1/***********************************************************************
2*                                                                      *
3*               This software is part of the ast package               *
4*          Copyright (c) 1992-2012 AT&T Intellectual Property          *
5*                      and is licensed under the                       *
6*                 Eclipse Public License, Version 1.0                  *
7*                    by AT&T Intellectual Property                     *
8*                                                                      *
9*                A copy of the License is available at                 *
10*          http://www.eclipse.org/org/documents/epl-v10.html           *
11*         (with md5 checksum b35adb5213ca9657e911e9befb180842)         *
12*                                                                      *
13*              Information and Software Systems Research               *
14*                            AT&T Research                             *
15*                           Florham Park NJ                            *
16*                                                                      *
17*                 Glenn Fowler <gsf@research.att.com>                  *
18*                  David Korn <dgk@research.att.com>                   *
19*                                                                      *
20***********************************************************************/
21#pragma prototyped
22/*
23 * David Korn
24 * Glenn Fowler
25 * AT&T Bell Laboratories
26 *
27 * cmp
28 */
29
30static const char usage[] =
31"[-?\n@(#)$Id: cmp (AT&T Research) 2010-04-11 $\n]"
32USAGE_LICENSE
33"[+NAME?cmp - compare two files]"
34"[+DESCRIPTION?\bcmp\b compares two files \afile1\a and \afile2\a. "
35    "\bcmp\b writes no output if the files are the same. By default, if the "
36    "files differ, the byte and line number at which the first difference "
37    "occurred are written to standard output. Bytes and lines are numbered "
38    "beginning with 1.]"
39"[+?If \askip1\a or \askip2\a are specified, or the \b-i\b option is "
40    "specified, initial bytes of the corresponding file are skipped before "
41    "beginning the compare. The skip values are in bytes or can have a "
42    "suffix of \bk\b for kilobytes or \bm\b for megabytes.]"
43"[+?If either \afile1\a or \afiles2\a is \b-\b, \bcmp\b uses standard "
44    "input starting at the current location.]"
45"[b:print-bytes?Print differing bytes as 3 digit octal values.]"
46"[c:print-chars?Print differing bytes as follows: non-space printable "
47    "characters as themselves; space and control characters as \b^\b "
48    "followed by a letter of the alphabet; and characters with the high bit "
49    "set as the lower 7 bit character prefixed by \bM^\b for 7 bit space and "
50    "non-printable characters and \bM-\b for all other characters. If the 7 "
51    "bit character encoding is not ASCII then the characters are converted "
52    "to ASCII to determine \ahigh bit set\a, and if set it is cleared and "
53    "converted back to the native encoding. Multibyte characters in the "
54    "current locale are treated as printable characters.]"
55"[d:differences?Print at most \adifferences\a differences using "
56    "\b--verbose\b output format. \b--differences=0\b is equivalent to "
57    "\b--silent\b.]#[differences]"
58"[i:ignore-initial|skip?Skip the the first \askip1\a bytes in \afile1\a "
59    "and the first \askip2\a bytes in \afile2\a. If \askip2\a is omitted "
60    "then \askip1\a is used.]:[skip1[::skip2]]:=0::0]"
61"[l:verbose?Write the decimal byte number and the differing bytes (in "
62    "octal) for each difference.]"
63"[n:count|bytes?Compare at most \acount\a bytes.]#[count]"
64"[s:quiet|silent?Write nothing for differing files; return non-zero exit "
65    "status only.]"
66"\n"
67"\nfile1 file2 [skip1 [skip2]]\n"
68"\n"
69"[+EXIT STATUS?]"
70    "{"
71        "[+0?The files or portions compared are identical.]"
72        "[+1?The files are different.]"
73        "[+>1?An error occurred.]"
74    "}"
75"[+SEE ALSO?\bcomm\b(1), \bdiff\b(1), \bcat\b(1)]"
76;
77
78#include <cmd.h>
79#include <ls.h>
80#include <ctype.h>
81#include <ccode.h>
82
83#define CMP_VERBOSE	0x01
84#define CMP_SILENT	0x02
85#define CMP_CHARS	0x04
86#define CMP_BYTES	0x08
87
88static void
89pretty(Sfio_t *out, int o, int delim, int flags)
90{
91	int	c;
92	int	m;
93	char*	s;
94	char	buf[10];
95
96	s = buf;
97	if ((flags & CMP_BYTES) || !(flags & CMP_CHARS))
98	{
99		*s++ = ' ';
100		if ((flags & CMP_CHARS) && delim != -1)
101			*s++ = ' ';
102		*s++ = '0' + ((o >> 6) & 07);
103		*s++ = '0' + ((o >> 3) & 07);
104		*s++ = '0' + (o & 07);
105	}
106	if (flags & CMP_CHARS)
107	{
108		*s++ = ' ';
109		c = ccmapc(o, CC_NATIVE, CC_ASCII);
110		if (c & 0x80)
111		{
112			m = 1;
113			*s++ = 'M';
114			c &= 0x7f;
115			o = ccmapc(c, CC_ASCII, CC_NATIVE);
116		}
117		else
118			m = 0;
119		if (isspace(o) || !isprint(o))
120		{
121			if (!m)
122				*s++ = ' ';
123			*s++ = '^';
124			c ^= 0x40;
125			o = ccmapc(c, CC_ASCII, CC_NATIVE);
126		}
127		else if (m)
128			*s++ = '-';
129		else
130		{
131			*s++ = ' ';
132			*s++ = ' ';
133		}
134		*s++ = o;
135	}
136	*s = 0;
137	sfputr(out, buf, delim);
138}
139
140/*
141 * compare two files
142 */
143
144static int
145cmp(const char* file1, Sfio_t* f1, const char* file2, Sfio_t* f2, int flags, Sfoff_t count, Sfoff_t differences)
146{
147	register int		c1;
148	register int		c2;
149	register unsigned char*	p1 = 0;
150	register unsigned char*	p2 = 0;
151	register Sfoff_t	lines = 1;
152	register unsigned char*	e1 = 0;
153	register unsigned char*	e2 = 0;
154	Sfoff_t			pos = 0;
155	int			n1 = 0;
156	int			ret = 0;
157	unsigned char*		last;
158
159	for (;;)
160	{
161		if ((c1 = e1 - p1) <= 0)
162		{
163			if (count > 0 && !(count -= n1))
164				return ret;
165			if (!(p1 = (unsigned char*)sfreserve(f1, SF_UNBOUND, 0)) || (c1 = sfvalue(f1)) <= 0)
166			{
167				if ((e2 - p2) > 0 || sfreserve(f2, SF_UNBOUND, 0) && sfvalue(f2) > 0)
168				{
169					ret = 1;
170					if (!(flags & CMP_SILENT))
171						error(ERROR_exit(1), "EOF on %s", file1);
172				}
173				return ret;
174			}
175			if (count > 0 && c1 > count)
176				c1 = (int)count;
177			e1 = p1 + c1;
178			n1 = c1;
179		}
180		if ((c2 = e2 - p2) <= 0)
181		{
182			if (!(p2 = (unsigned char*)sfreserve(f2, SF_UNBOUND, 0)) || (c2 = sfvalue(f2)) <= 0)
183			{
184				if (!(flags & CMP_SILENT))
185					error(ERROR_exit(1), "EOF on %s", file2);
186				return 1;
187			}
188			e2 = p2 + c2;
189		}
190		if (c1 > c2)
191			c1 = c2;
192		pos += c1;
193		if (flags & CMP_SILENT)
194		{
195			if (memcmp(p1, p2, c1))
196				return 1;
197			p1 += c1;
198			p2 += c1;
199		}
200		else
201		{
202			last = p1 + c1;
203			while (p1 < last)
204			{
205				if ((c1 = *p1++) != *p2++)
206				{
207					if (differences >= 0)
208					{
209						if (!differences)
210							return 1;
211						differences--;
212					}
213#if 0
214					if (!flags)
215						sfprintf(sfstdout, "%s %s differ: char %I*d, line %I*u\n", file1, file2, sizeof(pos), pos - (last - p1), sizeof(lines), lines);
216					else
217					{
218						sfprintf(sfstdout, "%6I*d", sizeof(pos), pos - (last - p1));
219						pretty(sfstdout, c1, -1, flags);
220						pretty(sfstdout, *(p2-1), '\n', flags);
221					}
222#else
223					if (flags & CMP_VERBOSE)
224						sfprintf(sfstdout, "%6I*d", sizeof(pos), pos - (last - p1));
225					else
226						sfprintf(sfstdout, "%s %s differ: char %I*d, line %I*u", file1, file2, sizeof(pos), pos - (last - p1), sizeof(lines), lines);
227					if (flags & (CMP_BYTES|CMP_CHARS|CMP_VERBOSE))
228					{
229						sfputc(sfstdout, (flags & CMP_VERBOSE) ? ' ' : ',');
230						pretty(sfstdout, c1, -1, flags);
231						pretty(sfstdout, *(p2-1), '\n', flags);
232					}
233					else
234						sfputc(sfstdout, '\n');
235#endif
236					if (!differences || differences < 0 && !(flags & CMP_VERBOSE))
237						return 1;
238					ret = 1;
239				}
240				if (c1 == '\n')
241					lines++;
242			}
243		}
244	}
245}
246
247int
248b_cmp(int argc, register char** argv, Shbltin_t* context)
249{
250	char*		s;
251	char*		e;
252	char*		file1;
253	char*		file2;
254	int		n;
255	struct stat	s1;
256	struct stat	s2;
257
258	Sfio_t*		f1 = 0;
259	Sfio_t*		f2 = 0;
260	Sfoff_t		o1 = 0;
261	Sfoff_t		o2 = 0;
262	Sfoff_t		count = -1;
263	Sfoff_t		differences = -1;
264	int		flags = 0;
265
266	NoP(argc);
267	cmdinit(argc, argv, context, ERROR_CATALOG, 0);
268	for (;;)
269	{
270		switch (optget(argv, usage))
271		{
272		case 'b':
273			flags |= CMP_BYTES;
274			continue;
275		case 'c':
276			flags |= CMP_CHARS;
277			continue;
278		case 'd':
279			flags |= CMP_VERBOSE;
280			differences = opt_info.number;
281			continue;
282		case 'i':
283			o1 = strtoll(opt_info.arg, &e, 0);
284			if (*e == ':')
285				o2 = strtoll(e + 1, &e, 0);
286			else
287				o2 = o1;
288			if (*e)
289			{
290				error(2, "%s: skip1:skip2 expected", opt_info.arg);
291				break;
292			}
293			continue;
294		case 'l':
295			flags |= CMP_VERBOSE;
296			continue;
297		case 'n':
298			count = opt_info.number;
299			continue;
300		case 's':
301			flags |= CMP_SILENT;
302			continue;
303		case ':':
304			error(2, "%s", opt_info.arg);
305			break;
306		case '?':
307			error(ERROR_usage(2), "%s", opt_info.arg);
308			break;
309		}
310		break;
311	}
312	argv += opt_info.index;
313	if (error_info.errors || !(file1 = *argv++) || !(file2 = *argv++))
314		error(ERROR_usage(2), "%s", optusage(NiL));
315	n = 2;
316	if (streq(file1, "-"))
317		f1 = sfstdin;
318	else if (!(f1 = sfopen(NiL, file1, "r")))
319	{
320		if (!(flags & CMP_SILENT))
321			error(ERROR_system(0), "%s: cannot open", file1);
322		goto done;
323	}
324	if (streq(file2, "-"))
325		f2 = sfstdin;
326	else if (!(f2 = sfopen(NiL, file2, "r")))
327	{
328		if (!(flags & CMP_SILENT))
329			error(ERROR_system(0), "%s: cannot open", file2);
330		goto done;
331	}
332	if (s = *argv++)
333	{
334		o1 = strtoll(s, &e, 0);
335		if (*e)
336		{
337			error(ERROR_exit(0), "%s: %s: invalid skip", file1, s);
338			goto done;
339		}
340		if (s = *argv++)
341		{
342			o2 = strtoll(s, &e, 0);
343			if (*e)
344			{
345				error(ERROR_exit(0), "%s: %s: invalid skip", file2, s);
346				goto done;
347			}
348		}
349		if (*argv)
350		{
351			error(ERROR_usage(0), "%s", optusage(NiL));
352			goto done;
353		}
354	}
355	if (o1 && sfseek(f1, o1, SEEK_SET) != o1)
356	{
357		if (!(flags & CMP_SILENT))
358			error(ERROR_exit(0), "EOF on %s", file1);
359		n = 1;
360		goto done;
361	}
362	if (o2 && sfseek(f2, o2, SEEK_SET) != o2)
363	{
364		if (!(flags & CMP_SILENT))
365			error(ERROR_exit(0), "EOF on %s", file2);
366		n = 1;
367		goto done;
368	}
369	if (fstat(sffileno(f1), &s1))
370		error(ERROR_system(0), "%s: cannot stat", file1);
371	else if (fstat(sffileno(f2), &s2))
372		error(ERROR_system(0), "%s: cannot stat", file1);
373	else if (s1.st_ino == s2.st_ino && s1.st_dev == s2.st_dev && o1 == o2)
374		n = 0;
375	else
376		n = ((flags & CMP_SILENT) && S_ISREG(s1.st_mode) && S_ISREG(s2.st_mode) && (s1.st_size - o1) != (s2.st_size - o2)) ? 1 : cmp(file1, f1, file2, f2, flags, count, differences);
377 done:
378	if (f1 && f1 != sfstdin)
379		sfclose(f1);
380	if (f2 && f2 != sfstdin)
381		sfclose(f2);
382	return n;
383}
384