Add some original tests

lib/Microformats/Parser.php

@@ -355,10 +355,9 @@ class Parser {
      * @param \DOMElement $node The element to start searching from, including itself
      */
     protected function getRootCandidates(\DOMElement $node): void {
-        $query = [];
-        $query[] = './/*[contains(concat(" ", normalize-space(@class)), " h-")]';
+        $query = [".", ".//*[contains(concat(' ', normalize-space(@class)), ' h-')]"];
         foreach (array_keys(static::BACKCOMPAT_ROOTS) as $root) {
-            $query[] = './/*[contains(concat(" ", normalize-space(@class), " "), " '.$root.' ")]';
+            $query[] = ".//*[contains(concat(' ', normalize-space(@class), ' '), ' $root ')]";
         }
         $query = implode("|", $query);
         $this->roots = iterator_to_array($this->xpath->query($query, $node));
@@ -711,7 +710,7 @@ class Parser {
                     # else if .h-x>abbr:only-child[title]:not([title=""]):not[.h-*] then use that abbr title for name
                     $name = $set->item(0)->getAttribute("title");
                 } elseif (
-                    ($set = $this->xpath->query("./*[not(template) and count(../*) = 1]", $root))->length
+                    ($set = $this->xpath->query("./*[local-name() != 'template' and count(../*) = 1]", $root))->length
                     && !$this->hasRoots($set->item(0))
                     && ($set = $this->xpath->query("./img[@alt and @alt != '' and count(../*) = 1]", $set->item(0)))->length
                     && !$this->hasRoots($set->item(0))
@@ -719,7 +718,7 @@ class Parser {
                     # else if .h-x>:only-child:not[.h-*]>img:only-child[alt]:not([alt=""]):not[.h-*] then use that img’s alt for name
                     $name = $set->item(0)->getAttribute("alt");
                 } elseif (
-                    ($set = $this->xpath->query("./*[not(template) and count(../*) = 1]", $root))->length
+                    ($set = $this->xpath->query("./*[local-name() != 'template' and count(../*) = 1]", $root))->length
                     && !$this->hasRoots($set->item(0))
                     && ($set = $this->xpath->query("./area[@alt and @alt != '' and count(../*) = 1]", $set->item(0)))->length
                     && !$this->hasRoots($set->item(0))
@@ -727,7 +726,7 @@ class Parser {
                     # else if .h-x>:only-child:not[.h-*]>area:only-child[alt]:not([alt=""]):not[.h-*] then use that area’s alt for name
                     $name = $set->item(0)->getAttribute("alt");
                 } elseif (
-                    ($set = $this->xpath->query("./*[not(template) and count(../*) = 1]", $root))->length
+                    ($set = $this->xpath->query("./*[local-name() != 'template' and count(../*) = 1]", $root))->length
                     && !$this->hasRoots($set->item(0))
                     && ($set = $this->xpath->query("./abbr[@title and @title != '' and count(../*) = 1]", $set->item(0)))->length
                     && !$this->hasRoots($set->item(0))
@@ -758,7 +757,7 @@ class Parser {
                     # else if .h-x>object[data]:only-of-type:not[.h-*] then use that object’s data for photo
                     $photo = $set->item(0)->getAttribute("data");
                 } elseif (
-                    ($set = $this->xpath->query("./*[not(template) and count(../*) = 1]", $root))->length
+                    ($set = $this->xpath->query("./*[local-name() != 'template' and count(../*) = 1]", $root))->length
                     && !$this->hasRoots($set->item(0))
                     && ($set = $this->xpath->query("./img[@src and count(../img) = 1]", $set->item(0)))->length
                     && !$this->hasRoots($set->item(0))
@@ -766,7 +765,7 @@ class Parser {
                     # else if .h-x>:only-child:not[.h-*]>img[src]:only-of-type:not[.h-*], then use the result of "parse an img element for src and alt" (see Sec.1.5) for photo
                     $out['properties']['photo'] = [$this->parseImg($set->item(0))];
                 } elseif (
-                    ($set = $this->xpath->query("./*[not(template) and count(../*) = 1]", $root))->length
+                    ($set = $this->xpath->query("./*[local-name() != 'template' and count(../*) = 1]", $root))->length
                     && !$this->hasRoots($set->item(0))
                     && ($set = $this->xpath->query("./object[@data and count(../object) = 1]", $set->item(0)))->length
                     && !$this->hasRoots($set->item(0))
@@ -797,7 +796,7 @@ class Parser {
                     # else if .h-x>area[href]:only-of-type:not[.h-*], then use that [href] for url
                     $url = $set->item(0)->getAttribute("href");
                 } elseif (
-                    ($set = $this->xpath->query("./*[not(template) and count(../*) = 1]", $root))->length
+                    ($set = $this->xpath->query("./*[local-name() != 'template' and count(../*) = 1]", $root))->length
                     && !$this->hasRoots($set->item(0))
                     && ($set = $this->xpath->query("./a[@href and count(../a) = 1]", $set->item(0)))->length
                     && !$this->hasRoots($set->item(0))
@@ -805,7 +804,7 @@ class Parser {
                     # else if .h-x>:only-child:not[.h-*]>a[href]:only-of-type:not[.h-*], then use that [href] for url
                     $url = $set->item(0)->getAttribute("href");
                 } elseif (
-                    ($set = $this->xpath->query("./*[not(template) and count(../*) = 1]", $root))->length
+                    ($set = $this->xpath->query("./*[local-name() != 'template' and count(../*) = 1]", $root))->length
                     && !$this->hasRoots($set->item(0))
                     && ($set = $this->xpath->query("./area[@href and count(../area) = 1]", $set->item(0)))->length
                     && !$this->hasRoots($set->item(0))

tests/cases/StandardTest.php

@@ -29,32 +29,36 @@ class StandardTest extends \PHPUnit\Framework\TestCase {
         if (isset(self::SUPPRESSED[$name])) {
             $this->markTestIncomplete(self::SUPPRESSED[$name]);
         }
+        // parse input
+        $act = Microformats::fromFile($path.".html", "text/html; charset=UTF-8", "http://example.com/", $options);
         // read expectation data
         $exp = json_decode(file_get_contents($path.".json"), true);
-        // fix up expectation where necessary
-        array_walk_recursive($exp, function(&$v) {
-            // URLs differ trivially from output of our normalization library
-            $v = preg_replace('#^https?://[^/]+$#', "$0/", $v);
-        });
-        // URLs also need fixing as keys in rel-urls
-        foreach ($exp['rel-urls'] as $k => $v) {
-            $fixed = preg_replace('#^https?://[^/]+$#', "$0/", $k);
-            $exp['rel-urls'][$fixed] = $v;
-            if ($fixed !== $k) {
-                unset($exp['rel-urls'][$k]);
+        if ($exp) {
+            // fix up expectation where necessary
+            array_walk_recursive($exp, function(&$v) {
+                // URLs differ trivially from output of our normalization library
+                $v = preg_replace('#^https?://[^/]+$#', "$0/", $v);
+            });
+            // URLs also need fixing as keys in rel-urls
+            foreach ($exp['rel-urls'] as $k => $v) {
+                $fixed = preg_replace('#^https?://[^/]+$#', "$0/", $k);
+                $exp['rel-urls'][$fixed] = $v;
+                if ($fixed !== $k) {
+                    unset($exp['rel-urls'][$k]);
+                }
             }
+            // perform some further monkey-patching on specific tests
+            $exp = $this->fixTests($exp, $name);
+        } else {
+            // if there are no expectations we're probably developing a new test; print the output as JSON
+            echo Microformats::toJson($act, \JSON_PRETTY_PRINT | \JSON_UNESCAPED_SLASHES | \JSON_UNESCAPED_UNICODE);
+            exit;
         }
-        // perform some further monkey-patching on specific tests
-        $exp = $this->fixTests($exp, $name);
-        // parse input
-        $act = Microformats::fromFile($path.".html", "text/html; charset=UTF-8", "http://example.com/", $options);
         // sort both arrays
         $this->ksort($exp);
         $this->ksort($act);
         // run comparison
         if (!$exp) {
-            echo json_encode($act, \JSON_PRETTY_PRINT | \JSON_UNESCAPED_SLASHES | \JSON_UNESCAPED_UNICODE);
-            exit;
         }
         $this->assertSame($exp, $act);
     }
@@ -64,6 +68,9 @@ class StandardTest extends \PHPUnit\Framework\TestCase {
         yield from $this->provideTestList([\MensBeam\Microformats\BASE."vendor-bin/phpunit/vendor/mf2/tests/tests/"], ['simpleTrim' => true]);
         // tests from php-mf2
         yield from $this->provideTestList([\MensBeam\Microformats\BASE."tests/cases/third-party/"], []);
+        // tests from our own corpus
+        yield from $this->provideTestList([\MensBeam\Microformats\BASE."tests/cases/mensbeam/default-settings/"], []);
+        yield from $this->provideTestList([\MensBeam\Microformats\BASE."tests/cases/mensbeam/lang-true/"], ['lang' => true]);
     }
 
     protected function provideTestList(array $tests, ?array $options = null): \Generator {

tests/cases/mensbeam/default-settings/entry-date.html

@@ -0,0 +1,9 @@
+<!--
+    This tests the negative cases of "entry-date", when it is on the wrong
+    element date or the element does not have the correct attribute.
+-->
+<div class="hentry">
+    <span class="entry-date">2023-06-21</span>
+    <time class="entry-date">2023-06-22</time>
+    <time class="entry-date" datetime="2023-06-23">This one is correct</time>
+</div>

tests/cases/mensbeam/default-settings/entry-date.json

@@ -0,0 +1,16 @@
+{
+    "items": [
+        {
+            "type": [
+                "h-entry"
+            ],
+            "properties": {
+                "published": [
+                    "2023-06-23"
+                ]
+            }
+        }
+    ],
+    "rels": {},
+    "rel-urls": {}
+}

tests/cases/mensbeam/default-settings/implied-prop.html

@@ -0,0 +1,31 @@
+<!--
+    This only tests a few cases not covered by standard tests. In particular we
+    need to ensure that templates are not used as valid elements.
+-->
+<div class="h-test">
+    <template>
+        <img src="ook" alt="eek">
+    </template>
+    Template be gone!
+</div>
+
+<div class="h-test">
+    <template>
+        <area src="ook" alt="eek">
+    </template>
+    Template be gone!
+</div>
+
+<div class="h-test">
+    <template>
+        <img src="ook" alt="eek">
+    </template>
+    Template be gone!
+</div>
+
+<div class="h-test">
+    <template>
+        <abbr title="eek">ook</abbr>
+    </template>
+    Template be gone!
+</div>

tests/cases/mensbeam/default-settings/name-conflict.html

@@ -0,0 +1,6 @@
+<div class="h-test">
+    <span class="p-no-conflict">2023-06-06T05:32:12-04:00</span>
+    <span class="p-dt-wins dt-dt-wins">2023-06-06T05:32:12-04:00</span>
+    <span class="p-u-wins dt-u-wins u-u-wins">2023-06-06T05:32:12-04:00</span>
+    <span class="p-e-wins dt-e-wins u-e-wins e-e-wins">2023-06-06T05:32:12-04:00</span>
+</div>

tests/cases/mensbeam/default-settings/name-conflict.json

@@ -0,0 +1,28 @@
+{
+    "items": [
+        {
+            "type": [
+                "h-test"
+            ],
+            "properties": {
+                "no-conflict": [
+                    "2023-06-06T05:32:12-04:00"
+                ],
+                "dt-wins": [
+                    "2023-06-06 05:32:12-0400"
+                ],
+                "u-wins": [
+                    "http://example.com/2023-06-06T05:32:12-04:00"
+                ],
+                "e-wins": [
+                    {
+                        "html": "2023-06-06T05:32:12-04:00",
+                        "value": "2023-06-06T05:32:12-04:00"
+                    }
+                ]
+            }
+        }
+    ],
+    "rels": {},
+    "rel-urls": {}
+}

tests/cases/mensbeam/lang-true/lang.html

@@ -0,0 +1,15 @@
+<html class="h-test">
+
+<div lang="ja">
+    <div class="h-test">
+        <div class="p-name">日本語</div>
+        <div class="e-ook" lang="en">English</div>
+    </div>
+
+    <div class="h-test" lang="iu">
+        <div lang="fr">
+            <div class="e-ook">français</div>
+        </div>
+        <div class="p-name">ᐃᓄᒃᑎᑐᑦ</div>
+    </div>
+</div>

tests/cases/mensbeam/lang-true/lang.json

@@ -0,0 +1,50 @@
+{
+    "items": [
+        {
+            "type": [
+                "h-test"
+            ],
+            "properties": {},
+            "children": [
+                {
+                    "type": [
+                        "h-test"
+                    ],
+                    "properties": {
+                        "name": [
+                            "日本語"
+                        ],
+                        "ook": [
+                            {
+                                "html": "English",
+                                "value": "English",
+                                "lang": "en"
+                            }
+                        ]
+                    },
+                    "lang": "ja"
+                },
+                {
+                    "type": [
+                        "h-test"
+                    ],
+                    "properties": {
+                        "ook": [
+                            {
+                                "html": "français",
+                                "value": "français",
+                                "lang": "fr"
+                            }
+                        ],
+                        "name": [
+                            "ᐃᓄᒃᑎᑐᑦ"
+                        ]
+                    },
+                    "lang": "iu"
+                }
+            ]
+        }
+    ],
+    "rels": {},
+    "rel-urls": {}
+}