From e0c659b34b8992111bcc64d3f0eea75dfa193412 Mon Sep 17 00:00:00 2001
From: Alexander Kampmann <programmer@nurfuerspam.de>
Date: Thu, 1 Mar 2012 11:44:33 +0100
Subject: Enabled automated doc building. Changes: -build.xml, phing build file
 -@package tags at acl_selector.php and db_update.php -added documentation and
 some IDE files to .gitignore -automated tool to find files that corrupt the
 doc build process -removed ansi characters from SSH1.php, docBlox could not
 handle them

---
 util/docblox_errorchecker.php | 145 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 145 insertions(+)
 create mode 100644 util/docblox_errorchecker.php

(limited to 'util/docblox_errorchecker.php')

diff --git a/util/docblox_errorchecker.php b/util/docblox_errorchecker.php
new file mode 100644
index 000000000..af4c76444
--- /dev/null
+++ b/util/docblox_errorchecker.php
@@ -0,0 +1,145 @@
+<?php 
+/**
+ * When I installed docblox, I had the experience that it does not generate any output at all.
+ * This script may be used to find that kind of problems with the documentation build process.
+ * If docblox generates output, use another approach for debugging.
+ *
+ * Basically, docblox takes a list of files to build documentation from. This script assumes there is a file or set of files
+ * breaking the build when it is included in that list. It tries to calculate the smallest list containing these files.
+ * Unfortunatly, the original problem is NP-complete, so what the script does is a best guess only.
+ *
+ * So it starts with a list of all files in the project.
+ * If that list can't be build, it cuts it in two parts and tries both parts independently. If only one of them breaks,
+ * it takes that one and tries the same independently. If both break, it assumes this is the smallest set. This assumption
+ * is not necessarily true. Maybe the smallest set consists of two files and both of them were in different parts when
+ * the list was divided, but by now it is my best guess. To make this assumption better, the list is shuffled after every step. 
+ *
+ * After that, the script tries to remove a file from the list. It tests if the list breaks and if so, it
+ * assumes that the file it removed belongs to the set of errorneous files. 
+ * This is done for all files, so, in the end removing one file leads to a working doc build. 
+ *
+ * @package util
+ * @author Alexander Kampmann
+ */
+
+/**
+ * This function generates a comma seperated list of file names.
+ * 
+ * @package util
+ * 
+ * @param array $fileset Set of file names
+ * 
+ * @return string comma-seperated list of the file names
+ */
+function namesList($fileset) {
+	$fsparam="";
+	foreach($fileset as $file) {
+		$fsparam=$fsparam.",".$file;
+	}
+	return $fsparam;
+};
+
+/**
+ * This functions runs phpdoc on the provided list of files
+ * @package util
+ * 
+ * @param array $fileset Set of filenames
+ * 
+ * @return bool true, if that set can be built
+ */
+function runs($fileset) {
+	$fsParam=namesList($fileset);
+	exec('docblox -t phpdoc_out -f '.$fsParam);
+	if(file_exists("phpdoc_out/index.html")) {
+		echo "\n Subset ".$fsParam." is okay. \n";
+		exec('rm -r phpdoc_out');
+		return true;
+	} else {
+		echo "\n Subset ".$fsParam." failed. \n";
+		return false;
+	}
+};
+
+/**
+ * This functions cuts down a fileset by removing files until it finally works.
+ * it was meant to be recursive, but php's maximum stack size is to small. So it just simulates recursion.
+ *
+ * In that version, it does not necessarily generate the smallest set, because it may not alter the elements order enough.
+ * 
+ * @package util
+ * 
+ * @param array $fileset set of filenames
+ * @param int $ps number of files in subsets
+ * 
+ * @return array a part of $fileset, that crashes
+ */
+function reduce($fileset, $ps) {
+	//split array...
+	$parts=array_chunk($fileset, $ps);
+	//filter working subsets...
+	$parts=array_filter($parts, "runs");
+	//melt remaining parts together
+	if(is_array($parts)) {
+		return array_reduce($parts, "array_merge", array());
+	}
+	return array();
+};
+
+//return from util folder to frindica base dir
+$dir='..';
+
+//stack for dirs to search
+$dirstack=array();
+//list of source files
+$filelist=array();
+
+//loop over all files in $dir
+while($dh=opendir($dir)) {
+	while($file=readdir($dh)) {
+		if(is_dir($dir."/".$file)) {
+			//add to directory stack
+			if($file!=".." && $file!=".") {
+				array_push($dirstack, $dir."/".$file);
+				echo "dir ".$dir."/".$file."\n";
+			}
+		} else  {
+			//test if it is a source file and add to filelist
+			if(substr($file, strlen($file)-4)==".php") {
+				array_push($filelist, $dir."/".$file);
+				echo $dir."/".$file."\n";
+			}
+		}
+	}
+	//look at the next dir
+	$dir=array_pop($dirstack);
+}
+
+//check the entire set
+if(runs($filelist)) {
+	echo "I can not detect a problem. \n";
+	exit;
+}
+
+//check half of the set and discard if that half is okay
+$res=$filelist;
+$i=0;
+do {
+	$i=count($res);
+	echo $i."/".count($fileset)." elements remaining. \n";
+	$res=reduce($res, count($res)/2);
+	shuffle($res);
+} while(count($res)<$i);
+
+//check one file after another
+$needed=array();
+
+while(count($res)!=0) {
+	$file=array_pop($res);
+
+	if(runs(array_merge($res, $needed))) {
+		echo "needs: ".$file." and file count ".count($needed);
+		array_push($needed, $file);
+	}
+}
+
+echo "\nSmallest Set is: ".namesList($needed)." with ".count($needed)." files. ";
-- 
cgit v1.2.3