Refactores and add tests for url resolving

index.php

@@ -1,12 +1,14 @@
 <?php
 
-require_once "lib-htmgem.php";
+require_once "lib-htmgem.inc.php";
+require_once "lib-html.inc.php";
 
 # The url argument is always absolute compared to the document root.
 $url = @$_REQUEST["url"];
 $urlRewriting = @$_REQUEST["rw"]=="1";
 
-/* Installation page
+/**
+ * Installation page
  *
  * Accessing directly /htmgem will make display the self-hosted documentation
  * contained in "index.gmi". If it's removed, display an empty page with a
@@ -17,13 +19,16 @@ if (empty($url)) {
         http_response_code(403);
         die("<!-- index.gmi missing -->");
     }
-    $t = new \htmgem\GemTextTranslate_html(@file_get_contents("index.gmi"), true, "/htmgem");
-    echo $t->getFullHtml();
+    $gt_html = new \htmgem\GemTextTranslate_html(@file_get_contents("index.gmi"), true, "/htmgem");
+    echo \htmgem\html\getFullHtml($gt_html);
     exit();
 }
 
 $documentRoot = $_SERVER['DOCUMENT_ROOT'];
 
+/**
+ * Provides index.gmi if no page given
+ */
 if (!preg_match("/\.gmi$/", $url)) {
     if ($url[-1] == "/")
         $url = $url."index.gmi";
@@ -49,20 +54,14 @@ switch(true) {
 if ($go404) {
     error_log("HtmGem: 404 $url $filePath");
     http_response_code(404);
-    $page404 = <<<EOF
-# ⚠ Page non trouvée
-
-​**$url**
-
-=> .. 🔄 🔄
-EOF;
-    $t = new \htmgem\GemTextTranslate_html($page404);
-    echo $t->getFullHtml();
+    $page404 = \htmgem\html\get404GmiPage("Page not found", $url);
+    $gt_html = new \htmgem\GemTextTranslate_html($page404);
+    echo \htmgem\html\getFullHtml($gt_html);
     exit();
 }
 
 # to false only if textDecoration=0 in the URL
-$textDecoration = "0" != @$_REQUEST['textDecoration'];
+$gt_htmlextDecoration = "0" != @$_REQUEST['textDecoration'];
 
 $fileContents = @file_get_contents($filePath);
 # Removes the Byte Order Mark
@@ -106,11 +105,11 @@ if ($urlRewriting)
     $baseUrl = null;
 else
     $baseUrl = dirname($url);
-$t = new \htmgem\GemTextTranslate_html($fileContents, $textDecoration, $baseUrl);
+$gt_html = new \htmgem\GemTextTranslate_html($fileContents, $gt_htmlextDecoration, $baseUrl);
 if ("none" == $style) {
-    $t->addCss("");
+    $gt_html->addCss("");
 } elseif ("/" == @$style[0]) {
-    $t->addCss($style);
+    $gt_html->addCss($style);
 } elseif (empty($style)) {
     $parts = pathinfo($filePath);
     $localCss = $parts["filename"].".css";
@@ -118,12 +117,12 @@ if ("none" == $style) {
     if (file_exists($localCssFilePath)) {
         # Warning, using htmhem.php?url=… will make $localCss not found
         # as the path is relative to htmgem.php and not / !
-        $t->addCss($localCss);
+        $gt_html->addCss($localCss);
     }
 } else { #TODO: regex check for $style
-    $t->addCss("/htmgem/css/$style.css");
+    $gt_html->addCss("/htmgem/css/$style.css");
 }
 
-echo $t->getFullHtml();
+echo \htmgem\html\getFullHtml($gt_html);
 
 ?>

lib-htmgem.inc.php

@@ -197,6 +197,9 @@ class GemtextTranslate_html {
         $this->cssList []= $css;
     }
 
+    function getCss() { return $this->cssList; }
+    function getTitle() { return $this->pageTitle; }
+
     const NARROW_NO_BREAK_SPACE = " ";
     const DASHES
         ="‒" # U+2012 Figure Dash
@@ -263,14 +266,20 @@ class GemtextTranslate_html {
         $text = preg_replace("/  +/", " ", $text);
     }
 
-    protected static function resolve_path($path) {
+    /**
+     * Resolve $path interpretating / . and ..
+     * @param $path str
+     * @returns "/" if .. goes above the limit
+     */
+    public static function resolve_path($path) {
+        if (empty($path)) return "";
         $absolute = "/"==$path[0];
         $parts = array_filter(explode("/", $path), 'strlen');
         $chuncks = array();
         foreach ($parts as $part) {
             if ('.' == $part) continue;
             if ('..' == $part) {
-                array_pop($chuncks);
+                if (is_null(array_pop($chuncks))) return "/";
             } else {
                 $chuncks[] = $part;
             }
@@ -374,34 +383,6 @@ class GemtextTranslate_html {
         $this->translatedGemtext = $output;
     }
 
-    function getFullHtml() {
-        if (!$this->cssList)
-            $css = array("/htmgem/css/htmgem.css");
-        else
-            $css = $this->cssList;
-        $output = <<<EOL
-<!DOCTYPE html>
-<html lang="">
-<head>
-<title>{$this->pageTitle}</title>
-<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
-EOL;
-        foreach ($css as $c) {
-            $output .= "\n<link type='text/css' rel='StyleSheet' href='$c'>\n";
-        }
-        $output .= <<<EOL
-</head>
-<body>\n
-EOL;
-        $output .= $this->translatedGemtext;
-        $output .= "</body>\n</html>\n";
-
-        echo $output;
-    }
-
-    public function __toString() {
-        return $this->translatedGemtext;
-    }
 } // GemTextTranslate_html
 
 ?>

lib-html.inc.php

@@ -0,0 +1,44 @@
+<?php
+
+namespace htmgem\html;
+
+mb_internal_encoding("UTF-8");
+mb_regex_encoding("UTF-8");
+
+/**
+ * Returns a full HTML page base
+ */
+function getFullHtml(\htmgem\GemtextTranslate_html $gt_html) {
+    $css = $gt_html->getCss();
+    if (!$css) $css = array("/htmgem/css/htmgem.css");
+    $output = <<<EOL
+<!DOCTYPE html>
+<html lang="">
+<head>
+<title>{$gt_html->getTitle()}</title>
+<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
+EOL;
+    foreach ($css as $c) {
+        $output .= "\n<link type='text/css' rel='StyleSheet' href='$c'>\n";
+    }
+    $output .= <<<EOL
+</head>
+<body>\n
+EOL;
+    $output .= $gt_html->translatedGemtext;
+    $output .= "</body>\n</html>\n";
+
+    return $output;
+}
+
+function get404Gmipage($message, $url) {
+    return <<<EOF
+# ⚠ $message
+
+​**$url**
+
+=> .. 🔄 🔄
+EOF;
+}
+
+?>

tests/cli/parse_gemtext.php

@@ -2,7 +2,7 @@
 
 $fileName = $argv[1];
 
-require_once dirname(__FILE__)."/../lib-htmgem.php";
+require_once dirname(__FILE__)."/../lib-htmgem.inc.php";
 
 $text = file_get_contents($fileName);
 $parsedGemtext = \htmgem\gemtextParser($text);

tests/cli/translate_to_gemtext.php

@@ -2,7 +2,7 @@
 
 $fileName = $argv[1];
 
-require_once dirname(__FILE__)."/../../lib-htmgem.php";
+require_once dirname(__FILE__)."/../../lib-htmgem.inc.php";
 
 $text = file_get_contents($fileName);
 $gt_gemtext = new \htmgem\GemtextTranslate_gemtext($text);

tests/cli/translate_to_html.php

@@ -2,7 +2,7 @@
 
 $fileName = $argv[1];
 
-require_once dirname(__FILE__)."/../../lib-htmgem.php";
+require_once dirname(__FILE__)."/../../lib-htmgem.inc.php";
 
 $text = file_get_contents($fileName);
 $gt_gemtext = new \htmgem\GemtextTranslate_html($text);

tests/miscTest.php

@@ -0,0 +1,96 @@
+<?php declare(strict_types=1);
+use PHPUnit\Framework\TestCase;
+
+require_once dirname(__FILE__)."/../lib-htmgem.inc.php";
+
+final class miscTest extends TestCase {
+
+    public function test_resolveLink(): void {
+        $this->assertSame(
+            \htmgem\GemtextTranslate_html::resolve_path(""),
+            "",
+            "empty link"
+        );
+        $this->assertSame(
+            \htmgem\GemtextTranslate_html::resolve_path("test"),
+            "test",
+            "single word"
+        );
+        $this->assertSame(
+            \htmgem\GemtextTranslate_html::resolve_path(" "),
+            " ",
+            "single space"
+        );
+        $this->assertSame(
+            \htmgem\GemtextTranslate_html::resolve_path(" A B "),
+            " A B ",
+            "several space"
+        );
+        $this->assertSame(
+            \htmgem\GemtextTranslate_html::resolve_path("/"),
+            "/",
+            "one slash"
+        );
+        $this->assertSame(
+            \htmgem\GemtextTranslate_html::resolve_path("//"),
+            "/",
+            "two slashes"
+        );
+        $this->assertSame(
+            \htmgem\GemtextTranslate_html::resolve_path("/////"),
+            "/",
+            "five slashes"
+        );
+        $this->assertSame(
+            \htmgem\GemtextTranslate_html::resolve_path("one/"),
+            "one",
+            "strip the last slash"
+        );
+        $this->assertSame(
+            \htmgem\GemtextTranslate_html::resolve_path("/two"),
+            "/two",
+            "slash at the beginning"
+        );
+        $this->assertSame(
+            \htmgem\GemtextTranslate_html::resolve_path("/two/"),
+            "/two",
+            "slash at the beginning and the end"
+        );
+        $this->assertSame(
+            \htmgem\GemtextTranslate_html::resolve_path("one/two/"),
+            "one/two",
+            "only the last slash remains"
+        );
+        $this->assertSame(
+            \htmgem\GemtextTranslate_html::resolve_path("one/two/three//"),
+            "one/two/three",
+            "strip the last slashes"
+        );
+        $this->assertSame(
+            \htmgem\GemtextTranslate_html::resolve_path("one/../"),
+            "",
+            "empty one"
+        );
+        $this->assertSame(
+            \htmgem\GemtextTranslate_html::resolve_path("one/two/../"),
+            "one",
+            "empty one two"
+        );
+        $this->assertSame(
+            \htmgem\GemtextTranslate_html::resolve_path("one/two/../.."),
+            "",
+            "empty one two twice"
+        );
+        $this->assertSame(
+            \htmgem\GemtextTranslate_html::resolve_path("one/../two/./../three"),
+            "three",
+            "waltz"
+        );
+        $this->assertSame(
+            \htmgem\GemtextTranslate_html::resolve_path("one/../.."),
+            "/",
+            "directory traversal"
+        );
+    }
+
+}

tests/parserTest.php

@@ -1,7 +1,7 @@
 <?php declare(strict_types=1);
 use PHPUnit\Framework\TestCase;
 
-require_once dirname(__FILE__)."/../lib-htmgem.php";
+require_once dirname(__FILE__)."/../lib-htmgem.inc.php";
 
 function parse($text): array {
     return iterator_to_array(\htmgem\gemtextParser($text));

tests/translateToGemtextTest.php

@@ -2,7 +2,7 @@
 use PHPUnit\Framework\TestCase;
 
 $dirname_file = dirname(__FILE__);
-require_once "$dirname_file/../lib-htmgem.php";
+require_once "$dirname_file/../lib-htmgem.inc.php";
 require_once "$dirname_file/utils.inc.php";
 
 function translate($text): string {

tests/translateToHtmlTest.php

@@ -2,11 +2,12 @@
 use PHPUnit\Framework\TestCase;
 
 $dirname_file = dirname(__FILE__);
-require_once "$dirname_file/../lib-htmgem.php";
+require_once "$dirname_file/../lib-htmgem.inc.php";
 require_once "$dirname_file/utils.inc.php";
 
 function translateHtml($text): string {
-    return strval(new htmgem\GemtextTranslate_html($text));
+    $gt_html = new htmgem\GemtextTranslate_html($text);
+    return strval($gt_html->translatedGemtext);
 }
 
 final class translateToHtmlTest extends TestCase {