jenkins-bot has submitted this change. ( https://gerrit.wikimedia.org/r/c/pywikibot/core/+/609297 )

Change subject: [bugfix] Avoid HTML entity substitution in <syntaxhighlight> ......................................................................

[bugfix] Avoid HTML entity substitution in <syntaxhighlight>

Bug: T254350 Change-Id: I70db180f6755e8b92675500e35969dc6f205800a --- M pywikibot/cosmetic_changes.py M tests/cosmetic_changes_tests.py 2 files changed, 10 insertions(+), 3 deletions(-)

Approvals: Xqt: Looks good to me, approved jenkins-bot: Verified

diff --git a/pywikibot/cosmetic_changes.py b/pywikibot/cosmetic_changes.py index 1baba11..e8dc359 100755 --- a/pywikibot/cosmetic_changes.py +++ b/pywikibot/cosmetic_changes.py @@ -646,7 +646,10 @@ if self.template: ignore += [32] # Space ( ) ignore += [58] # Colon (:) - text = pywikibot.html2unicode(text, ignore=ignore, exceptions=['code']) + # TODO: T254350 - what other extension tags should be avoided? + # (graph, math, score, timeline, etc.) + text = pywikibot.html2unicode( + text, ignore=ignore, exceptions=['comment', 'source']) return text

def removeEmptySections(self, text): diff --git a/tests/cosmetic_changes_tests.py b/tests/cosmetic_changes_tests.py index bf56944..07bb02e 100644 --- a/tests/cosmetic_changes_tests.py +++ b/tests/cosmetic_changes_tests.py @@ -63,8 +63,12 @@ self.cct.resolveHtmlEntities( '&amp;#&nbsp;#&#32;#&#48;#&#62;#&#120;')) self.assertEqual( - '<code>&#32;</code>', - self.cct.resolveHtmlEntities('<code>&#32;</code>')) + '<syntaxhighlight>&#32;</syntaxhighlight>', + self.cct.resolveHtmlEntities( + '<syntaxhighlight>&#32;</syntaxhighlight>')) + self.assertEqual( + '<!-- &ndash; -->', + self.cct.resolveHtmlEntities('<!-- &ndash; -->'))

def test_removeUselessSpaces(self): """Test removeUselessSpaces method."""