Fix for em/strong corner cases (#1853)

Fixes #1775

docs/src/markdown/about/changelog.md

@@ -1,5 +1,9 @@
 # Changelog
 
+## 9.7.1
+
+- **FIX**: BetterEm: Fix case where `**` nested between `*` would be handled in an unexpected way.
+
 ## 9.7
 
 - **NEW**: Tabbed: Add new syntax to allow forcing a specific tab to be select by default.

docs/src/markdown/extensions/betterem.md

@@ -60,6 +60,19 @@ BetterEm allows for a more natural nested token feel.
         ***I'm bold and italic!** I am just italic.*
         ```
 
+BetterEm will try to prioritize the more sane option when nesting bold (`**`) between italic (`*`).
+
+!!! example "Prioritize Best Example"
+
+    === "Output"
+        *I'm italic. **I'm bold and italic.** I'm also just italic.*
+
+    === "Markdown"
+
+        ```
+        *I'm italic. **I'm bold and italic.** I'm also just italic.*
+        ```
+
 BetterEm will ensure smart mode doesn't terminate in scenarios where there are a large amount of consecutive tokens
 inside.
 
@@ -68,12 +81,8 @@ inside.
     === "Output"
         ___A lot of underscores____________is okay___
 
-        ___A lot of underscores____________is okay___
-
     === "Markdown"
         ```
-        ___A lot of underscores____________is okay___
-
         ___A lot of underscores____________is okay___
         ```
 
@@ -113,7 +122,7 @@ BetterEm will allow non-smart emphasis to contain "floating" like tokens.
 
          *All will not* be italic*
 
-         *All will not ** be italic*
+         *All will ** be italic*
 
          **All will * be bold**
 

pymdownx/__meta__.py

@@ -185,5 +185,5 @@ def parse_version(ver, pre=False):
     return Version(major, minor, micro, release, pre, post, dev)
 
 
-__version_info__ = Version(9, 7, 0, "final")
+__version_info__ = Version(9, 7, 1, "final")
 __version__ = __version_info__._get_canonical()

pymdownx/betterem.py

@@ -53,6 +53,12 @@
 STAR_STRONG = r'(\*{2})(?!\s)%s(?<!\s)\1' % STAR_CONTENT2
 # __strong__
 UNDER_STRONG = r'(_{2})(?!\s)%s(?<!\s)\1' % UNDER_CONTENT2
+
+# Prioritize *value* when **value** is nested within
+STAR_EM2 = r'(?<!\*)(\*)(?![\*\s])(.+?)(?<![\*\s])(\*)(?!\*)'
+# Prioritize _value_ when __value__ is nested within
+UNDER_EM2 = r'(?<!_)(_)(?![_\s])(.+?)(?<![_\s])(_)(?!_)'
+
 # *emphasis*
 STAR_EM = r'(\*)(?!\s)%s(?<!\s)\1' % STAR_CONTENT
 # _emphasis_
@@ -71,6 +77,8 @@
 SMART_UNDER_STRONG = r'(?<!\w)(_{2})(?![\s_])%s(?<!\s)\1(?!\w)' % SMART_UNDER_CONTENT
 # SMART _em_
 SMART_UNDER_EM = r'(?<!\w)(_)(?![\s_])%s(?<!\s)\1(?!\w)' % SMART_UNDER_CONTENT
+# Prioritize _value_ when __value__ is nested within
+SMART_UNDER_EM2 = r'(?<![\w_])(_)(?![_\s])(.+?)(?<![_\s])(_)(?![_\w])'
 
 # Smart rules for when "smart asterisk" is enabled
 # SMART: ***strong,em***
@@ -89,6 +97,8 @@
 SMART_STAR_STRONG = r'(?:(?<=_)|(?<![\w\*]))(\*{2})(?![\s\*])%s(?<!\s)\1(?:(?=_)|(?![\w\*]))' % SMART_STAR_CONTENT
 # SMART *em*
 SMART_STAR_EM = r'(?:(?<=_)|(?<![\w\*]))(\*)(?![\s\*])%s(?<!\s)\1(?:(?=_)|(?![\w\*]))' % SMART_STAR_CONTENT
+# Prioritize *value* when **value** is nested within
+SMART_STAR_EM2 = r'(?<![\w\*])(\*)(?![\*\s])(.+?)(?<![\*\s])(\*)(?![\*\w])'
 
 
 class AsteriskProcessor(util.PatternSequenceProcessor):
@@ -100,6 +110,7 @@ class AsteriskProcessor(util.PatternSequenceProcessor):
         util.PatSeqItem(re.compile(STAR_STRONG_EM2, re.DOTALL | re.UNICODE), 'double', 'strong,em'),
         util.PatSeqItem(re.compile(STAR_STRONG_EM3, re.DOTALL | re.UNICODE), 'double2', 'strong,em'),
         util.PatSeqItem(re.compile(STAR_STRONG, re.DOTALL | re.UNICODE), 'single', 'strong'),
+        util.PatSeqItem(re.compile(STAR_EM2, re.DOTALL | re.UNICODE), 'single', 'em', True),
         util.PatSeqItem(re.compile(STAR_EM, re.DOTALL | re.UNICODE), 'single', 'em')
     ]
 
@@ -112,6 +123,7 @@ class SmartAsteriskProcessor(util.PatternSequenceProcessor):
         util.PatSeqItem(re.compile(SMART_STAR_EM_STRONG, re.DOTALL | re.UNICODE), 'double', 'em,strong'),
         util.PatSeqItem(re.compile(SMART_STAR_STRONG_EM2, re.DOTALL | re.UNICODE), 'double', 'strong,em'),
         util.PatSeqItem(re.compile(SMART_STAR_STRONG, re.DOTALL | re.UNICODE), 'single', 'strong'),
+        util.PatSeqItem(re.compile(SMART_STAR_EM2, re.DOTALL | re.UNICODE), 'single', 'em', True),
         util.PatSeqItem(re.compile(SMART_STAR_EM, re.DOTALL | re.UNICODE), 'single', 'em')
     ]
 
@@ -125,6 +137,7 @@ class UnderscoreProcessor(util.PatternSequenceProcessor):
         util.PatSeqItem(re.compile(UNDER_STRONG_EM2, re.DOTALL | re.UNICODE), 'double', 'strong,em'),
         util.PatSeqItem(re.compile(UNDER_STRONG_EM3, re.DOTALL | re.UNICODE), 'double2', 'strong,em'),
         util.PatSeqItem(re.compile(UNDER_STRONG, re.DOTALL | re.UNICODE), 'single', 'strong'),
+        util.PatSeqItem(re.compile(UNDER_EM2, re.DOTALL | re.UNICODE), 'single', 'em', True),
         util.PatSeqItem(re.compile(UNDER_EM, re.DOTALL | re.UNICODE), 'single', 'em')
     ]
 
@@ -137,6 +150,7 @@ class SmartUnderscoreProcessor(util.PatternSequenceProcessor):
         util.PatSeqItem(re.compile(SMART_UNDER_EM_STRONG, re.DOTALL | re.UNICODE), 'double', 'em,strong'),
         util.PatSeqItem(re.compile(SMART_UNDER_STRONG_EM2, re.DOTALL | re.UNICODE), 'double', 'strong,em'),
         util.PatSeqItem(re.compile(SMART_UNDER_STRONG, re.DOTALL | re.UNICODE), 'single', 'strong'),
+        util.PatSeqItem(re.compile(SMART_UNDER_EM2, re.DOTALL | re.UNICODE), 'single', 'em', True),
         util.PatSeqItem(re.compile(SMART_UNDER_EM, re.DOTALL | re.UNICODE), 'single', 'em')
     ]
 
@@ -170,10 +184,9 @@ def make_better(self, md):
 
         config = self.getConfigs()
         enabled = config["smart_enable"]
-        if enabled:
-            enable_all = enabled == "all"
-            enable_under = enabled == "underscore" or enable_all
-            enable_star = enabled == "asterisk" or enable_all
+        enable_all = enabled == "all"
+        enable_under = enabled == "underscore" or enable_all
+        enable_star = enabled == "asterisk" or enable_all
 
         # If we don't have to move an existing extension, use the same priority,
         # but if we do have to, move it closely to the relative needed position.

pymdownx/util.py

@@ -159,50 +159,55 @@ def parse_url(url):
     return (scheme, netloc, path, params, query, fragment, is_url, is_absolute)
 
 
-class PatSeqItem(namedtuple('PatSeqItem', ['pattern', 'builder', 'tags'])):
+class PatSeqItem(namedtuple('PatSeqItem', ['pattern', 'builder', 'tags', 'full_recursion'])):
     """Pattern sequence item item."""
 
+    def __new__(cls, pattern, builder, tags, full_recursion=False):
+        """Create object."""
+
+        return super(PatSeqItem, cls).__new__(cls, pattern, builder, tags, full_recursion)
+
 
 class PatternSequenceProcessor(InlineProcessor):
     """Processor for handling complex nested patterns such as strong and em matches."""
 
     PATTERNS = []
 
-    def build_single(self, m, tag, idx):
+    def build_single(self, m, tag, full_recursion, idx):
         """Return single tag."""
         el1 = etree.Element(tag)
         text = m.group(2)
-        self.parse_sub_patterns(text, el1, None, idx)
+        self.parse_sub_patterns(text, el1, None, full_recursion, idx)
         return el1
 
-    def build_double(self, m, tags, idx):
+    def build_double(self, m, tags, full_recursion, idx):
         """Return double tag."""
 
         tag1, tag2 = tags.split(",")
         el1 = etree.Element(tag1)
         el2 = etree.Element(tag2)
         text = m.group(2)
-        self.parse_sub_patterns(text, el2, None, idx)
+        self.parse_sub_patterns(text, el2, None, full_recursion, idx)
         el1.append(el2)
         if len(m.groups()) == 3:
             text = m.group(3)
-            self.parse_sub_patterns(text, el1, el2, idx)
+            self.parse_sub_patterns(text, el1, el2, full_recursion, idx)
         return el1
 
-    def build_double2(self, m, tags, idx):
+    def build_double2(self, m, tags, full_recursion, idx):
         """Return double tags (variant 2): `<strong>text <em>text</em></strong>`."""
 
         tag1, tag2 = tags.split(",")
         el1 = etree.Element(tag1)
         el2 = etree.Element(tag2)
         text = m.group(2)
-        self.parse_sub_patterns(text, el1, None, idx)
+        self.parse_sub_patterns(text, el1, None, full_recursion, idx)
         text = m.group(3)
         el1.append(el2)
-        self.parse_sub_patterns(text, el2, None, idx)
+        self.parse_sub_patterns(text, el2, None, full_recursion, idx)
         return el1
 
-    def parse_sub_patterns(self, data, parent, last, idx):
+    def parse_sub_patterns(self, data, parent, last, full_recursion, idx):
         """
         Parses sub patterns.
 
@@ -231,7 +236,7 @@ def parse_sub_patterns(self, data, parent, last, idx):
                 # See if the we can match an emphasis/strong pattern
                 for index, item in enumerate(self.PATTERNS):
                     # Only evaluate patterns that are after what was used on the parent
-                    if index <= idx:
+                    if not full_recursion and index <= idx:
                         continue
                     m = item.pattern.match(data, pos)
                     if m:
@@ -245,7 +250,7 @@ def parse_sub_patterns(self, data, parent, last, idx):
                                 last.tail = text
                             else:
                                 parent.text = text
-                        el = self.build_element(m, item.builder, item.tags, index)
+                        el = self.build_element(m, item.builder, item.tags, item.full_recursion, index)
                         parent.append(el)
                         last = el
                         # Move our position past the matched hunk
@@ -266,15 +271,15 @@ def parse_sub_patterns(self, data, parent, last, idx):
             else:
                 parent.text = text
 
-    def build_element(self, m, builder, tags, index):
+    def build_element(self, m, builder, tags, full_recursion, index):
         """Element builder."""
 
         if builder == 'double2':
-            return self.build_double2(m, tags, index)
+            return self.build_double2(m, tags, full_recursion, index)
         elif builder == 'double':
-            return self.build_double(m, tags, index)
+            return self.build_double(m, tags, full_recursion, index)
         else:
-            return self.build_single(m, tags, index)
+            return self.build_single(m, tags, full_recursion, index)
 
     def handleMatch(self, m, data):
         """Parse patterns."""
@@ -288,7 +293,7 @@ def handleMatch(self, m, data):
             if m1:
                 start = m1.start(0)
                 end = m1.end(0)
-                el = self.build_element(m1, item.builder, item.tags, index)
+                el = self.build_element(m1, item.builder, item.tags, item.full_recursion, index)
                 break
         return el, start, end
 

tests/extensions/betterem/betterem (normal).html

@@ -11,7 +11,7 @@
 <p>Test: <em>All will * be italic</em></p>
 <p>Test: <em>All will *be italic</em></p>
 <p>Test: <em>All will not</em> be italic*</p>
-<p>Test: <em>All will not *</em> be italic*</p>
+<p>Test: <em>All will not ** be italic</em></p>
 <p>Test: <strong>All will * be bold</strong></p>
 <p>Test: <em>All will *be italic</em>*</p>
 <p>Test: <strong>All will not</strong>* be bold**</p>

tests/extensions/betterem/betterem (reverse).html

@@ -11,7 +11,7 @@
 <p>Test: <em>All will _ be italic</em></p>
 <p>Test: <em>All will _be italic</em></p>
 <p>Test: <em>All will not</em> be italic_</p>
-<p>Test: <em>All will not _</em> be italic_</p>
+<p>Test: <em>All will not __ be italic</em></p>
 <p>Test: <strong>All will _ be bold</strong></p>
 <p>Test: <em>All will _be italic</em>_</p>
 <p>Test: <strong>All will not</strong>_ be bold__</p>