auto merge of #4311 : pcwalton/servo/text-transform, r=SimonSapin

The Unicode awareness of `text-transform` is implemented as well as
possible given the Rust standard library's Unicode support. In
particular, the notion of an alphabetic character is used instead of a
letter.

Gecko has a subclass of text run to handle text transforms, but I
implemented this in a simpler way.

r? @SimonSapin

components/layout/text.rs

@@ -22,7 +22,7 @@ use servo_util::smallvec::{SmallVec, SmallVec1};
 use std::collections::DList;
 use std::mem;
 use style::ComputedValues;
-use style::computed_values::{line_height, text_orientation, white_space};
+use style::computed_values::{line_height, text_orientation, text_transform, white_space};
 use style::style_structs::Font as FontStyle;
 use sync::Arc;
 
@@ -104,6 +104,7 @@ impl TextRunScanner {
         let run = {
             let fontgroup;
             let compression;
+            let text_transform;
             {
                 let in_fragment = self.clump.front().unwrap();
                 let font_style = in_fragment.style().get_font_arc();
@@ -111,7 +112,8 @@ impl TextRunScanner {
                 compression = match in_fragment.white_space() {
                     white_space::normal | white_space::nowrap => CompressWhitespaceNewline,
                     white_space::pre => CompressNone,
-                }
+                };
+                text_transform = in_fragment.style().get_inheritedtext().text_transform;
             }
 
             // First, transform/compress text of all the nodes.
@@ -136,6 +138,10 @@ impl TextRunScanner {
                 char_total = char_total + added_chars;
             }
 
+            // Account for `text-transform`. (Confusingly, this is not handled in "text
+            // transformation" above, but we follow Gecko in the naming.)
+            self.apply_style_transform_if_necessary(&mut run_text, text_transform);
+
             // Now create the run.
             //
             // TextRuns contain a cycle which is usually resolved by the teardown sequence.
@@ -176,6 +182,55 @@ impl TextRunScanner {
 
         last_whitespace
     }
+
+    /// Accounts for `text-transform`.
+    ///
+    /// FIXME(#4311, pcwalton): Case mapping can change length of the string; case mapping should
+    /// be language-specific; `full-width`; use graphemes instead of characters.
+    fn apply_style_transform_if_necessary(&mut self,
+                                          string: &mut String,
+                                          text_transform: text_transform::T) {
+        match text_transform {
+            text_transform::none => {}
+            text_transform::uppercase => {
+                let length = string.len();
+                let original = mem::replace(string, String::with_capacity(length));
+                for character in original.chars() {
+                    string.push(character.to_uppercase())
+                }
+            }
+            text_transform::lowercase => {
+                let length = string.len();
+                let original = mem::replace(string, String::with_capacity(length));
+                for character in original.chars() {
+                    string.push(character.to_lowercase())
+                }
+            }
+            text_transform::capitalize => {
+                let length = string.len();
+                let original = mem::replace(string, String::with_capacity(length));
+                let mut capitalize_next_letter = true;
+                for character in original.chars() {
+                    // FIXME(#4311, pcwalton): Should be the CSS/Unicode notion of a *typographic
+                    // letter unit*, not an *alphabetic* character:
+                    //
+                    //    http://dev.w3.org/csswg/css-text/#typographic-letter-unit
+                    if capitalize_next_letter && character.is_alphabetic() {
+                        string.push(character.to_uppercase());
+                        capitalize_next_letter = false;
+                        continue
+                    }
+
+                    string.push(character);
+
+                    // FIXME(#4311, pcwalton): Try UAX29 instead of just whitespace.
+                    if character.is_whitespace() {
+                        capitalize_next_letter = true
+                    }
+                }
+            }
+        }
+    }
 }
 
 struct NewLinePositions(Vec<CharIndex>);

components/style/properties/mod.rs.mako

@@ -1185,6 +1185,9 @@ pub mod longhands {
 
     ${single_keyword("white-space", "normal pre nowrap")}
 
+    // TODO(pcwalton): `full-width`
+    ${single_keyword("text-transform", "none capitalize uppercase lowercase")}
+
     // CSS 2.1, Section 17 - Tables
     ${new_style_struct("Table", is_inherited=False)}
 

tests/ref/basic.list

@@ -187,3 +187,7 @@ fragment=top != ../html/acid2.html acid2_ref.html
 == incremental_float_a.html incremental_float_ref.html
 == opacity_simple_a.html opacity_simple_ref.html
 == opacity_stacking_context_a.html opacity_stacking_context_ref.html
+== text_transform_none_a.html text_transform_none_ref.html
+== text_transform_uppercase_a.html text_transform_uppercase_ref.html
+== text_transform_lowercase_a.html text_transform_lowercase_ref.html
+== text_transform_capitalize_a.html text_transform_capitalize_ref.html

tests/ref/text_transform_capitalize_a.html

@@ -0,0 +1,8 @@
+<!DOCTYPE html>
+<html>
+<!-- Tests that `text-transform: capitalize` works. -->
+<body>
+<h1 style='text-transform: capitalize; font-family: Hiragino Maru Gothic Pro'>ュゥゥゥゥ can do ányThing at ゾムボ.cOm</h1>
+</body>
+</html>
+

tests/ref/text_transform_capitalize_ref.html

@@ -0,0 +1,9 @@
+<!DOCTYPE html>
+<html>
+<!-- Tests that `text-transform: capitalize` works. -->
+<body>
+<h1 style='font-family: Hiragino Maru Gothic Pro'>ュゥゥゥゥ Can Do ÁnyThing At ゾムボ.cOm</h1>
+</body>
+</html>
+
+

tests/ref/text_transform_lowercase_a.html

@@ -0,0 +1,8 @@
+<!DOCTYPE html>
+<html>
+<!-- Tests that `text-transform: lowercase` works. -->
+<body>
+<h1 style='text-transform: lowercase; font-family: Hiragino Maru Gothic Pro'>YoU CaN dO ÁnYtHiNg At ゾムボ.cOm</h1>
+</body>
+</html>
+

tests/ref/text_transform_lowercase_ref.html

@@ -0,0 +1,9 @@
+<!DOCTYPE html>
+<html>
+<!-- Tests that `text-transform: lowercase` works. -->
+<body>
+<h1 style='font-family: Hiragino Maru Gothic Pro'>you can do ánything at ゾムボ.com</h1>
+</body>
+</html>
+
+

tests/ref/text_transform_none_a.html

@@ -0,0 +1,8 @@
+<!DOCTYPE html>
+<html>
+<!-- Tests that `text-transform: none` works. -->
+<body>
+<h1 style='text-transform: none; font-family: Hiragino Maru Gothic Pro'>You can do anything at ゾムボ.com</h1>
+</body>
+</html>
+

tests/ref/text_transform_none_ref.html

@@ -0,0 +1,9 @@
+<!DOCTYPE html>
+<html>
+<!-- Tests that `text-transform: none` works. -->
+<body>
+<h1 style='font-family: Hiragino Maru Gothic Pro'>You can do anything at ゾムボ.com</h1>
+</body>
+</html>
+
+

tests/ref/text_transform_uppercase_a.html

@@ -0,0 +1,8 @@
+<!DOCTYPE html>
+<html>
+<!-- Tests that `text-transform: uppercase` works. -->
+<body>
+<h1 style='text-transform: uppercase; font-family: Hiragino Maru Gothic Pro'>You çan do anything at ゾムボ.com</h1>
+</body>
+</html>
+

tests/ref/text_transform_uppercase_ref.html

@@ -0,0 +1,9 @@
+<!DOCTYPE html>
+<html>
+<!-- Tests that `text-transform: uppercase` works. -->
+<body>
+<h1 style='font-family: Hiragino Maru Gothic Pro'>YOU ÇAN DO ANYTHING AT ゾムボ.COM</h1>
+</body>
+</html>
+
+