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 <stdbool.h>
20 #include <stdint.h>
21 #include <stdio.h>
22 #include <stdlib.h>
23 #include <unistd.h>
24 #include <ctype.h>
25 
26 #include <arraylist.h>
27 #include <diff_main.h>
28 
29 #include "diff_internal.h"
30 #include "diff_debug.h"
31 
32 unsigned int
33 diff_atom_hash_update(unsigned int hash, unsigned char atom_byte)
34 {
35 	return hash * 23 + atom_byte;
36 }
37 
38 static int
39 diff_data_atomize_text_lines_fd(struct diff_data *d)
40 {
41 	off_t pos = 0;
42 	const off_t end = pos + d->len;
43 	unsigned int array_size_estimate = d->len / 50;
44 	unsigned int pow2 = 1;
45 	bool ignore_whitespace = (d->diff_flags & DIFF_FLAG_IGNORE_WHITESPACE);
46 
47 	while (array_size_estimate >>= 1)
48 		pow2++;
49 
50 	ARRAYLIST_INIT(d->atoms, 1 << pow2);
51 
52 	if (fseek(d->root->f, 0L, SEEK_SET) == -1)
53 		return errno;
54 
55 	while (pos < end) {
56 		off_t line_end = pos;
57 		unsigned int hash = 0;
58 		unsigned char buf[512];
59 		size_t r, i;
60 		struct diff_atom *atom;
61 		int eol = 0;
62 
63 		while (eol == 0 && line_end < end) {
64 			r = fread(buf, sizeof(char), sizeof(buf), d->root->f);
65 			if (r == 0 && ferror(d->root->f))
66 				return errno;
67 			i = 0;
68 			while (eol == 0 && i < r) {
69 				if (buf[i] != '\r' && buf[i] != '\n') {
70 					if (!ignore_whitespace
71 					    || !isspace(buf[i]))
72 						hash = diff_atom_hash_update(
73 						    hash, buf[i]);
74 					line_end++;
75 				} else
76 					eol = buf[i];
77 				i++;
78 			}
79 		}
80 
81 		/* When not at the end of data, the line ending char ('\r' or
82 		 * '\n') must follow */
83 		if (line_end < end)
84 			line_end++;
85 		/* If that was an '\r', also pull in any following '\n' */
86 		if (line_end < end && eol == '\r') {
87 			if (fseeko(d->root->f, line_end, SEEK_SET) == -1)
88 				return errno;
89 			r = fread(buf, sizeof(char), sizeof(buf), d->root->f);
90 			if (r == 0 && ferror(d->root->f))
91 				return errno;
92 			if (r == 1 && buf[0] == '\n' )
93 				line_end++;
94 		}
95 
96 		/* Record the found line as diff atom */
97 		ARRAYLIST_ADD(atom, d->atoms);
98 		if (!atom)
99 			return ENOMEM;
100 
101 		*atom = (struct diff_atom){
102 			.root = d,
103 			.pos = pos,
104 			.at = NULL,	/* atom data is not memory-mapped */
105 			.len = line_end - pos,
106 			.hash = hash,
107 		};
108 
109 		/* Starting point for next line: */
110 		pos = line_end;
111 		if (fseeko(d->root->f, pos, SEEK_SET) == -1)
112 			return errno;
113 	}
114 
115 	return DIFF_RC_OK;
116 }
117 
118 static int
119 diff_data_atomize_text_lines_mmap(struct diff_data *d)
120 {
121 	const uint8_t *pos = d->data;
122 	const uint8_t *end = pos + d->len;
123 	bool ignore_whitespace = (d->diff_flags & DIFF_FLAG_IGNORE_WHITESPACE);
124 
125 	unsigned int array_size_estimate = d->len / 50;
126 	unsigned int pow2 = 1;
127 	while (array_size_estimate >>= 1)
128 		pow2++;
129 
130 	ARRAYLIST_INIT(d->atoms, 1 << pow2);
131 
132 	while (pos < end) {
133 		const uint8_t *line_end = pos;
134 		unsigned int hash = 0;
135 
136 		while (line_end < end && *line_end != '\r' && *line_end != '\n') {
137 			if (!ignore_whitespace
138 			    || !isspace(*line_end))
139 				hash = hash * 23 + *line_end;
140 			line_end++;
141 		}
142 
143 		/* When not at the end of data, the line ending char ('\r' or
144 		 * '\n') must follow */
145 		if (line_end < end)
146 			line_end++;
147 		/* If that was an '\r', also pull in any following '\n' */
148 		if (line_end < end - 1 && line_end[0] == '\r' &&
149 		    line_end[1] == '\n')
150 			line_end++;
151 
152 		/* Record the found line as diff atom */
153 		struct diff_atom *atom;
154 		ARRAYLIST_ADD(atom, d->atoms);
155 		if (!atom)
156 			return ENOMEM;
157 
158 		*atom = (struct diff_atom){
159 			.root = d,
160 			.pos = (off_t)(pos - d->data),
161 			.at = pos,
162 			.len = line_end - pos,
163 			.hash = hash,
164 		};
165 
166 		/* Starting point for next line: */
167 		pos = line_end;
168 	}
169 
170 	return DIFF_RC_OK;
171 }
172 
173 static int
174 diff_data_atomize_text_lines(struct diff_data *d)
175 {
176 	if (d->data == NULL)
177 		return diff_data_atomize_text_lines_fd(d);
178 	else
179 		return diff_data_atomize_text_lines_mmap(d);
180 }
181 
182 int
183 diff_atomize_text_by_line(void *func_data, struct diff_data *d)
184 {
185 	return diff_data_atomize_text_lines(d);
186 }