Documentation Code

sourceMakeup

sourceMakeup is a source code viewer written in PHP on the backend, and jQuery and jKit on the frontend. sourceMakeup was heavely inspired by the first versions of Docco, with some additional features that enhance the frontend visually and feature wise. It was mainly developed as a source code viewer for the jQuery based UI toolkit jKit. It's main use is to create beautiful source documentation for JavaScript libraries and plugins, but it can be used to document small PHP code and CSS markup as well, however, with some limitations.

Comments for sourceMakeup should be written in Markdown. This way sourceMakeup can correctly and beautifully format the code and add a navigation menu linked to the different parts of your code.

  • (c) 2013 by Fredi Bach
  • Home

License

sourceMakeup is open source and MIT licensed. For more informations read the license.txt file

Settings

  • $dir contains the path to the documented files
  • $file is the main file
  • $files contains all the files, leave it empty if you don't want to restrict the list to certain files
  • $extensions contains a comma separated list of allowed file extensions
  • $filter filters out any files that contain a certain string
  • $dev mode ignores the cache file and always generates a new one
39
40
41
42
43
44
45
46
47
48
49
50
    $dir = './';
    $file = 'sourcemakeup.php';
    $files = array(
                'sourcemakeup.php',
                'js/jquery.jkit.1.1.28.js',
                'js/jquery.jkit.1.1.29.js',
                'js/sourcemakeup.js',
                'css/sourcemakeup.css'
    );
    $extensions = 'js,php,css';
    $filter = '.min.js';
    $dev = false;

The Source

Is the main file the one from the settings or another one?

56
57
58
    if (isset($_GET['file']) && $_GET['file'] != ''){
        $file = $_GET['file'];
    }

Define the path for our cached file:

62
    $cachefile = 'cache/'.md5($file).'.txt';

Create a list of all files we want to document. In case $files is not empty, those files will be used and the following code block will be ignored.

67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
    if (count($files) == 0){
        if ($dir != ''){
            $tempfiles = scandir($dir);
            $files = array();
            foreach($tempfiles as $f){
                if ($f != '.' && $f != '..' && stripos($f, $filter) === false){
                    $allowed = explode(',', $extensions);
                    $ext = pathinfo($f, PATHINFO_EXTENSION);
                    if (in_array($ext, $allowed)) {
                        $files[] = $f;
                    }
                }
            }
        } else {
            $files = array($file);
        }
    }

Do we have to load the cahed file or create a new one?

87
    if (file_exists($cachefile) && !$dev){

Load the cached file:

91
92
93
        $output = file_get_contents($cachefile);
    
    } else {

Ok, so we need to create new output. To do that we first need to include the Markdown class:

97
        include_once "libs/markdown/markdown.php";

Load the file and create an array of all lines:

101
102
        $data = file_get_contents($file);
        $lines = explode("\n", $data);

Create some variables we need later:

106
107
108
109
        $blocks = array();
        $iscomment = false;
        $currentblock = 0;
        $linenumber = 0;

Step through all lines and combine them into codeblocks. Each codeblock can have following contents:

  • comment contains all comments of this codeblock
  • code contains all code of this block
  • lnr contains the original linenumbers for the above code
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
        foreach($lines as $line){
        
            $linenumber++;
        
            if (substr(trim($line),0,2) == '//'){
                if (!$iscomment){
                    $currentblock++;
                }
                $blocks[$currentblock]['comment'] .= "\n".substr(trim($line),3);
                $blocks[$currentblock]['code'] .= "";
                $iscomment = true;
                if (substr(trim(substr(trim($line),3)),0,3) == '{!}'){
                    $currentblock++;
                }
            } else {
                $blocks[$currentblock]['code'] .= "\n".str_replace("\t", "    ", $line);
                $blocks[$currentblock]['lnr'] .= "\n".$linenumber;
                $blocks[$currentblock]['comment'] .= "";
                $iscomment = false;
            }
    
        }

If this is a PHP script with <?php on the first line, we're going to remove it as it's only ruining our design:

142
143
144
        if (trim($blocks[0]['code']) == '<?php'){
            $blocks[0]['code'] = '';
        }

We have now all we need to construct the table that contains all of our code. The table has two columns, the first contains all the comments and the second one contains all the code.

149
        $output = '';

Loop through all blocks and construct the rows of our table:

153
        foreach($blocks as $block){

Each block is a row with a comment and a code column. Sometimes only one of them actually has content and if both have no content, simply skip this block.

158
159
160
161
162
            if (trim($block['comment']) != '' || trim($block['code']) != ''){
                
                $output .= '<tr class="docu">';
                
                if (substr(trim($block['comment']),0,3) == '{!}'){

This block is an instruction block. Instruction blocks can be created like this:

{!} instructiontype: instruction

Instructions are normally hidden, but can be made visible. They are ment to contain special instructions either for a script or for code contributors. If you use them, make sure to supply the icons for each of your instruction type. They have to ba named like this:

sourcemakeup.instructiontype.png
175
176
177
178
179
180
181
182
183
184
                    $split = explode(':', substr(trim($block['comment']),3));
                    
                    $block['comment'] = '<span class="itype">';
                    $block['comment'] .= '<img src="imgs/sourcemakeup.'.trim($split[0]).'.png"> '.$split[0].'</span>';
                    $block['comment'] .= ':<span class="iopt">'.$split[1].'</span>';
                    
                    $output .= '<td class="comment instruction">&nbsp;</td>';
                    $output .= '<td class="code instruction">'.$block['comment'].'</td>';
                    
                } else {

This is a regular codeblock. As we don't want to display any uneeded linebreaks, we use a special funtion to strip them away. The function returns an array that not only contains the stripped code, it tells us exactly how many lines where removed from the beginning and from the end. This way we can strip exactly the same number of lines from the linenumber string.

192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
                    $codedata = trimLinebreaks($block['code']);
                
                    $templines = explode("\n", $block['lnr']);
                
                    if ($codedata['shifted'] > 0){
                        for ($i=1; $i<=$codedata['shifted']; $i++){
                            array_shift($templines);
                        }
                    }
                
                    if ($codedata['popped'] > 0){
                        for ($i=1; $i<=$codedata['popped']; $i++){
                            array_pop($templines);
                        }
                    }
                
                    $lnr = implode("\n", $templines);

For the comment column, we basically do two things. We parse the Markdown formatting into nice HTML with the Markdown library, and we than add ids to all HTML headers so that we can create an anchor navigation.

213
                    $output .= '<td class="comment">'.addHeaderAnchors(Markdown(trim($block['comment']))).'</td>';

All we need to do for the code is escape the HTML special chars. The formatting is done by the Google Code Prettifier JavaScript.

217
218
219
220
221
222
223
224
                    $output .= '<td class="code"><pre class="linenumbers">'.$lnr.'</pre>';
                    $output .= '<pre class="prettyprint">'.htmlspecialchars($codedata['code']).'</pre></td>';
                }
                
                $output .= '</tr>';
                
            }
        }

Now as we created our output, we save it in a cache for later use:

228
229
230
231
232
        $fh = fopen($cachefile, 'w') or die("can't open file");
        fwrite($fh, $output);
        fclose($fh);
    
    }

And lastly, we inlude the template to generate our output:

236
    include 'sourcemakeup.template.php';

Functions

The addHeaderAnchors function and the regex callback function addHeaderAnchorsCallback add ids to each header in the supplied HTML based on the title of the header.

244
245
246
247
248
249
250
    function addHeaderAnchors($text){
        return preg_replace_callback('/\<h([0-6]{1})\>(.+)\<\/h[0-6]{1}\>/', 'addHeaderAnchorsCallback', $text);
    }
    
    function addHeaderAnchorsCallback($matches){
        return '<h'.$matches[1].' id="'.createURLSlug($matches[2]).'">'.$matches[2].'</h'.$matches[1].'>';
    }

The createURLSlug function makes a nice url slug where all special characters are removed and spaces replaced with the "-" character.

255
256
257
258
259
260
261
262
263
264
265
266
    function createURLSlug($str, $replace=array(), $delimiter='-'){
        if( !empty($replace) ) {
            $str = str_replace((array)$replace, ' ', $str);
        }

        $clean = iconv('UTF-8', 'ASCII//TRANSLIT', $str);
        $clean = preg_replace("/[^a-zA-Z0-9\/_|+ -]/", '', $clean);
        $clean = strtolower(trim($clean, '-'));
        $clean = preg_replace("/[\/_|+ -]+/", $delimiter, $clean);

        return $clean;
    }

The trimLinebreaks function removes all unneaded linebreaks before an after a string and tells us in the return array exactly how many of them.

271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
    function trimLinebreaks($string){
    
        $lines = explode("\n", $string);
    
        $shifted = 0;
        $popped = 0;
    
        while(trim($lines[0]) == '' && count($lines) > 0){
            array_shift($lines);
            $shifted++;
        }
    
        while( trim( $lines[count($lines)-1] ) == '' && count($lines) > 0){
            array_pop($lines);
            $popped++;
        }
    
        $arr = array('code' => implode("\n", $lines), 'shifted' => $shifted, 'popped' => $popped);
    
        return $arr;
    
    }