Parse srcset attribute values

components/script/dom/htmlimageelement.rs

@@ -57,15 +57,34 @@ use std::char;
 use std::default::Default;
 use std::i32;
 use std::sync::{Arc, Mutex};
-use style::attr::{AttrValue, LengthOrPercentageOrAuto};
+use style::attr::{AttrValue, LengthOrPercentageOrAuto, parse_double, parse_unsigned_integer};
 use style::context::QuirksMode;
 use style::media_queries::MediaQuery;
 use style::parser::ParserContext;
+use style::str::is_ascii_digit;
 use style::values::specified::{Length, ViewportPercentageLength};
 use style::values::specified::length::NoCalcLength;
 use style_traits::ParsingMode;
 use task_source::TaskSource;
 
+enum ParseState {
+    InDescriptor,
+    InParens,
+    AfterDescriptor,
+}
+
+#[derive(Debug, PartialEq)]
+pub struct ImageSource {
+    pub url: String,
+    pub descriptor: Descriptor,
+}
+
+#[derive(Debug, PartialEq)]
+pub struct Descriptor {
+    pub wid: Option<u32>,
+    pub den: Option<f64>,
+}
+
 #[derive(Clone, Copy, HeapSizeOf, JSTraceable)]
 #[allow(dead_code)]
 enum State {
@@ -1047,3 +1066,177 @@ fn image_dimension_setter(element: &Element, attr: LocalName, value: u32) {
     let value = AttrValue::Dimension(value.to_string(), dim);
     element.set_attribute(&attr, value);
 }
+
+/// Collect sequence of code points
+pub fn collect_sequence_characters<F>(s: &str, predicate: F) -> (&str, &str)
+    where F: Fn(&char) -> bool
+{
+    for (i, ch) in s.chars().enumerate() {
+        if !predicate(&ch) {
+            return (&s[0..i], &s[i..])
+        }
+    }
+
+    return (s, "");
+}
+
+/// Parse an `srcset` attribute - https://html.spec.whatwg.org/multipage/#parsing-a-srcset-attribute.
+pub fn parse_a_srcset_attribute(input: &str) -> Vec<ImageSource> {
+    let mut url_len = 0;
+    let mut candidates: Vec<ImageSource> = vec![];
+    while url_len < input.len() {
+        let position = &input[url_len..];
+        let (spaces, position) = collect_sequence_characters(position, |c| *c == ',' || char::is_whitespace(*c));
+        // add the length of the url that we parse to advance the start index
+        let space_len = spaces.char_indices().count();
+        url_len += space_len;
+        if position.is_empty() {
+            return candidates;
+        }
+        let (url, spaces) = collect_sequence_characters(position, |c| !char::is_whitespace(*c));
+        // add the counts of urls that we parse to advance the start index
+        url_len += url.chars().count();
+        let comma_count = url.chars().rev().take_while(|c| *c == ',').count();
+        let url: String = url.chars().take(url.chars().count() - comma_count).collect();
+        // add 1 to start index, for the comma
+        url_len += comma_count + 1;
+        let (space, position) = collect_sequence_characters(spaces, |c| char::is_whitespace(*c));
+        let space_len = space.len();
+        url_len += space_len;
+        let mut descriptors = Vec::new();
+        let mut current_descriptor = String::new();
+        let mut state = ParseState::InDescriptor;
+        let mut char_stream = position.chars().enumerate();
+        let mut buffered: Option<(usize, char)> = None;
+        loop {
+            let next_char = buffered.take().or_else(|| char_stream.next());
+            if next_char.is_some() {
+                url_len += 1;
+            }
+            match state {
+                ParseState::InDescriptor => {
+                    match next_char {
+                        Some((_, ' ')) => {
+                            if !current_descriptor.is_empty() {
+                                descriptors.push(current_descriptor.clone());
+                                current_descriptor = String::new();
+                                state = ParseState::AfterDescriptor;
+                            }
+                            continue;
+                        }
+                        Some((_, ',')) => {
+                            if !current_descriptor.is_empty() {
+                                descriptors.push(current_descriptor.clone());
+                            }
+                            break;
+                        }
+                        Some((_, c @ '(')) => {
+                            current_descriptor.push(c);
+                            state = ParseState::InParens;
+                            continue;
+                        }
+                        Some((_, c)) => {
+                            current_descriptor.push(c);
+                        }
+                        None => {
+                            if !current_descriptor.is_empty() {
+                                descriptors.push(current_descriptor.clone());
+                            }
+                            break;
+                        }
+                    }
+                }
+                ParseState::InParens => {
+                    match next_char {
+                        Some((_, c @ ')')) => {
+                            current_descriptor.push(c);
+                            state = ParseState::InDescriptor;
+                            continue;
+                        }
+                        Some((_, c)) => {
+                            current_descriptor.push(c);
+                            continue;
+                        }
+                        None => {
+                            if !current_descriptor.is_empty() {
+                                descriptors.push(current_descriptor.clone());
+                            }
+                            break;
+                        }
+                    }
+                }
+                ParseState::AfterDescriptor => {
+                    match next_char {
+                        Some((_, ' ')) => {
+                            state = ParseState::AfterDescriptor;
+                            continue;
+                        }
+                        Some((idx, c)) => {
+                            state = ParseState::InDescriptor;
+                            buffered = Some((idx, c));
+                            continue;
+                        }
+                        None => {
+                            if !current_descriptor.is_empty() {
+                                descriptors.push(current_descriptor.clone());
+                            }
+                            break;
+                        }
+                    }
+                }
+            }
+        }
+
+        let mut error = false;
+        let mut width: Option<u32> = None;
+        let mut density: Option<f64> = None;
+        let mut future_compat_h: Option<u32> = None;
+        for descriptor in descriptors {
+            let (digits, remaining) = collect_sequence_characters(&descriptor, |c| is_ascii_digit(c) || *c == '.');
+            let valid_non_negative_integer = parse_unsigned_integer(digits.chars());
+            let has_w = remaining == "w";
+            let valid_floating_point = parse_double(digits);
+            let has_x = remaining == "x";
+            let has_h = remaining == "h";
+            if valid_non_negative_integer.is_ok() && has_w {
+                let result = valid_non_negative_integer;
+                error = result.is_err();
+                if width.is_some() || density.is_some() {
+                    error = true;
+                }
+                if let Ok(w) = result {
+                    width = Some(w);
+                }
+            } else if valid_floating_point.is_ok() && has_x {
+                let result = valid_floating_point;
+                error = result.is_err();
+                if width.is_some() || density.is_some() || future_compat_h.is_some() {
+                    error = true;
+                }
+                if let Ok(x) = result {
+                    density = Some(x);
+                }
+            } else if valid_non_negative_integer.is_ok() && has_h {
+                let result = valid_non_negative_integer;
+                error = result.is_err();
+                if density.is_some() || future_compat_h.is_some() {
+                    error = true;
+                }
+                if let Ok(h) = result {
+                    future_compat_h = Some(h);
+                }
+            } else {
+                error = true;
+            }
+        }
+        if future_compat_h.is_some() && width.is_none() {
+            error = true;
+        }
+        if !error {
+            let descriptor = Descriptor { wid: width, den: density };
+            let image_source = ImageSource { url: url, descriptor: descriptor };
+            candidates.push(image_source);
+        }
+    }
+    candidates
+}

components/script/test.rs

@@ -62,3 +62,7 @@ pub mod size_of {
         size_of::<Text>()
     }
 }
+
+pub mod srcset {
+    pub use dom::htmlimageelement::{parse_a_srcset_attribute, ImageSource, Descriptor};
+}

components/style/str.rs

@@ -58,7 +58,8 @@ pub fn split_commas<'a>(s: &'a str) -> Filter<Split<'a, char>, fn(&&str) -> bool
     s.split(',').filter(not_empty as fn(&&str) -> bool)
 }
 
-fn is_ascii_digit(c: &char) -> bool {
+/// Character is ascii digit
+pub fn is_ascii_digit(c: &char) -> bool {
     match *c {
         '0'...'9' => true,
         _ => false,

tests/unit/script/htmlimageelement.rs

@@ -4,6 +4,7 @@
 
 use script::test::DOMString;
 use script::test::sizes::{parse_a_sizes_attribute, Size};
+use script::test::srcset::{Descriptor, ImageSource, parse_a_srcset_attribute};
 use style::media_queries::{MediaQuery, MediaQueryType};
 use style::media_queries::Expression;
 use style::servo::media_queries::{ExpressionKind, Range};
@@ -99,3 +100,82 @@ fn extra_whitespace() {
         DOMString::from("(max-width: 900px) 1000px,   (max-width: 900px) 50px"),
         None), a);
 }
+
+#[test]
+fn no_value() {
+    let new_vec = Vec::new();
+    assert_eq!(parse_a_srcset_attribute(" "), new_vec);
+}
+
+#[test]
+fn width_one_value() {
+    let first_descriptor = Descriptor { wid: Some(320), den: None };
+    let first_imagesource = ImageSource { url: "small-image.jpg".to_string(), descriptor: first_descriptor };
+    let sources = &[first_imagesource];
+    assert_eq!(parse_a_srcset_attribute("small-image.jpg, 320w"), sources);
+}
+
+#[test]
+fn width_two_value() {
+    let first_descriptor = Descriptor { wid: Some(320), den: None };
+    let first_imagesource = ImageSource { url: "small-image.jpg".to_string(), descriptor: first_descriptor };
+    let second_descriptor = Descriptor { wid: Some(480), den: None };
+    let second_imagesource = ImageSource { url: "medium-image.jpg".to_string(), descriptor: second_descriptor };
+    let sources = &[first_imagesource, second_imagesource];
+    assert_eq!(parse_a_srcset_attribute("small-image.jpg 320w, medium-image.jpg 480w"), sources);
+}
+
+#[test]
+fn width_three_value() {
+    let first_descriptor = Descriptor { wid: Some(320), den: None };
+    let first_imagesource = ImageSource { url: "smallImage.jpg".to_string(), descriptor: first_descriptor };
+    let second_descriptor = Descriptor { wid: Some(480), den: None };
+    let second_imagesource = ImageSource { url: "mediumImage.jpg".to_string(), descriptor: second_descriptor };
+    let third_descriptor = Descriptor { wid: Some(800), den: None };
+    let third_imagesource = ImageSource { url: "largeImage.jpg".to_string(), descriptor: third_descriptor };
+    let sources = &[first_imagesource, second_imagesource, third_imagesource];
+    assert_eq!(parse_a_srcset_attribute("smallImage.jpg 320w,
+                                        mediumImage.jpg 480w,
+                                        largeImage.jpg 800w"), sources);
+}
+
+#[test]
+fn density_value() {
+    let first_descriptor = Descriptor { wid: None, den: Some(1.0) };
+    let first_imagesource = ImageSource { url: "small-image.jpg".to_string(), descriptor: first_descriptor };
+    let sources = &[first_imagesource];
+    assert_eq!(parse_a_srcset_attribute("small-image.jpg 1x"), sources);
+}
+
+#[test]
+fn without_descriptor() {
+    let first_descriptor = Descriptor { wid: None, den: None };
+    let first_imagesource = ImageSource { url: "small-image.jpg".to_string(), descriptor: first_descriptor };
+    let sources = &[first_imagesource];
+    assert_eq!(parse_a_srcset_attribute("small-image.jpg"), sources);
+}
+
+//Does not parse an ImageSource when both width and density descriptor present
+#[test]
+fn two_descriptor() {
+    let empty_vec = Vec::new();
+    assert_eq!(parse_a_srcset_attribute("small-image.jpg 320w 1.1x"), empty_vec);
+}
+
+#[test]
+fn decimal_descriptor() {
+    let first_descriptor = Descriptor { wid: None, den: Some(2.2) };
+    let first_imagesource = ImageSource { url: "small-image.jpg".to_string(), descriptor: first_descriptor };
+    let sources = &[first_imagesource];
+    assert_eq!(parse_a_srcset_attribute("small-image.jpg 2.2x"), sources);
+}
+
+#[test]
+fn different_descriptor() {
+    let first_descriptor = Descriptor { wid: Some(320), den: None };
+    let first_imagesource = ImageSource { url: "small-image.jpg".to_string(), descriptor: first_descriptor };
+    let second_descriptor = Descriptor { wid: None, den: Some(2.2) };
+    let second_imagesource = ImageSource { url: "medium-image.jpg".to_string(), descriptor: second_descriptor };
+    let sources = &[first_imagesource, second_imagesource];
+    assert_eq!(parse_a_srcset_attribute("small-image.jpg 320w, medium-image.jpg 2.2x"), sources);
+}