1 /* Split source by line breaks, and calculate a simplistic checksum. */
2 /*
3  * Copyright (c) 2020 Neels Hofmeyr <neels@hofmeyr.de>
4  *
5  * Permission to use, copy, modify, and distribute this software for any
6  * purpose with or without fee is hereby granted, provided that the above
7  * copyright notice and this permission notice appear in all copies.
8  *
9  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
10  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
11  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
12  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
13  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
14  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
15  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
16  */
17 
18 #include <errno.h>
19 #include <inttypes.h>
20 #include <stdbool.h>
21 #include <stdio.h>
22 #include <stdlib.h>
23 #include <unistd.h>
24 
25 #include <arraylist.h>
26 #include <diff_main.h>
27 
28 #include "diff_internal.h"
29 #include "diff_debug.h"
30 
31 static int
32 diff_data_atomize_text_lines_fd(struct diff_data *d)
33 {
34 	off_t pos = 0;
35 	const off_t end = pos + d->len;
36 	unsigned int array_size_estimate = d->len / 50;
37 	unsigned int pow2 = 1;
38 
39 	while (array_size_estimate >>= 1)
40 		pow2++;
41 
42 	ARRAYLIST_INIT(d->atoms, 1 << pow2);
43 
44 	if (fseek(d->root->f, 0L, SEEK_SET) == -1)
45 		return errno;
46 
47 	while (pos < end) {
48 		off_t line_end = pos;
49 		unsigned int hash = 0;
50 		unsigned char buf[512];
51 		size_t r, i;
52 		struct diff_atom *atom;
53 		int eol = 0;
54 
55 		while (eol == 0 && line_end < end) {
56 			r = fread(buf, sizeof(char), sizeof(buf), d->root->f);
57 			if (r == 0 && ferror(d->root->f))
58 				return errno;
59 			i = 0;
60 			while (eol == 0 && i < r) {
61 				if (buf[i] != '\r' && buf[i] != '\n') {
62 					hash = hash * 23 + buf[i];
63 					line_end++;
64 				} else
65 					eol = buf[i];
66 				i++;
67 			}
68 		}
69 
70 		/* When not at the end of data, the line ending char ('\r' or
71 		 * '\n') must follow */
72 		if (line_end < end)
73 			line_end++;
74 		/* If that was an '\r', also pull in any following '\n' */
75 		if (line_end < end && eol == '\r') {
76 			if (fseeko(d->root->f, line_end, SEEK_SET) == -1)
77 				return errno;
78 			r = fread(buf, sizeof(char), sizeof(buf), d->root->f);
79 			if (r == 0 && ferror(d->root->f))
80 				return errno;
81 			if (r == 1 && buf[0] == '\n' )
82 				line_end++;
83 		}
84 
85 		/* Record the found line as diff atom */
86 		ARRAYLIST_ADD(atom, d->atoms);
87 		if (!atom)
88 			return ENOMEM;
89 
90 		*atom = (struct diff_atom){
91 			.d = d,
92 			.pos = pos,
93 			.at = NULL,	/* atom data is not memory-mapped */
94 			.len = line_end - pos,
95 			.hash = hash,
96 		};
97 
98 		/* Starting point for next line: */
99 		pos = line_end;
100 		if (fseeko(d->root->f, pos, SEEK_SET) == -1)
101 			return errno;
102 	}
103 
104 	return DIFF_RC_OK;
105 }
106 
107 static int
108 diff_data_atomize_text_lines_mmap(struct diff_data *d)
109 {
110 	const uint8_t *pos = d->data;
111 	const uint8_t *end = pos + d->len;
112 
113 	unsigned int array_size_estimate = d->len / 50;
114 	unsigned int pow2 = 1;
115 	while (array_size_estimate >>= 1)
116 		pow2++;
117 
118 	ARRAYLIST_INIT(d->atoms, 1 << pow2);
119 
120 	while (pos < end) {
121 		const uint8_t *line_end = pos;
122 		unsigned int hash = 0;
123 
124 		while (line_end < end && *line_end != '\r' && *line_end != '\n') {
125 			hash = hash * 23 + *line_end;
126 			line_end++;
127 		}
128 
129 		/* When not at the end of data, the line ending char ('\r' or
130 		 * '\n') must follow */
131 		if (line_end < end)
132 			line_end++;
133 		/* If that was an '\r', also pull in any following '\n' */
134 		if (line_end[0] == '\r'
135 		    && line_end < end && line_end[1] == '\n')
136 			line_end++;
137 
138 		/* Record the found line as diff atom */
139 		struct diff_atom *atom;
140 		ARRAYLIST_ADD(atom, d->atoms);
141 		if (!atom)
142 			return ENOMEM;
143 
144 		*atom = (struct diff_atom){
145 			.d = d,
146 			.pos = (off_t)(pos - d->data),
147 			.at = pos,
148 			.len = line_end - pos,
149 			.hash = hash,
150 		};
151 
152 		/* Starting point for next line: */
153 		pos = line_end;
154 	}
155 
156 	return DIFF_RC_OK;
157 }
158 
159 static int
160 diff_data_atomize_text_lines(struct diff_data *d)
161 {
162 	if (d->data == NULL)
163 		return diff_data_atomize_text_lines_fd(d);
164 	else
165 		return diff_data_atomize_text_lines_mmap(d);
166 }
167 
168 int
169 diff_atomize_text_by_line(void *func_data, struct diff_data *left,
170 			  struct diff_data *right)
171 {
172 	int rc;
173 	rc = diff_data_atomize_text_lines(left);
174 	if (rc != DIFF_RC_OK)
175 		return rc;
176 	return diff_data_atomize_text_lines(right);
177 }