |
| 219 | |
| 220 | static PyObject * |
| 221 | stringlib_zfill(PyObject *self, PyObject *args) |
| 222 | { |
| 223 | Py_ssize_t fill; |
| 224 | PyObject *s; |
| 225 | char *p; |
| 226 | Py_ssize_t width; |
| 227 | |
| 228 | if (!PyArg_ParseTuple(args, "n:zfill", &width)) |
| 229 | return NULL; |
| 230 | |
| 231 | if (STRINGLIB_LEN(self) >= width) { |
| 232 | if (STRINGLIB_CHECK_EXACT(self)) { |
| 233 | #if STRINGLIB_MUTABLE |
| 234 | /* We're defined as returning a copy; If the object is mutable |
| 235 | * that means we must make an identical copy. */ |
| 236 | return STRINGLIB_NEW(STRINGLIB_STR(self), STRINGLIB_LEN(self)); |
| 237 | #else |
| 238 | Py_INCREF(self); |
| 239 | return (PyObject*) self; |
| 240 | #endif |
| 241 | } |
| 242 | else |
| 243 | return STRINGLIB_NEW( |
| 244 | STRINGLIB_STR(self), |
| 245 | STRINGLIB_LEN(self) |
| 246 | ); |
| 247 | } |
| 248 | |
| 249 | fill = width - STRINGLIB_LEN(self); |
| 250 | |
| 251 | s = pad(self, fill, 0, '0'); |
| 252 | |
| 253 | if (s == NULL) |
| 254 | return NULL; |
| 255 | |
| 256 | p = STRINGLIB_STR(s); |
| 257 | if (p[fill] == '+' || p[fill] == '-') { |
| 258 | /* move sign to beginning of string */ |
| 259 | p[0] = p[fill]; |
| 260 | p[fill] = '0'; |
| 261 | } |
| 262 | |
| 263 | return (PyObject*) s; |
| 264 | } |