Minimize bundled user-agent CSS: html.css

third_party/WebKit/Source/build/scripts/minimize_css.py

Index: third_party/WebKit/Source/build/scripts/minimize_css.py
diff --git a/third_party/WebKit/Source/build/scripts/minimize_css.py b/third_party/WebKit/Source/build/scripts/minimize_css.py
new file mode 100755
index 0000000000000000000000000000000000000000..b2cd54feef4fb4c69ec2d85d1decbe544e19cce8
--- /dev/null
+++ b/third_party/WebKit/Source/build/scripts/minimize_css.py
@@ -0,0 +1,128 @@
+#!/usr/bin/env python
+# Copyright 2016 The Chromium Authors. All rights reserved.
+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+import os.path
+import re
+import sys
+
+import in_generator
+
+
+class CSSMinimizer(object):
+
+    INITIAL = 0
+    MAYBE_COMMENT_START = 1
+    INSIDE_COMMENT = 2
+    MAYBE_COMMENT_END = 3
+    INSIDE_SINGLE_QUOTE = 4
+    INSIDE_SINGLE_QUOTE_ESCAPE = 5
+    INSIDE_DOUBLE_QUOTE = 6
+    INSIDE_DOUBLE_QUOTE_ESCAPE = 7
+
+    WHITESPACE_PATTERN = re.compile(r"\s+", re.MULTILINE)

[Timothy Loh, 2016/08/10 05:37:39]

re.MULTILINE doesn't affect either of these regexes. Also I think it's better to
not compile these and just use them directly (iirc the regular python
interpreter will cache them anyway, and for the second regex it's easier to see
the grouping reference)

[kouhei (in TOK), 2016/11/01 10:49:16]

Done.

+    OP_PADDING_PATTERN = re.compile(r";?\s*(?P<op>[\{\};])\s*", re.MULTILINE)

[Timothy Loh, 2016/08/10 05:37:39]

I don't think you want the backslashes in [\{\};]

[kouhei (in TOK), 2016/11/01 10:49:16]

Done.

+
+    def __init__(self):
+        self._output = ''
+        self._codeblock = ''
+
+    def flush_codeblock(self):
+        stripped = re.sub(self.WHITESPACE_PATTERN, ' ', self._codeblock)
+        stripped = re.sub(self.OP_PADDING_PATTERN, r'\g<op>', stripped)
+        self._output += stripped
+        self._codeblock = ''
+
+    def parse(self, content):
+        state = self.INITIAL
+        for c in content:
+            if state == self.INITIAL:
+                if c == '/':
+                    state = self.MAYBE_COMMENT_START

[Timothy Loh, 2016/08/10 05:37:38]

"'"

[kouhei (in TOK), 2016/11/01 10:49:16]

Done.

+                elif c == '\'':
+                    self.flush_codeblock()
+                    self._output += c
+                    state = self.INSIDE_SINGLE_QUOTE
+                elif c == '\"':

[Timothy Loh, 2016/08/10 05:37:38]

'"'

[kouhei (in TOK), 2016/11/01 10:49:16]

Done.

+                    self.flush_codeblock()
+                    self._output += c
+                    state = self.INSIDE_DOUBLE_QUOTE
+                else:
+                    self._codeblock += c
+            elif state == self.MAYBE_COMMENT_START:
+                if c == '*':
+                    self.flush_codeblock()
+                    state = self.INSIDE_COMMENT
+                else:
+                    self._codeblock += '/' + c
+                    state = self.INITIAL
+            elif state == self.INSIDE_COMMENT:
+                if c == '*':
+                    state = self.MAYBE_COMMENT_END
+                else:
+                    pass
+            elif state == self.MAYBE_COMMENT_END:
+                if c == '/':
+                    state = self.INITIAL
+                else:
+                    state = self.INSIDE_COMMENT
+            elif state == self.INSIDE_SINGLE_QUOTE:
+                if c == '\\':
+                    self._output += c
+                    state = self.INSIDE_SINGLE_QUOTE_ESCAPE
+                elif c == '\'':

[Timothy Loh, 2016/08/10 05:37:38]

"'"

[kouhei (in TOK), 2016/11/01 10:49:16]

Done.

+                    self._output += c
+                    state = self.INITIAL
+                else:
+                    self._output += c
+            elif state == self.INSIDE_SINGLE_QUOTE_ESCAPE:
+                self._output += c
+                state = self.INSIDE_SINGLE_QUOTE
+            elif state == self.INSIDE_DOUBLE_QUOTE:
+                if c == '\\':
+                    self._output += c
+                    state = self.INSIDE_DOUBLE_QUOTE_ESCAPE
+                elif c == '"':
+                    self._output += c
+                    state = self.INITIAL
+                else:
+                    self._output += c
+            elif state == self.INSIDE_DOUBLE_QUOTE_ESCAPE:
+                self._output += c
+                state = self.INSIDE_DOUBLE_QUOTE
+
+        self.flush_codeblock()
+        self._output = self._output.strip()
+        return self._output
+
+    @property
+    def output(self):
+        return self._output
+
+    @classmethod
+    def minimize_css(cls, content):
+        minimizer = CSSMinimizer()
+        minimizer.parse(content)
+        return minimizer.output
+
+
+class CSSMinimizerWriter(in_generator.GenericWriter):
+
+    def __init__(self, in_file_paths):
+        super(CSSMinimizerWriter, self).__init__(in_file_paths)
+
+        self._outputs = {}
+        for in_file_path in in_file_paths:
+            out_path = os.path.basename(in_file_path)
+            self._outputs[out_path] = lambda fp=in_file_path: self.generate_implementation(fp)

[Timothy Loh, 2016/08/10 05:37:39]

This line is a bit hard for me to read, I'd use partial from functools instead

[kouhei (in TOK), 2016/11/01 10:49:15]

Done.

+
+    def generate_implementation(self, in_file_path):
+        content = ''
+        with open(os.path.abspath(in_file_path)) as in_file:
+            content = in_file.read()
+        return CSSMinimizer.minimize_css(content)
+
+
+if __name__ == '__main__':
+    in_generator.Maker(CSSMinimizerWriter).main(sys.argv)