inp.c revision 246074 
1/* $FreeBSD: head/usr.bin/patch/inp.c 246074 2013-01-29 17:03:18Z gabor $ */
2/*-
3 *
4 * Copyright 1986, Larry Wall
5 *
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following condition is met:
8 * 1. Redistributions of source code must retain the above copyright notice,
9 * this condition and the following disclaimer.
10 *
11 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND ANY
12 * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
13 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
14 * DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR
15 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
16 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
17 * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
18 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
19 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
20 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
21 * SUCH DAMAGE.
22 *
23 * patch - a program to apply diffs to original files
24 *
25 * -C option added in 1998, original code by Marc Espie, based on FreeBSD
26 * behaviour
27 *
28 * $OpenBSD: inp.c,v 1.34 2006/03/11 19:41:30 otto Exp $
29 */
30
31#include <sys/types.h>
32#include <sys/file.h>
33#include <sys/stat.h>
34#include <sys/mman.h>
35
36#include <ctype.h>
37#include <libgen.h>
38#include <limits.h>
39#include <stddef.h>
40#include <stdio.h>
41#include <stdlib.h>
42#include <string.h>
43#include <unistd.h>
44
45#include "common.h"
46#include "util.h"
47#include "pch.h"
48#include "inp.h"
49
50
51/* Input-file-with-indexable-lines abstract type */
52
53static size_t	i_size;		/* size of the input file */
54static char	*i_womp;	/* plan a buffer for entire file */
55static char	**i_ptr;	/* pointers to lines in i_womp */
56static char	empty_line[] = { '\0' };
57
58static int	tifd = -1;	/* plan b virtual string array */
59static char	*tibuf[2];	/* plan b buffers */
60static LINENUM	tiline[2] = {-1, -1};	/* 1st line in each buffer */
61static LINENUM	lines_per_buf;	/* how many lines per buffer */
62static int	tireclen;	/* length of records in tmp file */
63
64static bool	rev_in_string(const char *);
65static bool	reallocate_lines(size_t *);
66
67/* returns false if insufficient memory */
68static bool	plan_a(const char *);
69
70static void	plan_b(const char *);
71
72/* New patch--prepare to edit another file. */
73
74void
75re_input(void)
76{
77	if (using_plan_a) {
78		free(i_ptr);
79		i_ptr = NULL;
80		if (i_womp != NULL) {
81			munmap(i_womp, i_size);
82			i_womp = NULL;
83		}
84		i_size = 0;
85	} else {
86		using_plan_a = true;	/* maybe the next one is smaller */
87		close(tifd);
88		tifd = -1;
89		free(tibuf[0]);
90		free(tibuf[1]);
91		tibuf[0] = tibuf[1] = NULL;
92		tiline[0] = tiline[1] = -1;
93		tireclen = 0;
94	}
95}
96
97/* Construct the line index, somehow or other. */
98
99void
100scan_input(const char *filename)
101{
102	if (!plan_a(filename))
103		plan_b(filename);
104	if (verbose) {
105		say("Patching file %s using Plan %s...\n", filename,
106		    (using_plan_a ? "A" : "B"));
107	}
108}
109
110static bool
111reallocate_lines(size_t *lines_allocated)
112{
113	char	**p;
114	size_t	new_size;
115
116	new_size = *lines_allocated * 3 / 2;
117	p = realloc(i_ptr, (new_size + 2) * sizeof(char *));
118	if (p == NULL) {	/* shucks, it was a near thing */
119		munmap(i_womp, i_size);
120		i_womp = NULL;
121		free(i_ptr);
122		i_ptr = NULL;
123		*lines_allocated = 0;
124		return false;
125	}
126	*lines_allocated = new_size;
127	i_ptr = p;
128	return true;
129}
130
131/* Try keeping everything in memory. */
132
133static bool
134plan_a(const char *filename)
135{
136	int		ifd, statfailed;
137	char		*p, *s, lbuf[INITLINELEN];
138	struct stat	filestat;
139	ptrdiff_t	sz;
140	size_t		i;
141	size_t		iline, lines_allocated;
142
143#ifdef DEBUGGING
144	if (debug & 8)
145		return false;
146#endif
147
148	if (filename == NULL || *filename == '\0')
149		return false;
150
151	statfailed = stat(filename, &filestat);
152	if (statfailed && ok_to_create_file) {
153		if (verbose)
154			say("(Creating file %s...)\n", filename);
155
156		/*
157		 * in check_patch case, we still display `Creating file' even
158		 * though we're not. The rule is that -C should be as similar
159		 * to normal patch behavior as possible
160		 */
161		if (check_only)
162			return true;
163		makedirs(filename, true);
164		close(creat(filename, 0666));
165		statfailed = stat(filename, &filestat);
166	}
167	if (statfailed && check_only)
168		fatal("%s not found, -C mode, can't probe further\n", filename);
169	/* For nonexistent or read-only files, look for RCS or SCCS versions.  */
170	if (statfailed ||
171	    /* No one can write to it.  */
172	    (filestat.st_mode & 0222) == 0 ||
173	    /* I can't write to it.  */
174	    ((filestat.st_mode & 0022) == 0 && filestat.st_uid != getuid())) {
175		const char	*cs = NULL, *filebase, *filedir;
176		struct stat	cstat;
177		char *tmp_filename1, *tmp_filename2;
178
179		tmp_filename1 = strdup(filename);
180		tmp_filename2 = strdup(filename);
181		if (tmp_filename1 == NULL || tmp_filename2 == NULL)
182			fatal("strdupping filename");
183		filebase = basename(tmp_filename1);
184		filedir = dirname(tmp_filename2);
185
186		/* Leave room in lbuf for the diff command.  */
187		s = lbuf + 20;
188
189#define try(f, a1, a2, a3) \
190	(snprintf(s, buf_size - 20, f, a1, a2, a3), stat(s, &cstat) == 0)
191
192		if (try("%s/RCS/%s%s", filedir, filebase, RCSSUFFIX) ||
193		    try("%s/RCS/%s%s", filedir, filebase, "") ||
194		    try("%s/%s%s", filedir, filebase, RCSSUFFIX)) {
195			snprintf(buf, buf_size, CHECKOUT, filename);
196			snprintf(lbuf, sizeof lbuf, RCSDIFF, filename);
197			cs = "RCS";
198		} else if (try("%s/SCCS/%s%s", filedir, SCCSPREFIX, filebase) ||
199		    try("%s/%s%s", filedir, SCCSPREFIX, filebase)) {
200			snprintf(buf, buf_size, GET, s);
201			snprintf(lbuf, sizeof lbuf, SCCSDIFF, s, filename);
202			cs = "SCCS";
203		} else if (statfailed)
204			fatal("can't find %s\n", filename);
205
206		free(tmp_filename1);
207		free(tmp_filename2);
208
209		/*
210		 * else we can't write to it but it's not under a version
211		 * control system, so just proceed.
212		 */
213		if (cs) {
214			if (!statfailed) {
215				if ((filestat.st_mode & 0222) != 0)
216					/* The owner can write to it.  */
217					fatal("file %s seems to be locked "
218					    "by somebody else under %s\n",
219					    filename, cs);
220				/*
221				 * It might be checked out unlocked.  See if
222				 * it's safe to check out the default version
223				 * locked.
224				 */
225				if (verbose)
226					say("Comparing file %s to default "
227					    "%s version...\n",
228					    filename, cs);
229				if (system(lbuf))
230					fatal("can't check out file %s: "
231					    "differs from default %s version\n",
232					    filename, cs);
233			}
234			if (verbose)
235				say("Checking out file %s from %s...\n",
236				    filename, cs);
237			if (system(buf) || stat(filename, &filestat))
238				fatal("can't check out file %s from %s\n",
239				    filename, cs);
240		}
241	}
242	filemode = filestat.st_mode;
243	if (!S_ISREG(filemode))
244		fatal("%s is not a normal file--can't patch\n", filename);
245	if ((uint64_t)filestat.st_size > SIZE_MAX) {
246		say("block too large to mmap\n");
247		return false;
248	}
249	i_size = (size_t)filestat.st_size;
250	if (out_of_mem) {
251		set_hunkmax();	/* make sure dynamic arrays are allocated */
252		out_of_mem = false;
253		return false;	/* force plan b because plan a bombed */
254	}
255	if ((ifd = open(filename, O_RDONLY)) < 0)
256		pfatal("can't open file %s", filename);
257
258	if (i_size) {
259		i_womp = mmap(NULL, i_size, PROT_READ, MAP_PRIVATE, ifd, 0);
260		if (i_womp == MAP_FAILED) {
261			perror("mmap failed");
262			i_womp = NULL;
263			close(ifd);
264			return false;
265		}
266	} else {
267		i_womp = NULL;
268	}
269
270	close(ifd);
271	if (i_size)
272		madvise(i_womp, i_size, MADV_SEQUENTIAL);
273
274	/* estimate the number of lines */
275	lines_allocated = i_size / 25;
276	if (lines_allocated < 100)
277		lines_allocated = 100;
278
279	if (!reallocate_lines(&lines_allocated))
280		return false;
281
282	/* now scan the buffer and build pointer array */
283	iline = 1;
284	i_ptr[iline] = i_womp;
285	/* test for NUL too, to maintain the behavior of the original code */
286	for (s = i_womp, i = 0; i < i_size && *s != '\0'; s++, i++) {
287		if (*s == '\n') {
288			if (iline == lines_allocated) {
289				if (!reallocate_lines(&lines_allocated))
290					return false;
291			}
292			/* these are NOT NUL terminated */
293			i_ptr[++iline] = s + 1;
294		}
295	}
296	/* if the last line contains no EOL, append one */
297	if (i_size > 0 && i_womp[i_size - 1] != '\n') {
298		last_line_missing_eol = true;
299		/* fix last line */
300		sz = s - i_ptr[iline];
301		p = malloc(sz + 1);
302		if (p == NULL) {
303			free(i_ptr);
304			i_ptr = NULL;
305			munmap(i_womp, i_size);
306			i_womp = NULL;
307			return false;
308		}
309
310		memcpy(p, i_ptr[iline], sz);
311		p[sz] = '\n';
312		i_ptr[iline] = p;
313		/* count the extra line and make it point to some valid mem */
314		i_ptr[++iline] = empty_line;
315	} else
316		last_line_missing_eol = false;
317
318	input_lines = iline - 1;
319
320	/* now check for revision, if any */
321
322	if (revision != NULL) {
323		if (!rev_in_string(i_womp)) {
324			if (force) {
325				if (verbose)
326					say("Warning: this file doesn't appear "
327					    "to be the %s version--patching anyway.\n",
328					    revision);
329			} else if (batch) {
330				fatal("this file doesn't appear to be the "
331				    "%s version--aborting.\n",
332				    revision);
333			} else {
334				ask("This file doesn't appear to be the "
335				    "%s version--patch anyway? [n] ",
336				    revision);
337				if (*buf != 'y')
338					fatal("aborted\n");
339			}
340		} else if (verbose)
341			say("Good.  This file appears to be the %s version.\n",
342			    revision);
343	}
344	return true;		/* plan a will work */
345}
346
347/* Keep (virtually) nothing in memory. */
348
349static void
350plan_b(const char *filename)
351{
352	FILE	*ifp;
353	size_t	i = 0, j, maxlen = 1;
354	char	*p;
355	bool	found_revision = (revision == NULL);
356
357	using_plan_a = false;
358	if ((ifp = fopen(filename, "r")) == NULL)
359		pfatal("can't open file %s", filename);
360	unlink(TMPINNAME);
361	if ((tifd = open(TMPINNAME, O_EXCL | O_CREAT | O_WRONLY, 0666)) < 0)
362		pfatal("can't open file %s", TMPINNAME);
363	while (fgets(buf, buf_size, ifp) != NULL) {
364		if (revision != NULL && !found_revision && rev_in_string(buf))
365			found_revision = true;
366		if ((i = strlen(buf)) > maxlen)
367			maxlen = i;	/* find longest line */
368	}
369	last_line_missing_eol = i > 0 && buf[i - 1] != '\n';
370	if (last_line_missing_eol && maxlen == i)
371		maxlen++;
372
373	if (revision != NULL) {
374		if (!found_revision) {
375			if (force) {
376				if (verbose)
377					say("Warning: this file doesn't appear "
378					    "to be the %s version--patching anyway.\n",
379					    revision);
380			} else if (batch) {
381				fatal("this file doesn't appear to be the "
382				    "%s version--aborting.\n",
383				    revision);
384			} else {
385				ask("This file doesn't appear to be the %s "
386				    "version--patch anyway? [n] ",
387				    revision);
388				if (*buf != 'y')
389					fatal("aborted\n");
390			}
391		} else if (verbose)
392			say("Good.  This file appears to be the %s version.\n",
393			    revision);
394	}
395	fseek(ifp, 0L, SEEK_SET);	/* rewind file */
396	lines_per_buf = BUFFERSIZE / maxlen;
397	tireclen = maxlen;
398	tibuf[0] = malloc(BUFFERSIZE + 1);
399	if (tibuf[0] == NULL)
400		fatal("out of memory\n");
401	tibuf[1] = malloc(BUFFERSIZE + 1);
402	if (tibuf[1] == NULL)
403		fatal("out of memory\n");
404	for (i = 1;; i++) {
405		p = tibuf[0] + maxlen * (i % lines_per_buf);
406		if (i % lines_per_buf == 0)	/* new block */
407			if (write(tifd, tibuf[0], BUFFERSIZE) < BUFFERSIZE)
408				pfatal("can't write temp file");
409		if (fgets(p, maxlen + 1, ifp) == NULL) {
410			input_lines = i - 1;
411			if (i % lines_per_buf != 0)
412				if (write(tifd, tibuf[0], BUFFERSIZE) < BUFFERSIZE)
413					pfatal("can't write temp file");
414			break;
415		}
416		j = strlen(p);
417		/* These are '\n' terminated strings, so no need to add a NUL */
418		if (j == 0 || p[j - 1] != '\n')
419			p[j] = '\n';
420	}
421	fclose(ifp);
422	close(tifd);
423	if ((tifd = open(TMPINNAME, O_RDONLY)) < 0)
424		pfatal("can't reopen file %s", TMPINNAME);
425}
426
427/*
428 * Fetch a line from the input file, \n terminated, not necessarily \0.
429 */
430char *
431ifetch(LINENUM line, int whichbuf)
432{
433	if (line < 1 || line > input_lines) {
434		if (warn_on_invalid_line) {
435			say("No such line %ld in input file, ignoring\n", line);
436			warn_on_invalid_line = false;
437		}
438		return NULL;
439	}
440	if (using_plan_a)
441		return i_ptr[line];
442	else {
443		LINENUM	offline = line % lines_per_buf;
444		LINENUM	baseline = line - offline;
445
446		if (tiline[0] == baseline)
447			whichbuf = 0;
448		else if (tiline[1] == baseline)
449			whichbuf = 1;
450		else {
451			tiline[whichbuf] = baseline;
452
453			if (lseek(tifd, (off_t) (baseline / lines_per_buf *
454			    BUFFERSIZE), SEEK_SET) < 0)
455				pfatal("cannot seek in the temporary input file");
456
457			if (read(tifd, tibuf[whichbuf], BUFFERSIZE) < 0)
458				pfatal("error reading tmp file %s", TMPINNAME);
459		}
460		return tibuf[whichbuf] + (tireclen * offline);
461	}
462}
463
464/*
465 * True if the string argument contains the revision number we want.
466 */
467static bool
468rev_in_string(const char *string)
469{
470	const char	*s;
471	size_t		patlen;
472
473	if (revision == NULL)
474		return true;
475	patlen = strlen(revision);
476	if (strnEQ(string, revision, patlen) && isspace((unsigned char)string[patlen]))
477		return true;
478	for (s = string; *s; s++) {
479		if (isspace((unsigned char)*s) && strnEQ(s + 1, revision, patlen) &&
480		    isspace((unsigned char)s[patlen + 1])) {
481			return true;
482		}
483	}
484	return false;
485}
486