Merge pull request #69 from CloudCannon/fix/html-unescaping

bglw · web-flow · commit 72b5dc807203 · 2024-12-03T10:07:34.000+13:00
Stop Rosey unescaping escaped HTML
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -9,6 +9,8 @@
 
 ## Unreleased
 
+* Fix issue where Rosey would unescape characters that were entity encoded in the source HTML.
+
 ## v2.3.0 (October 31, 2024)
 
 * Added `data-rosey-ignore` attribute to ignore hrefs when rewriting pages.
diff --git a/rosey/features/build/rosey-build-code-blocks.feature b/rosey/features/build/rosey-build-code-blocks.feature
@@ -0,0 +1,29 @@
+Feature: Rosey Build Code Blocks
+  Background:
+    Given I have the environment variables:
+      | ROSEY_SOURCE | dist/site            |
+      | ROSEY_DEST   | dist/translated_site |
+
+  Scenario: Rosey doesn't ruin innocent escaped code
+    Given I have a "dist/site/index.html" file with the content:
+      """
+      <html>
+      <body>
+      <div data-rosey="seal">Kiss From A Rose</div>
+      <pre><code>&lt;configuration&gt;&lt;setting v='true'/&gt;&lt;/configuration&gt;</code></pre>
+      </body>
+      </html>
+      """
+    And I have a "rosey/locales/manuka.json" file with the content:
+      """
+      {
+        "seal": "Pollen From A Tree"
+      }
+      """
+    When I run my program with the flags:
+      | build |
+    Then I should see a selector '[data-rosey="seal"]' in "dist/translated_site/manuka/index.html" with the attributes:
+      | data-rosey | seal               |
+      | innerText  | Pollen From A Tree |
+    Then I should see a selector 'pre > code' in "dist/translated_site/manuka/index.html" with the attributes:
+      | innerText  | <configuration><setting v='true'/></configuration> |
diff --git a/rosey/features/generate/rosey-generate-code-blocks.feature b/rosey/features/generate/rosey-generate-code-blocks.feature
@@ -0,0 +1,22 @@
+Feature: Rosey Generate Code Blocks
+	Background:
+		Given I have the environment variables:
+			| ROSEY_SOURCE | dist/site            |
+			| ROSEY_DEST   | dist/translated_site |
+
+	Scenario: Rosey generates base.json files with escaped code
+		Given I have a "dist/site/index.html" file with the content:
+			"""
+            <html>
+                <body data-rosey-ns='home'>
+                    <div data-rosey="seal">Kiss From A Rose</div>
+                    <pre><code data-rosey="code">&lt;configuration&gt;&lt;setting v='true'/&gt;&lt;/configuration&gt;</code></pre>
+                </body>
+            </html>
+			"""
+		When I run my program with the flags:
+			| generate |
+		Then I should see "rosey/base.json" containing the values:
+			| version                 | int:2                                                                |
+			| keys.home:seal.original | Kiss From A Rose                                                     |
+			| keys.home:code.original | &lt;configuration&gt;&lt;setting v='true'/&gt;&lt;/configuration&gt; |
diff --git a/rosey/src/lib.rs b/rosey/src/lib.rs
@@ -342,3 +342,14 @@ pub fn inline_templates(dom: &kuchiki::NodeRef) {
         }
     })
 }
+
+pub fn escape_source_text(dom: &kuchiki::NodeRef) {
+    dom.children().for_each(|node| {
+        if let Some(text_node) = node.as_text() {
+            text_node.replace_with(|old| old.replace('<', "&lt;").replace('>', "&gt;"));
+        } else {
+            node.descendants()
+                .for_each(|node| escape_source_text(&node));
+        }
+    });
+}
diff --git a/rosey/src/runners/builder/html.rs b/rosey/src/runners/builder/html.rs
@@ -246,6 +246,7 @@ impl<'a> RoseyPage<'a> {
     ) -> Self {
         let dom = kuchiki::parse_html().one(content);
         crate::inline_templates(&dom);
+        crate::escape_source_text(&dom);
 
         RoseyPage {
             dom,
