SF.net SVN: docutils:[9743 ] trunk/docutils

Revision: 9743
          http://sourceforge.net/p/docutils/code/9743
Author:   milde
Date:     2024-06-06 14:03:21 +0000 (Thu, 06 Jun 2024)
Log Message:
-----------
xml-parser: Unindent continuation lines. Fixes xml round trip.

The XML writer supports formatting XML with newlines and indentation
(see the corresponding configuration settings).

In most reStructuredText markup blocks, indentation has a semantic
importance and is removed before passing the block for further parsing.
As a result, only a limited set of elements may have text with
spaces after a newline.

Remove indentation of continuation lines unless the containing element
is a `FixedTextElement`, `literal`, or `Unknown` instance.

Modified Paths:
--------------
    trunk/docutils/docutils/parsers/docutils_xml.py
    trunk/docutils/test/test_parsers/test_docutils_xml/test_parse.py

Modified: trunk/docutils/docutils/parsers/docutils_xml.py
===================================================================
--- trunk/docutils/docutils/parsers/docutils_xml.py	2024-06-06 14:03:12 UTC (rev 9742)
+++ trunk/docutils/docutils/parsers/docutils_xml.py	2024-06-06 14:03:21 UTC (rev 9743)
@@ -19,6 +19,7 @@
    without prior notice.
 """
 
+import re
 import xml.etree.ElementTree as ET
 
 from docutils import frontend, nodes, parsers, utils
@@ -97,10 +98,16 @@
     return element2node(root, document)
 
 
-def element2node(element, document=None):
+def element2node(element, document=None, unindent=True):
     """
     Convert an `etree` element and its children to Docutils doctree nodes.
 
+    :element:  `xml.etree` element
+    :document: see `parse_element()`
+    :unindent: Remove formatting indentation of follow-up text lines?
+               Cf. `append_text()`.
+               TODO: do we need an "unindent" configuration setting?
+
     Return a `docutils.nodes.Element` instance.
 
     Internal.
@@ -141,21 +148,28 @@
                 value = value.split()
             node.attributes[key] = value  # node becomes invalid!
 
-    # Append text (wrapped in a `nodes.Text` instance)
-    append_text(node, element.text)
-
-    # Append children and their tailing text
+    # Append content:
+    # update "unindent" flag: change line indentation?
+    unindent = unindent and not isinstance(
+                   node, (nodes.FixedTextElement, nodes.literal, Unknown))
+    # (leading) text
+    append_text(node, element.text, unindent)
+    # children and their tailing text
     for child in element:
-        node.append(element2node(child, document))
+        node.append(element2node(child, document, unindent))
         # Text after a child node
-        append_text(node, child.tail)
+        append_text(node, child.tail, unindent)
 
     return node
 
 
-def append_text(node, text):
+def append_text(node, text, unindent):
+    # Format `text`, wrap in a TextElement and append to `node`.
+    # Skip if `text` is empty or just formatting whitespace.
     if not text:
         return
+    if unindent:
+        text = re.sub('\n +', '\n', text)
     if isinstance(node, nodes.TextElement):
         node.append(nodes.Text(text))
     elif text.strip():

Modified: trunk/docutils/test/test_parsers/test_docutils_xml/test_parse.py
===================================================================
--- trunk/docutils/test/test_parsers/test_docutils_xml/test_parse.py	2024-06-06 14:03:12 UTC (rev 9742)
+++ trunk/docutils/test/test_parsers/test_docutils_xml/test_parse.py	2024-06-06 14:03:21 UTC (rev 9743)
@@ -59,12 +59,11 @@
         with two lines.</paragraph>
 </document>
 """,
-# TODO: unindent second line
 """\
 <document source="test sample">
     <paragraph>
         A paragraph
-                with two lines.
+        with two lines.
 """],
 ["""\
 <document>
@@ -111,6 +110,25 @@
             strong
          week
 """],
+["""\
+<document source="test data">
+    <literal_block xml:space="preserve">  Inline element
+with <strong>  space at start,
+    in the middle</strong>
+  and after end.</literal_block>
+</document>
+""",
+"""\
+<document source="test data">
+    <literal_block xml:space="preserve">
+          Inline element
+        with \n\
+        <strong>
+              space at start,
+                in the middle
+        \n\
+          and after end.
+"""],
 ]
 
 totest['attributes'] = [

This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.