fix(latex-renderer): make inline code more robust

Currently `\verb` is used to render inline code. However, that has some
downsides as it does not work in moving arguments or within macros like
\multicolumn, and it doesn't break long inline code lines. This patch
changes the implementation to use `\texttt`, which fixes those issues.

mistletoe/latex_renderer.py

@@ -3,19 +3,11 @@
 """
 
 import re
-import string
 from itertools import chain
 from urllib.parse import quote
 import mistletoe.latex_token as latex_token
 from mistletoe.base_renderer import BaseRenderer
 
-# (customizable) delimiters for inline code
-verb_delimiters = string.punctuation + string.digits
-for delimiter in '*':  # remove invalid delimiters
-    verb_delimiters.replace(delimiter, '')
-for delimiter in reversed('|!"\'=+'):  # start with most common delimiters
-    verb_delimiters = delimiter + verb_delimiters.replace(delimiter, '')
-
 
 class LaTeXRenderer(BaseRenderer):
     def __init__(self, *extras, **kwargs):
@@ -27,7 +19,6 @@ def __init__(self, *extras, **kwargs):
         """
         tokens = self._tokens_from_module(latex_token)
         self.packages = {}
-        self.verb_delimiters = verb_delimiters
         super().__init__(*chain(tokens, extras), **kwargs)
 
     def render_strong(self, token):
@@ -37,18 +28,15 @@ def render_emphasis(self, token):
         return '\\textit{{{}}}'.format(self.render_inner(token))
 
     def render_inline_code(self, token):
-        content = self.render_raw_text(token.children[0], escape=False)
-
-        # search for delimiter not present in content
-        for delimiter in self.verb_delimiters:
-            if delimiter not in content:
-                break
+        # fontenc to get better results for `_{}\` in `\texttt{}`
+        self.packages['fontenc'] = ['T1']
 
-        if delimiter in content:  # no delimiter found
-            raise RuntimeError('Unable to find delimiter for verb macro')
+        content = self.render_raw_text(token.children[0], escape=True)
+        # make \texttt behave like \verb w.r.t. whitespace in inline code
+        content = re.sub(r'\s{2,}', lambda m: '\\ '*len(m.group(0)), content)
 
-        template = '\\verb{delimiter}{content}{delimiter}'
-        return template.format(delimiter=delimiter, content=content)
+        template = '\\texttt{{{content}}}'
+        return template.format(content=content)
 
     def render_strikethrough(self, token):
         self.packages['ulem'] = ['normalem']

test/test_latex_renderer.py

@@ -27,21 +27,19 @@ def test_strong(self):
     def test_emphasis(self):
         self._test_token('Emphasis', '\\textit{inner}')
 
-    def test_inline_code(self):
+    @parameterized.expand([
+        ('inner', '\\texttt{inner}'),
+        ('a + b', '\\texttt{a + b}'),
+        ('a | b', '\\texttt{a | b}'),
+        ('|ab!|', '\\texttt{|ab!|}'),
+        ('two  spaces', '\\texttt{two\\ \\ spaces}'),
+        ('two\t whitespaces', '\\texttt{two\\ \\ whitespaces}'),
+    ])
+    def test_inline_code(self, content, expected):
         func_path = 'mistletoe.latex_renderer.LaTeXRenderer.render_raw_text'
 
-        for content, expected in {'inner': '\\verb|inner|',
-                                'a + b': '\\verb|a + b|',
-                                'a | b': '\\verb!a | b!',
-                                '|ab!|': '\\verb"|ab!|"',
-                               }.items():
-            with mock.patch(func_path, return_value=content):
-                self._test_token('InlineCode', expected, content=content)
-
-        content = mistletoe.latex_renderer.verb_delimiters
-        with self.assertRaises(RuntimeError):
-            with mock.patch(func_path, return_value=content):
-                self._test_token('InlineCode', None, content=content)
+        with mock.patch(func_path, return_value=content):
+            self._test_token('InlineCode', expected, content=content)
 
     def test_strikethrough(self):
         self._test_token('Strikethrough', '\\sout{inner}')