Rev 14 | Only display areas with differences | Regard whitespace | Details | Blame | Last modification | View Log | RSS feed
Rev 14 | Rev 15 | ||
---|---|---|---|
1 | #!/usr/bin/php |
1 | #!/usr/bin/php |
2 | <?php |
2 | <?php |
3 | 3 | ||
4 | /* |
4 | /* |
5 | Copyright 2020 Daniel Marschall, ViaThinkSoft |
5 | Copyright 2020-2022 Daniel Marschall, ViaThinkSoft |
6 | 6 | ||
7 | Licensed under the Apache License, Version 2.0 (the "License"); |
7 | Licensed under the Apache License, Version 2.0 (the "License"); |
8 | you may not use this file except in compliance with the License. |
8 | you may not use this file except in compliance with the License. |
9 | You may obtain a copy of the License at |
9 | You may obtain a copy of the License at |
10 | 10 | ||
11 | http://www.apache.org/licenses/LICENSE-2.0 |
11 | http://www.apache.org/licenses/LICENSE-2.0 |
12 | 12 | ||
13 | Unless required by applicable law or agreed to in writing, software |
13 | Unless required by applicable law or agreed to in writing, software |
14 | distributed under the License is distributed on an "AS IS" BASIS, |
14 | distributed under the License is distributed on an "AS IS" BASIS, |
15 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
15 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
16 | See the License for the specific language governing permissions and |
16 | See the License for the specific language governing permissions and |
17 | limitations under the License. |
17 | limitations under the License. |
18 | */ |
18 | */ |
19 | 19 | ||
20 | // TODO: make use of STDERR and return different exit codes |
- | |
21 | // TODO: On Windows file systems, accept file names case insensitively |
20 | // TODO: On Windows file systems, accept file names case insensitively |
22 | 21 | ||
23 | function utf8_normalize($str) { |
22 | function utf8_normalize($str) { |
24 | // This helps to handle decomposite Unicode endpoints (E.g. German Umlauts have different representations) |
23 | // This helps to handle decomposite Unicode endpoints (E.g. German Umlauts have different representations) |
25 | // Requires php-intl |
24 | // Requires php-intl |
26 | if (!class_exists('Normalizer')) return $str; |
25 | if (!class_exists('Normalizer')) return $str; |
27 | return Normalizer::normalize($str); |
26 | return Normalizer::normalize($str); |
28 | } |
27 | } |
29 | 28 | ||
30 | function convertToUTF8($str) { |
29 | function convertToUTF8($str) { |
31 | $enc = mb_detect_encoding($str); |
30 | $enc = mb_detect_encoding($str); |
32 | if ($enc && $enc != 'UTF-8') { |
31 | if ($enc && $enc != 'UTF-8') { |
33 | return iconv($enc, 'UTF-8', $str); |
32 | return iconv($enc, 'UTF-8', $str); |
34 | } else { |
33 | } else { |
35 | return $str; |
34 | return $str; |
36 | } |
35 | } |
37 | } |
36 | } |
38 | 37 | ||
39 | function testmd5($file) { |
38 | function testmd5($file) { |
40 | // TODO: warn if an entry is multiple times (with different checksums) in a single file |
39 | // TODO: warn if an entry is multiple times (with different checksums) in a single file |
41 | if (!file_exists($file)) { |
40 | if (!file_exists($file)) { |
42 | echo "ERROR: File $file does not exist.\n"; |
41 | fwrite(STDERR, "ERROR: File $file does not exist.\n"); |
43 | return; |
42 | return; |
44 | } |
43 | } |
45 | 44 | ||
46 | $files_checked = array(); |
45 | $files_checked = array(); |
47 | 46 | ||
48 | $lines = file($file); |
47 | $lines = file($file); |
49 | $is_first_line = true; |
48 | $is_first_line = true; |
50 | $force_utf8 = false; |
49 | $force_utf8 = false; |
51 | foreach ($lines as $line) { |
50 | foreach ($lines as $line) { |
52 | if ($is_first_line) { |
51 | if ($is_first_line) { |
53 | $tmp = 0; |
52 | $tmp = 0; |
54 | $line = str_replace("\xEF\xBB\xBF",'',$line,$tmp); |
53 | $line = str_replace("\xEF\xBB\xBF",'',$line,$tmp); |
55 | if ($tmp > 0) $force_utf8 = true; |
54 | if ($tmp > 0) $force_utf8 = true; |
56 | $is_first_line = false; |
55 | $is_first_line = false; |
57 | } |
56 | } |
58 | if (!$force_utf8) $line = convertToUTF8($line); |
57 | if (!$force_utf8) $line = convertToUTF8($line); |
59 | 58 | ||
60 | if (substr(trim($line),0,1) == ';') continue; |
59 | if (substr(trim($line),0,1) == ';') continue; |
61 | 60 | ||
62 | $line = trim($line); |
61 | $line = trim($line); |
63 | if ($line == '') continue; |
62 | if ($line == '') continue; |
64 | $line = str_replace('*', ' ', $line); |
63 | $line = str_replace('*', ' ', $line); |
65 | $line = str_replace("\t", ' ', $line); |
64 | $line = str_replace("\t", ' ', $line); |
66 | list($checksum, $origname) = explode(' ', $line, 2); |
65 | list($checksum, $origname) = explode(' ', $line, 2); |
67 | $origname = dirname($file) . '/' . trim($origname); |
66 | $origname = dirname($file) . '/' . trim($origname); |
68 | $checksum = trim($checksum); |
67 | $checksum = trim($checksum); |
69 | if (!file_exists($origname)) { |
68 | if (!file_exists($origname)) { |
70 | echo "WARNING: File vanished : $origname\n"; |
69 | fwrite(STDERR, "WARNING: File vanished : $origname\n"); |
71 | } else { |
70 | } else { |
72 | if (is_file($origname)) { |
71 | if (is_file($origname)) { |
73 | $checksum2 = md5_file($origname); |
72 | $checksum2 = md5_file($origname); |
74 | if (strtolower($checksum) != strtolower($checksum2)) { |
73 | if (strtolower($checksum) != strtolower($checksum2)) { |
75 | echo "CHECKSUM FAIL: $origname (expected $checksum, but is $checksum2)\n"; |
74 | fwrite(STDERR, "CHECKSUM FAIL: $origname (expected $checksum, but is $checksum2)\n"); |
76 | } else { |
75 | } else { |
77 | global $show_verbose; |
76 | global $show_verbose; |
78 | if ($show_verbose) echo "OK: $origname\n"; |
77 | if ($show_verbose) echo "OK: $origname\n"; |
79 | } |
78 | } |
80 | } else { |
79 | } else { |
81 | // For some reason, some files on a NTFS volume are "FIFO" pipe files?! |
80 | // For some reason, some files on a NTFS volume are "FIFO" pipe files?! |
82 | echo "Warning: $origname is not a regular file!\n"; |
81 | fwrite(STDERR, "Warning: $origname is not a regular file!\n"); |
83 | } |
82 | } |
84 | } |
83 | } |
85 | 84 | ||
86 | $origname = utf8_normalize(basename($origname)); |
85 | $origname = utf8_normalize(basename($origname)); |
87 | $files_checked[] = dirname($file) . '/' . $origname; |
86 | $files_checked[] = dirname($file) . '/' . $origname; |
88 | } |
87 | } |
89 | 88 | ||
90 | // Now check if files have vanished! |
89 | // Now check if files have vanished! |
91 | $directory = dirname($file); |
90 | $directory = dirname($file); |
92 | $sd = @scandir($directory); |
91 | $sd = @scandir($directory); |
93 | if ($sd === false) { |
92 | if ($sd === false) { |
94 | echo "Error: Cannot scan directory $directory\n"; |
93 | fwrite(STDERR, "Error: Cannot scan directory $directory\n"); |
95 | } else { |
94 | } else { |
96 | foreach ($sd as $file) { |
95 | foreach ($sd as $file) { |
97 | if ($file === '.') continue; |
96 | if ($file === '.') continue; |
98 | if ($file === '..') continue; |
97 | if ($file === '..') continue; |
- | 98 | if (substr($file,0,1) === '.') continue; |
|
99 | if (strtolower($file) === 'thumbs.db') continue; |
99 | if (strtolower($file) === 'thumbs.db') continue; |
100 | if (strtolower(substr($file, -4)) === '.md5') continue; |
100 | if (strtolower(substr($file, -4)) === '.md5') continue; |
101 | if (strtolower(substr($file, -4)) === '.sfv') continue; |
101 | if (strtolower(substr($file, -4)) === '.sfv') continue; |
102 | $fullpath = $directory . '/' . $file; |
102 | $fullpath = $directory . '/' . $file; |
103 | if (!is_dir($fullpath)) { |
103 | if (!is_dir($fullpath)) { |
104 | $fullpath = utf8_normalize($fullpath); |
104 | $fullpath = utf8_normalize($fullpath); |
105 | if (!in_array($fullpath,$files_checked)) { |
105 | if (!in_array($fullpath,$files_checked)) { |
106 | echo "Warning: File not in SFV checksum file: $fullpath\n"; |
106 | fwrite(STDERR, "Warning: File not in SFV checksum file: $fullpath\n"); |
107 | } |
107 | } |
108 | } |
108 | } |
109 | } |
109 | } |
110 | } |
110 | } |
111 | } |
111 | } |
112 | 112 | ||
113 | function _rec($directory) { |
113 | function _rec($directory) { |
114 | $directory = rtrim($directory, '/\\'); |
114 | $directory = rtrim($directory, '/\\'); |
115 | 115 | ||
116 | if (!is_dir($directory)) { |
116 | if (!is_dir($directory)) { |
117 | exit("Invalid directory path $directory\n"); |
117 | fwrite(STDERR, "Invalid directory path $directory\n"); |
- | 118 | return false; |
|
118 | } |
119 | } |
119 | 120 | ||
120 | if ($dont_add_files = count(glob("$directory/*.md5")) == 0) { |
121 | if ($dont_add_files = count(glob("$directory/*.md5")) == 0) { |
121 | global $show_verbose; |
122 | global $show_verbose; |
122 | if ($show_verbose) echo "Directory $directory has no MD5 file. Skipping.\n"; |
123 | if ($show_verbose) echo "Directory $directory has no MD5 file. Skipping.\n"; |
123 | } else { |
124 | } else { |
124 | $out = array(); |
125 | $out = array(); |
125 | 126 | ||
126 | global $show_verbose; |
127 | global $show_verbose; |
127 | if ($show_verbose) echo "Check directory $directory\n"; |
128 | if ($show_verbose) echo "Check directory $directory\n"; |
128 | $md5files = glob($directory.'/*.md5'); |
129 | $md5files = glob($directory.'/*.md5'); |
129 | foreach ($md5files as $md5file) { |
130 | foreach ($md5files as $md5file) { |
130 | testmd5($md5file); |
131 | testmd5($md5file); |
131 | } |
132 | } |
132 | } |
133 | } |
133 | 134 | ||
134 | $sd = @scandir($directory); |
135 | $sd = @scandir($directory); |
135 | if ($sd === false) { |
136 | if ($sd === false) { |
136 | echo "Error: Cannot scan directory $directory\n"; |
137 | fwrite(STDERR, "Error: Cannot scan directory $directory\n"); |
137 | return; |
138 | return false; |
138 | } |
139 | } |
139 | 140 | ||
140 | foreach ($sd as $file) { |
141 | foreach ($sd as $file) { |
141 | if ($file !== '.' && $file !== '..') { |
142 | if ($file !== '.' && $file !== '..') { |
142 | $file = $directory . '/' . $file; |
143 | $file = $directory . '/' . $file; |
143 | if (is_dir($file)) { |
144 | if (is_dir($file)) { |
144 | _rec($file); |
145 | _rec($file); |
145 | } |
146 | } |
146 | } |
147 | } |
147 | } |
148 | } |
- | 149 | ||
- | 150 | return true; |
|
148 | } |
151 | } |
149 | 152 | ||
150 | 153 | ||
151 | # --- |
154 | # --- |
152 | 155 | ||
153 | $show_verbose = false; |
156 | $show_verbose = false; |
154 | $dir = ''; |
157 | $dirs = array(); |
155 | 158 | ||
156 | for ($i=1; $i<$argc; $i++) { |
159 | for ($i=1; $i<$argc; $i++) { |
157 | if ($argv[$i] == '-v') { |
160 | if ($argv[$i] == '-v') { |
158 | $show_verbose = true; |
161 | $show_verbose = true; |
159 | } else { |
162 | } else { |
160 | $dir = $argv[$i]; |
163 | $dirs[] = $argv[$i]; |
161 | } |
164 | } |
162 | } |
165 | } |
163 | 166 | ||
164 | if (empty($dir)) { |
167 | if (count($dirs) == 0) { |
165 | echo "Syntax: $argv[0] [-v] <directory>\n"; |
168 | echo "Syntax: $argv[0] [-v] <directory> [<directory> [...]]\n"; |
166 | exit(2); |
169 | exit(2); |
167 | } |
170 | } |
168 | 171 | ||
- | 172 | $res = 0; |
|
169 | if (!is_dir($dir)) { |
173 | foreach ($dirs as $dir) { |
170 | echo "Directory not found\n"; |
174 | if (!_rec($dir)) $res = 1; |
171 | exit(1); |
- | |
172 | } |
175 | } |
173 | - | ||
174 | _rec($dir); |
- | |
175 | - | ||
176 | if ($show_verbose) echo "Done.\n"; |
176 | if ($show_verbose) echo "Done.\n"; |
- | 177 | exit($res); |
|
- | 178 | ||
177 | 179 |