Preserve HTML blocks in po2md output (#246)

* Preserve HTML blocks in po2md output
mondeja · Sep 20, 2022 · eedb0b5 · eedb0b5
1 parent a32dce6
commit eedb0b5
Show file tree

Hide file tree

Showing 10 changed files with 70 additions and 15 deletions.
diff --git a/docs/cli.rst b/docs/cli.rst
@@ -51,6 +51,7 @@ The output produced by :ref:`cli:po2md` is compatible with the following
    {
      "no-blanks-blockquote": false,
      "no-bare-urls": false,
+     "no-inline-html": false,
      "ul-indent": {
        "indent": 3
      }

diff --git a/docs/pre-commit-hooks.rst b/docs/pre-commit-hooks.rst
@@ -19,7 +19,7 @@ so you don't need to specify them.
       .. code-block:: yaml
 
          - repo: https://github.com/mondeja/mdpo
-           rev: v1.0.2
+           rev: v1.0.3
            hooks:
              - id: md2po
                args:
@@ -32,7 +32,7 @@ so you don't need to specify them.
       .. code-block:: yaml
 
          - repo: https://github.com/mondeja/mdpo
-           rev: v1.0.2
+           rev: v1.0.3
            hooks:
              - id: md2po
                files: ^README\.md
@@ -53,7 +53,7 @@ po2md
       .. code-block:: yaml
 
          - repo: https://github.com/mondeja/mdpo
-           rev: v1.0.2
+           rev: v1.0.3
            hooks:
              - id: po2md
                args:
@@ -68,7 +68,7 @@ po2md
       .. code-block:: yaml
 
          - repo: https://github.com/mondeja/mdpo
-           rev: v1.0.2
+           rev: v1.0.3
            hooks:
              - id: po2md
                files: ^README\.md
@@ -91,7 +91,7 @@ md2po2md
       .. code-block:: yaml
 
          - repo: https://github.com/mondeja/mdpo
-           rev: v1.0.2
+           rev: v1.0.3
            hooks:
              - id: md2po2md
                args:
@@ -107,7 +107,7 @@ md2po2md
       .. code-block:: yaml
 
          - repo: https://github.com/mondeja/mdpo
-           rev: v1.0.2
+           rev: v1.0.3
            hooks:
              - id: md2po2md
                files: ^README\.md
@@ -126,7 +126,7 @@ mdpo2html
       .. code-block:: yaml
 
          - repo: https://github.com/mondeja/mdpo
-           rev: v1.0.2
+           rev: v1.0.3
            hooks:
              - id: mdpo2html
                args:
@@ -141,7 +141,7 @@ mdpo2html
       .. code-block:: yaml
 
          - repo: https://github.com/mondeja/mdpo
-           rev: v1.0.2
+           rev: v1.0.3
            hooks:
              - id: mdpo2html
                files: ^README\.html

diff --git a/pyproject.toml b/pyproject.toml
@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "mdpo"
-version = "1.0.2"
+version = "1.0.3"
 description = "Markdown files translation using PO files."
 readme = "README.md"
 license = "BSD-3-Clause"

diff --git a/src/mdpo/po2md/__init__.py b/src/mdpo/po2md/__init__.py
@@ -247,7 +247,10 @@ def __init__(self, pofiles, ignore=frozenset(), po_encoding=None, **kwargs):
             self._leavespan_replacer[md4c.SpanType.WIKILINK.value] = \
                 self.wikilink_end_string
 
-        self._inside_htmlblock = False
+        self._inside_htmlblock = [
+            False,  # the parser is inside an HTML block
+            None,  # a mdpo command has been found in the HTML block
+        ]
         self._inside_codeblock = False
         self._inside_indented_codeblock = False
         self._inside_hblock = False
@@ -326,7 +329,9 @@ def command(self, mdpo_command, comment, original_command):
     def _process_command(self, text):
         original_command, comment = parse_mdpo_html_command(text)
         if original_command is None:
-            return
+            return False
+
+        self._inside_htmlblock[1] = True
 
         try:
             command = self.command_aliases[original_command]
@@ -336,6 +341,8 @@ def _process_command(self, text):
         # process solved command
         self.command(command, comment, original_command)
 
+        return True
+
     def _escape_translation(self, text):
         if self._aimg_title_inside_current_msgid:
             # escape '"' characters inside links and image titles
@@ -561,7 +568,7 @@ def enter_block(self, block, details):
             if self._current_list_type and not self._inside_quoteblock:
                 self._save_current_line()
         elif block is md4c.BlockType.HTML:
-            self._inside_htmlblock = True
+            self._inside_htmlblock = [True, False]
 
     def leave_block(self, block, details):
         # raise 'leave_block' event
@@ -681,7 +688,11 @@ def leave_block(self, block, details):
                 self._save_current_line()
             self._current_thead_aligns = []
         elif block is md4c.BlockType.HTML:
-            self._inside_htmlblock = False
+            if not self._inside_htmlblock[1]:
+                self.current_line += '\n'
+            else:
+                self.current_line = self.current_line.rstrip('\n')
+            self._inside_htmlblock[0] = False
 
     def enter_span(self, span, details):
         # raise 'enter_span' event
@@ -851,7 +862,7 @@ def text(self, block, text):
         ):
             return
 
-        if not self._inside_htmlblock:
+        if not self._inside_htmlblock[0]:
             if not self._inside_codeblock:
                 if self._inside_liblock and text == '\n':
                     text = ' '
@@ -903,7 +914,8 @@ def text(self, block, text):
                         self.current_line += indent
                 self.current_msgid += text
         else:
-            self._process_command(text)
+            if not self._process_command(text):
+                self.current_line += text
 
     def _append_link_references(self):
         if self.link_references:

diff --git a/tests/test_unit/test_md2po/extract-examples/markuptext/html.md b/tests/test_unit/test_md2po/extract-examples/markuptext/html.md
@@ -0,0 +1,7 @@
+Some text
+
+<div id="my-special-div">
+content
+</div>
+
+More text
diff --git a/tests/test_unit/test_md2po/extract-examples/markuptext/html.md.expect.po b/tests/test_unit/test_md2po/extract-examples/markuptext/html.md.expect.po
@@ -0,0 +1,9 @@
+#
+msgid ""
+msgstr ""
+
+msgid "Some text"
+msgstr ""
+
+msgid "More text"
+msgstr ""
diff --git a/tests/test_unit/test_po2md/.markdownlint-cli2.yaml b/tests/test_unit/test_po2md/.markdownlint-cli2.yaml
@@ -41,3 +41,6 @@ config:
   # Link fragments could not be valid in certain tests:
   # https://github.com/DavidAnson/markdownlint/blob/main/doc/Rules.md#md051
   link-fragments: false
+
+  # HTML blocks are preserved in output
+  no-inline-html: false
diff --git a/tests/test_unit/test_po2md/translate-examples/html.md b/tests/test_unit/test_po2md/translate-examples/html.md
@@ -0,0 +1,7 @@
+Some text
+
+<div id="my-special-div">
+content
+</div>
+
+More text
diff --git a/tests/test_unit/test_po2md/translate-examples/html.md.expect.md b/tests/test_unit/test_po2md/translate-examples/html.md.expect.md
@@ -0,0 +1,7 @@
+Algo de texto
+
+<div id="my-special-div">
+content
+</div>
+
+Más texto
diff --git a/tests/test_unit/test_po2md/translate-examples/html.po b/tests/test_unit/test_po2md/translate-examples/html.po
@@ -0,0 +1,9 @@
+#
+msgid ""
+msgstr ""
+
+msgid "Some text"
+msgstr "Algo de texto"
+
+msgid "More text"
+msgstr "Más texto"