From 95d962bb09180df82cfc1ec93ad476df3fa327ef Mon Sep 17 00:00:00 2001
From: Derek Rhodes <physci@gmail.com>
Date: Sun, 22 Dec 2024 17:31:13 -0500
Subject: [PATCH] This project now uses HTML as the base languages instead of
 XML.

XML is too strict.
---
 pyproject.toml             | 10 +++++++++-
 src/auxml/macro_manager.py |  8 +++++---
 src/auxml/main.py          |  2 +-
 src/auxml/parser.py        | 14 ++++++++------
 tests/macro_test.py        |  6 +++++-
 tests/macrofile1.html      |  9 +++++++++
 6 files changed, 37 insertions(+), 12 deletions(-)
 create mode 100644 tests/macrofile1.html

diff --git a/pyproject.toml b/pyproject.toml
index bbcb375..b43421c 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -6,7 +6,15 @@ build-backend = "setuptools.build_meta"
 [project]
 name = "auxml"
 version = "0.1.0"
-dependencies = [ "lxml", "pytest", "pudb", "pytest-watcher", "xmldiff", "livereload" ]
+dependencies = [
+ "lxml",
+ "pytest",
+ "pudb",
+ "pytest-watcher",
+ "xmldiff",
+ "livereload",
+ "beautifulsoup4",
+ ]
 requires-python = ">=3.8"
 authors = [ {name = "Derek A. Rhodes", email = "physci@gmail.com"} ]
 description = "An author friendly macro system for XML"
diff --git a/src/auxml/macro_manager.py b/src/auxml/macro_manager.py
index bef3c73..73fcfbe 100644
--- a/src/auxml/macro_manager.py
+++ b/src/auxml/macro_manager.py
@@ -4,7 +4,7 @@
 from auxml.macro import MacroDef, MacroCall
 from auxml.util import *
 from auxml.html_tags import is_an_html_tag
-from auxml.parser import parse_xml_file
+from auxml.parser import parse_html_file
 
 from auxml.err import SyntaxErrorAuXML
 
@@ -17,7 +17,7 @@ def register_directive(self, name, cls):
         self.directives[name] = cls
         
     def load_macro_file(self, infile):
-        macros = parse_xml_file(infile)
+        macros = parse_html_file(infile)
         for el in macros.findall(".//define-macro"):
             md = MacroDef(el)
             self.add_macro_def(md)
@@ -28,7 +28,9 @@ def add_macro_def(self, macdef):
         if macdef.name in self.macro_defs:
             raise Exception(f"macro already defined: {macdef.name}")
         if len(macdef.el.getchildren()) > 1:
-            raise SyntaxErrorAuXML(f"macro `{macdef.name}` may not have more than one child element")
+            msg = f"macro `{macdef.name}` may not have more than one child element\n"
+            msg += f"Got `{macdef.el.getchildren()}` child elements\n"
+            raise SyntaxErrorAuXML(msg)
         self.macro_defs[macdef.name] = macdef
 
     def cant_find(self, tag):
diff --git a/src/auxml/main.py b/src/auxml/main.py
index 01e4b66..e66a96a 100644
--- a/src/auxml/main.py
+++ b/src/auxml/main.py
@@ -18,7 +18,7 @@ def __init__(self):
         mm.load_macro_file(cmdline.macros())
 
         fname = cmdline.infile()
-        root = parse_xml_file(fname)
+        root = parse_html_file(fname)
         
         exp = mm.expand_all(fname, root)
         self.save(exp)
diff --git a/src/auxml/parser.py b/src/auxml/parser.py
index 83bde81..f6d001c 100644
--- a/src/auxml/parser.py
+++ b/src/auxml/parser.py
@@ -1,13 +1,15 @@
 from lxml import etree
-
+from bs4 import BeautifulSoup
 _html_parser = etree.HTMLParser(remove_blank_text=True, remove_comments=True)
 _xml_parser = etree.XMLParser(remove_blank_text=True, remove_comments=True)
 
 def parse_html_file(fname):
-    tree = etree.parse(fname, _html_parser)
-    return tree.getroot()
+    text = open(fname).read()
+    soup = BeautifulSoup(text, "html.parser")
+    tree = etree.fromstring(str(soup), _xml_parser)
+    return tree
    
-def parse_xml_file(fname):
-    tree = etree.parse(fname, _xml_parser)
-    return tree.getroot()
+# def parse_xml_file(fname):
+#     tree = etree.parse(fname, _xml_parser)
+#     return tree.getroot()
 
diff --git a/tests/macro_test.py b/tests/macro_test.py
index 01c0e9f..d8195b6 100644
--- a/tests/macro_test.py
+++ b/tests/macro_test.py
@@ -234,6 +234,10 @@ def test_mac_multi_rearrange_bug():
     e = '''<div><div>A</div>B</div>'''
     #pudb.set_trace()
     with_mm([m], c, e)
-    
 
+def with_mm_file(filename):
+    mm = MacroManager()
+    mm.load_macro_file(filename)
     
+def test_mac_multi_els_with_html_bug():
+    with_mm_file("tests/macrofile1.html")
diff --git a/tests/macrofile1.html b/tests/macrofile1.html
new file mode 100644
index 0000000..9bdd2ed
--- /dev/null
+++ b/tests/macrofile1.html
@@ -0,0 +1,9 @@
+<macros>
+  <define-macro name="weird">
+    <head>
+      <div>&</div>
+      <div></div>
+    </head>
+  </define-macro>
+</macros>
+