Ignore white-space should strip following white-space from Difference / Preview / API endpoints

changedetectionio/api/Watch.py

@@ -260,7 +260,10 @@ class WatchSingleHistory(Resource):
                 response = make_response("No content found", 404)
                 response.mimetype = "text/plain"
         else:
+            from changedetectionio import html_tools
             content = watch.get_history_snapshot(timestamp=timestamp)
+            if self.datastore.data['settings']['application'].get('ignore_whitespace', False):
+                content = html_tools.rstrip_snapshot_content(content)
             response = make_response(content, 200)
             response.mimetype = "text/plain"
 
@@ -328,8 +331,12 @@ class WatchHistoryDiff(Resource):
         no_markup = strtobool(request.args.get('no_markup', 'false'))
 
         # Retrieve snapshot contents
+        from changedetectionio import html_tools
         from_version_file_contents = watch.get_history_snapshot(from_timestamp)
         to_version_file_contents = watch.get_history_snapshot(to_timestamp)
+        if self.datastore.data['settings']['application'].get('ignore_whitespace', False):
+            from_version_file_contents = html_tools.rstrip_snapshot_content(from_version_file_contents)
+            to_version_file_contents = html_tools.rstrip_snapshot_content(to_version_file_contents)
 
         # Get diff preferences from query parameters (matching UI preferences in DIFF_PREFERENCES_CONFIG)
         # Support both 'type' (UI parameter) and 'word_diff' (API parameter) for backward compatibility

changedetectionio/blueprint/ui/preview.py

@@ -84,6 +84,9 @@ def construct_blueprint(datastore: ChangeDetectionStore):
                 versions = list(watch.history.keys())
                 content = watch.get_history_snapshot(timestamp=timestamp)
 
+                if datastore.data['settings']['application'].get('ignore_whitespace', False):
+                    content = html_tools.rstrip_snapshot_content(content)
+
                 triggered_line_numbers = html_tools.strip_ignore_text(content=content,
                                                                       wordlist=watch.get('trigger_text'),
                                                                       mode='line numbers'

changedetectionio/html_tools.py

@@ -631,6 +631,17 @@ def workarounds_for_obfuscations(content):
     return content
 
 
+def rstrip_snapshot_content(content: str) -> str:
+    """Strip trailing whitespace from each line of a snapshot.
+
+    Table-layout pages (common on older sites) cause inscriptis to pad every cell
+    to the maximum column width found in any row, producing lines of thousands of
+    trailing spaces from simple spacer/nav rows that sit alongside wide content rows.
+    This should be applied at display/output time, not at storage time.
+    """
+    return '\n'.join(line.rstrip() for line in content.splitlines())
+
+
 def get_triggered_text(content, trigger_text):
     triggered_text = []
     result = strip_ignore_text(content=content,

changedetectionio/processors/text_json_diff/difference.py

@@ -152,6 +152,11 @@ def render(watch, datastore, request, url_for, render_template, flash, redirect,
         logger.error(f"Unable to read watch history from-version for version {from_version}: {str(e)}")
         from_version_file_contents = f"Unable to read to-version {from_version}.\n"
 
+    if datastore.data['settings']['application'].get('ignore_whitespace', False):
+        from changedetectionio import html_tools
+        to_version_file_contents = html_tools.rstrip_snapshot_content(to_version_file_contents)
+        from_version_file_contents = html_tools.rstrip_snapshot_content(from_version_file_contents)
+
     screenshot_url = watch.get_screenshot()
 
     system_uses_webdriver = datastore.data['settings']['application']['fetch_backend'] == 'html_webdriver'

changedetectionio/tests/unit/test_html_to_text.py

@@ -8,7 +8,7 @@ import threading
 import unittest
 from queue import Queue
 
-from changedetectionio.html_tools import html_to_text
+from changedetectionio.html_tools import html_to_text, rstrip_snapshot_content
 
 
 class TestHtmlToText(unittest.TestCase):
@@ -622,6 +622,54 @@ var also = 1;
         assert 'Real content after the data attribute' in text
 
 
+class TestRstripSnapshotContent(unittest.TestCase):
+    """Tests for rstrip_snapshot_content — the display-layer whitespace trimmer."""
+
+    def test_strips_trailing_spaces(self):
+        """Lines with trailing spaces are rstripped."""
+        content = "hello   \nworld\nfoo   "
+        result = rstrip_snapshot_content(content)
+        assert result == "hello\nworld\nfoo"
+
+    def test_preserves_content_without_trailing_whitespace(self):
+        """Lines without trailing whitespace are left unchanged."""
+        content = "line one\nline two\nline three"
+        assert rstrip_snapshot_content(content) == content
+
+    def test_table_padding_artifact(self):
+        """Simulates inscriptis table-layout padding: a symbol followed by thousands of spaces."""
+        padded_line = "▲" + " " * 27000
+        content = padded_line + "\n" + " " * 27000 + "\nsome real content"
+        result = rstrip_snapshot_content(content)
+        lines = result.splitlines()
+        assert lines[0] == "▲"
+        assert lines[1] == ""
+        assert lines[2] == "some real content"
+
+    def test_preserves_leading_whitespace(self):
+        """Leading whitespace (indentation) is never touched."""
+        content = "    indented line   \n  another   "
+        result = rstrip_snapshot_content(content)
+        assert result == "    indented line\n  another"
+
+    def test_empty_string(self):
+        assert rstrip_snapshot_content("") == ""
+
+    def test_html_to_text_does_not_rstrip(self):
+        """html_to_text itself must NOT apply rstrip — that's the output layer's job.
+        A table with a narrow cell next to a wide cell produces trailing padding."""
+        html = """<html><body>
+        <table>
+          <tr><td>narrow</td><td>{}</td></tr>
+          <tr><td>x</td><td>y</td></tr>
+        </table>
+        </body></html>""".format("w" * 500)
+        text = html_to_text(html)
+        # At least one line should have trailing spaces (inscriptis table padding)
+        has_trailing = any(line != line.rstrip() for line in text.splitlines())
+        assert has_trailing, "Expected inscriptis to produce trailing-space padding on table rows"
+
+
 if __name__ == '__main__':
     # Can run this file directly for quick testing
     unittest.main()