1 /* $Id: dos2unix.c 3114 2017-10-29 18:02:04Z bird $ */
2 /** @file
3 * dos2unix - Line ending conversion routines.
4 */
5
6 /*
7 * Copyright (c) 2017 knut st. osmundsen <bird-kBuild-spamx@anduin.net>
8 *
9 * Permission is hereby granted, free of charge, to any person obtaining a
10 * copy of this software and associated documentation files (the "Software"),
11 * to deal in the Software without restriction, including without limitation
12 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
13 * and/or sell copies of the Software, and to permit persons to whom the
14 * Software is furnished to do so, subject to the following conditions:
15 *
16 * The above copyright notice and this permission notice shall be included
17 * in all copies or substantial portions of the Software.
18 *
19 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
20 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
21 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
22 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
23 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
24 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
25 * IN THE SOFTWARE.
26 *
27 * Alternatively, the content of this file may be used under the terms of the
28 * GPL version 2 or later, or LGPL version 2.1 or later.
29 */
30
31
32 /*********************************************************************************************************************************
33 * Header Files *
34 *********************************************************************************************************************************/
35 #include "dos2unix.h"
36 #include <k/kDefs.h>
37 #include <errno.h>
38 #include <fcntl.h>
39 #if K_OS == K_OS_WINDOWS
40 # include <io.h>
41 #else
42 # include <unistd.h>
43 #endif
44 #include <assert.h>
45
46 #ifndef O_BINARY
47 # ifdef _O_BINARY
48 # define O_BINARY _O_BINARY
49 # else
50 # define O_BINARY 0
51 # endif
52 #endif
53
54
55 /*********************************************************************************************************************************
56 * Defined Constants And Macros *
57 *********************************************************************************************************************************/
58 #define STACK_BUF_SIZE 0x20000
59
60 #define DOS2UNIX_LF 0x0a
61 #define DOS2UNIX_CR 0x0d
62
63
64
65 /**
66 * Does a line ending analysis of the given file.
67 *
68 * @returns 0 on success, errno value on open or read error.
69 * @param pszFilename The path to the file
70 * @param pfStyle Where to return the DOS2UNIX_STYLE_XXX and
71 * DOS2UNIX_F_XXX flags.
72 * @param pcDosEols Where to return the number of DOS end-of-line
73 * sequences found. Optional.
74 * @param pcUnixEols Where to return the number of UNIX end-of-line
75 * sequences found.
76 */
dos2unix_analyze_file(const char * pszFilename,KU32 * pfStyle,KSIZE * pcDosEols,KSIZE * pcUnixEols)77 int dos2unix_analyze_file(const char *pszFilename, KU32 *pfStyle, KSIZE *pcDosEols, KSIZE *pcUnixEols)
78 {
79 int iRet = 0;
80 int fd = open(pszFilename, O_RDONLY | O_BINARY);
81 if (fd >= 0)
82 {
83 iRet = dos2unix_analyze_fd(fd, pfStyle, pcDosEols, pcUnixEols);
84 close(fd);
85 }
86 else
87 {
88 iRet = errno;
89 *pfStyle = DOS2UNIX_STYLE_NONE;
90 if (pcUnixEols)
91 *pcUnixEols = 0;
92 if (pcDosEols)
93 *pcDosEols = 0;
94 }
95 return iRet;
96 }
97
98 /**
99 * Does a line ending analysis of the given file descriptor.
100 *
101 * @returns 0 on success, errno value on open or read error.
102 * @param fd The file descriptor to analyze. Caller must
103 * place this as the desired position.
104 * @param pfStyle Where to return the DOS2UNIX_STYLE_XXX and
105 * DOS2UNIX_F_XXX flags.
106 * @param pcDosEols Where to return the number of DOS end-of-line
107 * sequences found. Optional.
108 * @param pcUnixEols Where to return the number of UNIX end-of-line
109 * sequences found.
110 */
dos2unix_analyze_fd(int fd,KU32 * pfStyle,KSIZE * pcDosEols,KSIZE * pcUnixEols)111 int dos2unix_analyze_fd(int fd, KU32 *pfStyle, KSIZE *pcDosEols, KSIZE *pcUnixEols)
112 {
113 KSIZE cUnixEols = 0;
114 KSIZE cDosEols = 0;
115 KSIZE cLoneCrs = 0;
116 KBOOL fPendingCr = K_FALSE;
117 int iRet = 0;
118
119 /*
120 * Do the analysis.
121 */
122 *pfStyle = DOS2UNIX_STYLE_NONE;
123 for (;;)
124 {
125 char achBuf[STACK_BUF_SIZE];
126 int cchRead = read(fd, achBuf, sizeof(achBuf));
127 if (cchRead > 0)
128 {
129 int off = 0;
130 if (fPendingCr)
131 {
132 if (achBuf[0] == DOS2UNIX_LF)
133 {
134 off++;
135 cDosEols++;
136 }
137 else
138 cLoneCrs++;
139 fPendingCr = K_FALSE;
140 }
141
142 while (off < cchRead)
143 {
144 char ch = achBuf[off++];
145 if ((unsigned char)ch > (unsigned char)DOS2UNIX_CR)
146 { /* likely */ }
147 else if (ch == DOS2UNIX_CR)
148 {
149 if (off < cchRead && achBuf[off] == DOS2UNIX_CR)
150 cDosEols++;
151 else
152 {
153 fPendingCr = K_TRUE;
154 while (off < cchRead)
155 {
156 ch = achBuf[off++];
157 if (ch != DOS2UNIX_CR)
158 {
159 if (ch == DOS2UNIX_LF)
160 cDosEols++;
161 else
162 cLoneCrs++;
163 fPendingCr = K_FALSE;
164 break;
165 }
166 cLoneCrs++;
167 }
168 }
169 }
170 else if (ch == DOS2UNIX_LF)
171 cUnixEols++;
172 else if (ch == '\0')
173 *pfStyle |= DOS2UNIX_F_BINARY;
174 }
175 }
176 else
177 {
178 if (cchRead < 0)
179 iRet = errno;
180 if (fPendingCr)
181 cLoneCrs++;
182 break;
183 }
184 }
185
186 /*
187 * Set return values.
188 */
189 if (cUnixEols > 0 && cDosEols == 0)
190 *pfStyle |= DOS2UNIX_STYLE_UNIX;
191 else if (cDosEols > 0 && cUnixEols == 0)
192 *pfStyle |= DOS2UNIX_STYLE_DOS;
193 else if (cDosEols != 0 && cUnixEols != 0)
194 *pfStyle |= DOS2UNIX_STYLE_MIXED;
195 if (pcUnixEols)
196 *pcUnixEols = cUnixEols;
197 if (pcDosEols)
198 *pcDosEols = cDosEols;
199
200 return iRet;
201 }
202
203
204 /**
205 * Converts a buffer to unix line (LF) endings.
206 *
207 * @retval K_TRUE if pending CR. The caller must handle this case.
208 * @retval K_FALSE if no pending CR.
209 *
210 * @param pchSrc The input buffer.
211 * @param cchSrc Number of characters to convert from the input
212 * buffer.
213 * @param pchDst The output buffer. This must be at least as big as
214 * the input. It is okay if this overlaps with the
215 * source buffer, as long as this is at the same or a
216 * lower address.
217 * @param pcchDst Where to return the number of characters in the
218 * output buffer.
219 */
dos2unix_convert_to_unix(const char * pchSrc,KSIZE cchSrc,char * pchDst,KSIZE * pcchDst)220 KBOOL dos2unix_convert_to_unix(const char *pchSrc, KSIZE cchSrc, char *pchDst, KSIZE *pcchDst)
221 {
222 KSIZE offDst = 0;
223 while (cchSrc-- > 0)
224 {
225 char ch = *pchSrc++;
226 if ((unsigned char)ch != (unsigned char)DOS2UNIX_CR)
227 pchDst[offDst++] = ch;
228 else if (cchSrc > 0 && *pchSrc == DOS2UNIX_LF)
229 {
230 pchDst[offDst++] = DOS2UNIX_LF;
231 cchSrc--;
232 pchSrc++;
233 }
234 else if (cchSrc == 0)
235 {
236 *pcchDst = offDst;
237 return K_TRUE;
238 }
239 else
240 pchDst[offDst++] = ch;
241 }
242
243 *pcchDst = offDst;
244 return K_FALSE;
245 }
246
247
248 /**
249 * Converts a buffer to DOS (CRLF) endings.
250 *
251 * @retval K_TRUE if pending CR. The caller must handle this case.
252 * @retval K_FALSE if no pending CR.
253 *
254 * @param pchSrc The input buffer.
255 * @param cchSrc Number of characters to convert from the input
256 * buffer.
257 * @param pchDst The output buffer. This must be at least _twice_ as
258 * big as the input. It is okay if the top half of the
259 * buffer overlaps with the source buffer.
260 * @param pcchDst Where to return the number of characters in the
261 * output buffer.
262 */
dos2unix_convert_to_dos(const char * pchSrc,KSIZE cchSrc,char * pchDst,KSIZE * pcchDst)263 KBOOL dos2unix_convert_to_dos(const char *pchSrc, KSIZE cchSrc, char *pchDst, KSIZE *pcchDst)
264 {
265 KSIZE offDst = 0;
266 while (cchSrc-- > 0)
267 {
268 char ch = *pchSrc++;
269 if ((unsigned char)ch > (unsigned char)DOS2UNIX_CR)
270 pchDst[offDst++] = ch;
271 else if (ch == DOS2UNIX_CR)
272 {
273 /* We treat CR kind of like an escape character. */
274 do
275 {
276 if (cchSrc > 0)
277 {
278 pchDst[offDst++] = ch;
279 cchSrc--;
280 ch = *pchSrc++;
281 }
282 else
283 {
284 *pcchDst = offDst;
285 return K_TRUE;
286 }
287 } while (ch == DOS2UNIX_CR);
288 pchDst[offDst++] = ch;
289 }
290 else if (ch == DOS2UNIX_LF)
291 {
292 pchDst[offDst++] = DOS2UNIX_CR;
293 pchDst[offDst++] = DOS2UNIX_LF;
294 }
295 else
296 pchDst[offDst++] = ch;
297 }
298
299 *pcchDst = offDst;
300 return K_FALSE;
301 }
302
303