From cfbab91524b7f7718c264b52fcda565bb6d1ae61 Mon Sep 17 00:00:00 2001
From: Jerome Dumonteil <jerome.dumonteil@gmail.com>
Date: Sat, 30 Nov 2024 20:35:37 +0100
Subject: [PATCH] markdown: fix successive bold tags

---
 odfdo/mixin_md.py      | 16 +++++++++++++---
 tests/test_markdown.py | 20 ++++++++++----------
 2 files changed, 23 insertions(+), 13 deletions(-)

diff --git a/odfdo/mixin_md.py b/odfdo/mixin_md.py
index 39a7ddc..cdce858 100644
--- a/odfdo/mixin_md.py
+++ b/odfdo/mixin_md.py
@@ -21,12 +21,17 @@
 """
 from __future__ import annotations
 
+import re
 from copy import deepcopy
 from itertools import chain
 from typing import Any, Callable, NamedTuple
 
 MD_GLOBAL = {}
 
+RE_STAR6 = re.compile(r"(?<!\\)(\*{6})")
+RE_STAR4 = re.compile(r"(?<!\\)(\*{4})")
+RE_UND2 = re.compile(r"(?<!\\)(_{2})")
+
 
 class LIStyle(NamedTuple):
     name: str
@@ -74,7 +79,7 @@ def _release_list_counter(level: int) -> None:
 
 
 def _strip_left_spaces(text: str) -> str:
-    return text.lstrip(" ")
+    return RE_STAR4.sub("", RE_STAR6.sub("", RE_UND2.sub("", text.lstrip(" "))))
 
 
 def _md_swap_spaces(word: str) -> SplitSpace:
@@ -95,11 +100,16 @@ def _md_escape(text: str | None) -> str:
     if not text:
         return ""
     return (
-        text.replace(" ", " ")  # non break space is no understood as char
+        text.replace(" ", r" ")  # non break space is no understood as char
         .replace("#", r"\#")
         .replace(r"\*", "*")
         .replace("*", r"\*")
+        .replace(r"\_", r"_")
+        .replace("_", r"\_")
+        .replace("-", r"\-")
+        .replace(r"\`", "`")
         .replace("`", r"\`")
+        .replace(r"\~", "~")
         .replace("~", r"\~")
         .replace("|", r"\|")
     )
@@ -110,7 +120,7 @@ def _as_italic(text: str | None) -> str:
     if not text.strip():
         return text
     word = _md_swap_spaces(text)
-    return f"{word.start}*{word.word}*{word.end}"
+    return f"{word.start}_{word.word}_{word.end}"
 
 
 def _as_bold(text: str | None) -> str:
diff --git a/tests/test_markdown.py b/tests/test_markdown.py
index 5a1ae9d..852226e 100755
--- a/tests/test_markdown.py
+++ b/tests/test_markdown.py
@@ -275,7 +275,7 @@ def test_md_list_text(document_list):
         Some text
 
          -  une liste accentuée
-            -  un sous-élément
+            -  un sous\\-élément
 
          1. une liste numérotée
          2. et de deux !
@@ -354,7 +354,7 @@ def test_user_fields_text(document_uf):
     md = document_uf.to_markdown()
     expected = dedent(
         """\
-    A document with user-field declarations.
+    A document with user\\-field declarations.
 
     Paris
 
@@ -400,7 +400,7 @@ def test_md_case1_text(document_case1):  # FAIL
 
         This is a paragraph with a named style.
 
-        Some list :
+        Some list :
 
          -  item1
          -  item2
@@ -436,7 +436,7 @@ def test_md_case2_text(document_case2):  # FAIL
         """\
         # odfdo Test Case2
 
-        Some style `here` and *there*
+        Some style `here` and _there_
 
         Two spaces at start
 
@@ -450,11 +450,11 @@ def test_md_case2_text(document_case2):  # FAIL
 
         This is a **bold** **paragraph** with a named style.
 
-        Here  **bolded space** at *begin*.
+        Here  **bolded space** at _begin_.
 
-        Some *special* ***case*** *to* check if ~~barred~~
+        Some _special_ ***case*** _to_ check if ~~barred~~
 
-        Some list :
+        Some list :
 
          -  item1
          -  **item2**
@@ -473,7 +473,7 @@ def test_md_case2_text(document_case2):  # FAIL
 
             3. c
 
-         -  *item6*
+         -  _item6_
          -  [x] done
 
             1. u
@@ -483,7 +483,7 @@ def test_md_case2_text(document_case2):  # FAIL
          -  [ ] undone
          -
 
-        last paragraph *with italic*
+        last paragraph _with italic_
 
         \\#\\#\\#\\# To see **\\*\\*stars\\*\\*** or \\*\\*any\\*\\*
 
@@ -546,7 +546,7 @@ def test_md_case3_text(document_case3):  # FAIL
 
         ```
         another one,
-        with *thing* and other things
+        with _thing_ and other things
         in it
          tabs	in same place
               spaces