DjHTML Version 3

For this release, a substantial part the codebase has been refactored.
Here is a non-exhaustive summary of all the things that have changed:

- Support for Python 3.6 and 3.7 has been dropped
- A Django middleware class to indent HTML responses has been added
- Ignore both opening and closing comment tags
- New handling off both relative and absolute offsets
- Return correct `repr()` strings with the `--debug` option
- New token class "OpenDouble" and a revised indentation algorithm
- Additional `line` argument to `create_token()` methods
- Changed return value of `create_token()` methods
- Changed constructor arguments of Token and Line classes
- Refactored all `create_token()` methods to get rid of return statements
- Simplified test suite runner that reindents the expected output
- Don't indent the contents of template variables (closes #75)
- Improved handling of Django tags inside HTML elements
- Improved JavaScript `case` indentation (closes #76)
- Improved JavaScript method chaining (closes #59)
- Improved CSS multiline statements (closes #74)
- New multiline HTML element indentation (closes #50)
- New multiline HTML attribute value indentation
- Extensive test coverage with lots of edge cases

.github/workflows/main.yml

@@ -9,13 +9,11 @@ on:
 jobs:
   tests:
     name: Python ${{ matrix.python-version }}
-    runs-on: ubuntu-20.04
+    runs-on: ubuntu-22.04
 
     strategy:
       matrix:
         python-version:
-        - 3.6
-        - 3.7
         - 3.8
         - 3.9
         - '3.10'

README.md

@@ -2,8 +2,7 @@
 
 ***A pure-Python Django/Jinja template indenter without dependencies.***
 
-DjHTML is a fully automatic template indenter that works with mixed
-HTML/CSS/Javascript templates that contain
+DjHTML indents mixed HTML/CSS/Javascript templates that contain
 [Django](https://docs.djangoproject.com/en/stable/ref/templates/language/)
 or [Jinja](https://jinja.palletsprojects.com/templates/) template
 tags. It works similar to other code-formatting tools such as
@@ -59,10 +58,10 @@ Install DjHTML with the following command:
 
     $ pip install djhtml
 
-Note that [Windows still uses legacy code pages for the system
-encoding](https://docs.python.org/3/using/windows.html#win-utf8-mode).
-It is highly advised to set the environment variable `PYTHONUTF8` to
-`1` to avoid issues with indenting UTF-8 files. You can do so with the
+Note that
+[Windows still uses legacy code pages](https://docs.python.org/3/using/windows.html#win-utf8-mode)
+instead of UTF-8. It is highly advised to set the environment variable
+`PYTHONUTF8` to `1` with the
 [setx](https://learn.microsoft.com/en-us/windows-server/administration/windows-commands/setx)
 command:
 
@@ -78,12 +77,42 @@ command:
     reindented template.html
     1 template has been reindented.
 
+You can also run `djhtml .` to indent all HTML files beneath the
+current directory.
+
 An exit status of 0 means that everything went well, regardless of
 whether any files were changed. When the option `-c` / `--check` is
 used, the exit status is 1 when one or more files would have changed,
-but no changes are actually made. The exit status of 123 means that
-there was an error while indenting one or more files. All available
-options are given by `djthml -h` / `djthml --help`.
+but no changes are actually made. All available options are given by
+`djthml -h` / `djthml --help`.
+
+
+## New! Multi-line HTML tag alignment
+
+As of version 3, DjHTML indents multi-line HTML tags and multi-line
+attribute values like this:
+
+```html
+<blockquote cite="John Lennon"
+            style="color: dimgray;
+                   font-style: italic;
+                   border-left: 5px solid gray;
+                  ">
+    It's weird not to be weird.
+</blockquote>
+```
+
+
+## New! Django middleware
+
+To automatically indent all the HTML responses from your Django web
+application, add the following to your settings file:
+
+    TABWIDTH = 4
+    MIDDLEWARE += ['djhtml.middleware.DjHTMLMiddleware']
+
+(Caution: when used in production, it is advised to use some kind of
+[caching](https://docs.djangoproject.com/en/stable/topics/cache/).)
 
 
 ## `fmt:off` and `fmt:on`
@@ -98,7 +127,7 @@ You can exclude specific lines from being processed with the
      /     .\
      \_,--._/
     {# fmt:on #}
-"/>
+           "/>
 ```
 
 Contents inside `<pre> ... </pre>`, `<!-- ... --->`, `/* ... */`, and
@@ -124,7 +153,7 @@ The indenter operates in one of three different modes:
 
 ## pre-commit configuration
 
-The best way to use DjHTML is as a [pre-commit](https://pre-commit.com/)
+A great way to use DjHTML is as a [pre-commit](https://pre-commit.com/)
 hook, so all your HTML, CSS and JavaScript files will automatically be
 indented upon every commit.
 
@@ -203,6 +232,7 @@ happy, please do the following:
 
 Your feedback for improving DjHTML is very welcome!
 
+
 ## Development
 
 Use your preferred system for setting up a virtualenv, docker environment,

djhtml/__main__.py

@@ -53,9 +53,6 @@ def main():
 
         # Indent input file
         try:
-            if options.debug:
-                print(Mode(source).debug())
-                sys.exit()
             result = Mode(source).indent(options.tabwidth)
         except Exception:
             _error(
@@ -66,8 +63,7 @@ def main():
             )
             raise
 
-        changed = _verify_changed(source, result)
-        if changed:
+        if changed := _verify_changed(source, result):
             changed_files += 1
         else:
             unchanged_files += 1
@@ -102,6 +98,9 @@ def main():
             f"{problematic_files} template{s} could not be processed due to an error."
         )
 
+    if options.debug:
+        print(Mode(source).debug(), file=sys.stderr)
+
     # Exit with appropriate exit status
     if problematic_files:
         sys.exit(123)

djhtml/lines.py

@@ -2,33 +2,25 @@ class Line:
     """
     A single output line not including the final newline.
 
-    The behavior regarding final newlines has changed between DjHTML
-    v1.4.14 and v1.5.0. It used to always append the final newline,
-    but now this will only happen when the source file already
-    contains a final newline.
-
-    See https://github.com/rtts/djhtml/issues/56 for the discussion
-    that led to this change.
-
     """
 
-    def __init__(self, nr=1):
-        self.nr = nr
-        self.tokens = []
-        self.level = 0
+    def __init__(self, tokens=None, level=0, offset=0, ignore=False):
+        self.tokens = tokens or []
+        self.level = level
+        self.offset = offset
+        self.ignore = ignore
 
     def append(self, token):
         """
         Append tokens to the line.
 
         """
-        token.line_nr = self.nr
         self.tokens.append(token)
 
     @property
     def text(self):
         """
-        The unindented text of this line without leading/trailing spaces.
+        The text of this line without leading/trailing spaces.
 
         """
         return "".join([str(token) for token in self.tokens]).strip()
@@ -39,20 +31,22 @@ def indent(self, tabwidth):
         and optionally offset before calling this method.
 
         """
-        if self.tokens:
-            if self.tokens[0].ignore:
-                return "".join([str(token) for token in self.tokens])
-            elif self.text:
-                offset = self.tokens[0].offset * tabwidth
-                spaces = tabwidth * self.level + offset
-                return " " * spaces + self.text
+        if self.ignore:
+            return "".join([str(token) for token in self.tokens])
+        if self.text:
+            spaces = tabwidth * self.level + self.offset
+            return " " * spaces + self.text
         return ""
 
-    def __repr__(self):
-        return repr(self.tokens)
-
     def __bool__(self):
-        return bool(self.tokens and self.text)
+        return bool(self.tokens)
+
+    def __len__(self):
+        return len(self.text)
 
-    def __next__(self):
-        return Line(nr=self.nr + 1)
+    def __repr__(self):
+        kwargs = ""
+        for attr in ["level", "offset", "ignore"]:
+            if value := getattr(self, attr):
+                kwargs += f", {attr}={value}"
+        return f"Line({repr(self.tokens)}{kwargs})"

djhtml/middleware.py

@@ -0,0 +1,35 @@
+from django.conf import settings
+from django.core.exceptions import ImproperlyConfigured
+
+from djhtml.modes import DjHTML
+
+try:
+    TABWIDTH = settings.TABWIDTH
+except AttributeError:
+    TABWIDTH = 4
+
+
+class DjHTMLMiddleware:
+    """
+    Django middleware class to indent HTML responses.
+
+    """
+
+    def __init__(self, get_response):
+        self.get_response = get_response
+
+    def __call__(self, request):
+        response = self.get_response(request)
+        if (
+            "Content-Type" in response
+            and response["Content-Type"] == "text/html; charset=utf-8"
+        ):
+            if "Content-Length" in response:
+                raise ImproperlyConfigured(
+                    "Please load DjHTMLMiddleware _after_ CommonMiddleware (otherwise"
+                    ' the "Content-Length" header will be incorrect)'
+                )
+            response.content = (
+                DjHTML(response.content.decode()).indent(TABWIDTH).encode()
+            )
+        return response