WIP #17808

Rakhisharma · 2017-07-20T21:19:57Z

Parse sizes attribute implementation.

./mach build -d does not report any errors
./mach test-tidy does not report any errors
These changes fix #11416 (github issue number if applicable).

There are tests for these changes OR
These changes do not require tests because _____

This change is

highfive · 2017-07-20T21:20:03Z

Thanks for the pull request, and welcome! The Servo team is excited to review your changes, and you should hear from @asajeffrey (or someone else) soon.

highfive · 2017-07-20T21:20:04Z

Heads up! This PR modifies the following files:

@fitzgen: components/script/dom/htmlimageelement.rs
@KiChjang: components/script/dom/htmlimageelement.rs

highfive · 2017-07-20T21:20:05Z

Warning

These commits modify script code, but no tests are modified. Please consider adding a test!

bors-servo · 2017-07-21T23:55:51Z

☔ The latest upstream changes (presumably #17822) made this pull request unmergeable. Please resolve the merge conflicts.

jdm

Good start!

jdm · 2017-07-27T19:18:47Z

components/script/dom/htmlimageelement.rs

@@ -782,6 +797,71 @@ impl LayoutHTMLImageElementHelpers for LayoutJS<HTMLImageElement> {
    }
 }

+pub fn parse_a_sizes_attribute(input: DOMString, width: Option<u32>) -> Result<Vec<Size>, String> {


Since we only return Ok from this method (and not Err), there is no reason to return a Result value.

jdm · 2017-07-27T19:18:47Z

components/script/dom/htmlimageelement.rs

+        let trimmed: String = unparsed_size.chars().take(unparsed_size.chars().count() - whitespace).collect();
+
+        if trimmed.is_empty() {
+            // return Err("reason".to_owned());


Please remove code instead of commenting it out.

jdm · 2017-07-27T19:18:47Z

components/script/dom/htmlimageelement.rs

+            },
+        }
+    }
+    if sizes.len() == 0 {


nit: use sizes.is_empty() instead.

jdm · 2017-07-27T19:18:47Z

tests/unit/script/htmlimageelement.rs

+
+#[test]
+fn empty_vector() {
+    assert!(parse_a_sizes_attribute(DOMString::new(), None).is_ok());


We will want to assert that the result is an empty vector (as the test name implies ;))

@Rakhisharma As discussed, the test case depicts the situation where sizes is an empty string (DOMString::new()), so here we should also check if the vector contains a single Size with the value 100vw. (i.e., the default value returned from spec algorithm)

jdm · 2017-07-27T19:18:47Z

tests/unit/script/htmlimageelement.rs

+#[test]
+fn test_whitespace() {
+    assert!(parse_a_sizes_attribute(DOMString::from("     (min-width: 500px)"),
+            None).is_ok());


Let's think about what this test is checking. (min-width: 500px) by itself is not a valid size; it is a media query, and the sizes algorithm then tries to parse a length that follows it. Since there is no length present, this string would be considered a parse error. That's fine, but we should also verify in a separate test that a valid size with extra whitespace parses correctly and gives us a real parsed size.

jdm · 2017-07-27T19:18:47Z

components/script/dom/htmlimageelement.rs

+                let media_query = media_query_parser.try(|i| MediaQuery::parse(&context, i));
+                match media_query {
+                    Ok(query) => {
+                        let length = media_query_parser.try(|i| Length::parse_non_negative(&context, i));


We can use Length::parse_non_negative directly instead of Parser::try: let length = Length::parse_non_negative(&context, &mut media_query_parser);

jdm · 2017-07-27T19:18:47Z

components/script/dom/htmlimageelement.rs

+                match media_query {
+                    Ok(query) => {
+                        let length = media_query_parser.try(|i| Length::parse_non_negative(&context, i));
+                        if length.is_ok() {


Instead of using is_ok and unwrap, we can write if let Ok(length) = length {.

jdm · 2017-07-27T19:18:47Z

components/script/dom/htmlimageelement.rs

+                let mut media_query_parser = Parser::new(&mut input);
+                let media_query = media_query_parser.try(|i| MediaQuery::parse(&context, i));
+                match media_query {
+                    Ok(query) => {


Since the Err branch of this match is empty except for continue, we can write if let Ok(query) = media_query { and omit it entirely.

jdm · 2017-07-27T19:18:47Z

components/script/dom/htmlimageelement.rs

@@ -66,6 +76,11 @@ enum State {
    CompletelyAvailable,
    Broken,
 }
+#[allow(dead_code)]


This probably is not considered dead code any more, so we can remove it.

jdm · 2017-07-27T19:18:47Z

components/script/dom/htmlimageelement.rs

@@ -3,6 +3,8 @@
 * file, You can obtain one at http://mozilla.org/MPL/2.0/. */

 use app_units::{Au, AU_PER_PX};
+use core::ops::Deref;


This can be removed.

we are using this at line: 802: let unparsed_sizes = input.deref().split(',').collect::<Vec<_>>();

An explicit call to deref should not be necessary, if let unparsed_sizes = input.split(',').collect::<Vec<_>>(); does not work, try let unparsed_sizes = (*input).split(',').collect::<Vec<_>>();.

let unparsed_sizes = input.split(',').collect::<Vec<_>>(); worked! thanks.

creativcoder · 2017-07-30T04:47:08Z

components/script/dom/htmlimageelement.rs

+                                                        None,
+                                                        ParsingMode::empty(),
+                                                        QuirksMode::NoQuirks);
+        let url = ServoUrl::parse("about:blank").unwrap();


You already have url declared in line 813. So you can remove this.

creativcoder · 2017-07-30T04:50:45Z

tests/unit/script/lib.rs

@@ -12,4 +12,6 @@ extern crate servo_url;
 #[cfg(test)] mod textinput;
 #[cfg(test)] mod headers;
 #[cfg(test)] mod htmlareaelement;
+#[cfg(test)] mod htmlimageelement; 
+


nit: Remove the extra newline.

bors-servo · 2017-08-23T23:25:22Z

☔ The latest upstream changes (presumably #18179) made this pull request unmergeable. Please resolve the merge conflicts.

jdm

This is a good start! The tests can use a bit more refining, but I think we're close to having something that's ready to be merged!

jdm · 2017-08-31T16:50:05Z

components/script/dom/htmlimageelement.rs

+                                                   None,
+                                                   ParsingMode::empty(),
+                                                   QuirksMode::NoQuirks);
+        let length = Parser::new(&mut input).try(|i| Length::parse_non_negative(&context, i));


We should create a separate parser object that we can reuse later.

How to create a separate parser object here? I am not clear if we can create a Parser object without any input string and reuse it later.

let mut parser = Parser::new(&mut input); /* ... */ let length = parser.try(..); /* .. */ let something_else = parser.foo(..).

jdm · 2017-08-31T16:50:05Z

components/script/dom/htmlimageelement.rs

+                query: None
+            }),
+            Err(_) => {
+                let mut media_query_parser = Parser::new(&mut input);


We can reuse the earlier parser instead.

jdm · 2017-08-31T16:50:05Z

components/script/dom/htmlimageelement.rs

+                let media_query = media_query_parser.try(|i| MediaQuery::parse(&context, i));
+                if let Ok(query) = media_query {
+                        let length = Length::parse_non_negative(&context, &mut media_query_parser);
+                        if length.is_ok() {


We can use if let Ok(length) = length { here and avoid calling unwrap below.

jdm · 2017-08-31T16:50:05Z

components/script/dom/htmlimageelement.rs

@@ -782,6 +798,63 @@ impl LayoutHTMLImageElementHelpers for LayoutJS<HTMLImageElement> {
    }
 }

+pub fn parse_a_sizes_attribute(input: DOMString, width: Option<u32>) -> Vec<Size> {


Add a comment like // https://html.spec.whatwg.org/multipage/#parse-a-sizes-attribute so it's clear which algorithm this is based on.

jdm · 2017-08-31T16:50:05Z

components/script/dom/htmlimageelement.rs

+    make_getter!(Sizes, "sizes");
+    // https://html.spec.whatwg.org/multipage/#parse-a-sizes-attribute
+    make_setter!(SetSizes, "sizes");
+


nit: remove this extra newline.

jdm · 2017-08-31T16:50:05Z

tests/unit/script/htmlimageelement.rs

+
+#[test]
+fn no_default_provided() {
+    assert!(parse_a_sizes_attribute(DOMString::new(), None).len() == 1);


Instead of asserting the length, we should be asserting the actual value that is returned in all of these tests.

jdm · 2017-08-31T16:50:05Z

tests/unit/script/htmlimageelement.rs

+
+#[test]
+fn no_size() {
+    assert!(parse_a_sizes_attribute(DOMString::new(), None).len() == 1);


This is the same as no_default_provided.

jdm · 2017-08-31T16:50:05Z

tests/unit/script/htmlimageelement.rs

+    let length = test_length(545f32);
+    let size = Size { query: Some(media_query), length: length };
+    a.push(size);
+    assert_eq!(parse_a_sizes_attribute(DOMString::from(" (max-width: 200px) 545px"), None), a);


It's confusing how this is called no_extra_whitespace but the test includes a leading space.

Here what you mean with the extra whitespace, "does the middle spaces also count as extra whitespace or only the leading and trailing spaces counts as extra whitespace ?"

Extra whitespace generally means any additional whitespace beyond what is necessary to separate two tokens, regardless of where the whitespace exists in the string.

jdm · 2017-08-31T16:50:05Z

tests/unit/script/htmlimageelement.rs

+            (max-width: 900px) and (min-width: 400px) 50em,
+            100vw           "),
+            None);
+    assert_eq!(result.len(), 3);


We should definitely be checking the actual values that are returned here.

jdm · 2017-08-31T16:50:05Z

tests/unit/script/htmlimageelement.rs

+    let length = test_length(545f32);
+    let size = Size { query: Some(media_query), length: length };
+    a.push(size);
+    assert_eq!(parse_a_sizes_attribute(DOMString::from("(max-width: 320px) 280px,     (max-width: 480px) 440px, 800px"), None), a);


Why does this test pass? We are parsing a string that contains three different sizes, but the expected results only contain one size that does not match any of them.

KiChjang · 2017-09-01T18:45:35Z

Sorry, but which issue does this PR solve?

creativcoder · 2017-09-01T18:58:52Z

This is part of #11416


        Parse Sizes Attribute


        Parse Sizes Attribute

wafflespeanut

Nice work! This mostly looks good to me. Just a few nits, and we'll be good to go.

wafflespeanut · 2017-09-24T09:29:46Z

components/script/dom/htmlimageelement.rs

@@ -66,6 +74,12 @@ enum State {
    CompletelyAvailable,
    Broken,
 }
+#[derive(PartialEq)]


Nit: newline before types please

wafflespeanut · 2017-09-24T09:29:46Z

components/script/dom/htmlimageelement.rs

+pub fn parse_a_sizes_attribute(input: DOMString, width: Option<u32>) -> Vec<Size> {
+    let mut sizes = Vec::<Size>::new();
+    let unparsed_sizes = input.split(',').collect::<Vec<_>>();
+    for unparsed_size in &unparsed_sizes {


Since unparsed_sizes is used only in the loop, let's iterate on input.split directly (which avoids an iteration and vector allocation).

Not sure what you want me to do here. Would you please elaborate little more. Thanks :)

I meant,

for unparsed_size in input.split(',') {

This way, we perform lazy iteration.

wafflespeanut · 2017-09-24T09:29:46Z

components/script/dom/htmlimageelement.rs

+                let mut media_query_parser = parser;
+                let media_query = media_query_parser.try(|i| MediaQuery::parse(&context, i));
+                if let Ok(query) = media_query {
+                        let length = Length::parse_non_negative(&context, &mut media_query_parser);


Nit: Improper indentation (should be 4 spaces instead of 8).

wafflespeanut · 2017-09-24T09:29:46Z

tests/unit/script/htmlimageelement.rs

+    let size = Size { query: None, length: length };
+    a.push(size);
+    assert_eq!(parse_a_sizes_attribute(DOMString::new(), None), a);
+    println!("{:?}", parse_a_sizes_attribute(DOMString::new(), None));


Nit: Please remove println! usage.

Note for the future: println! is not useful in the testsuite because the stdout is captured by cargo. Instead, we should panic! so that cargo prints the captured stdout (which also has the panic message).

wafflespeanut · 2017-09-24T09:29:46Z

tests/unit/script/htmlimageelement.rs

+    let size = Size { query: None, length: length };
+    a.push(size);
+    assert_eq!(parse_a_sizes_attribute(DOMString::new(), Some(2)), a);
+    println!("{:?}", parse_a_sizes_attribute(DOMString::new(), Some(2)));


Nit: Ditto on println!

wafflespeanut · 2017-09-24T09:29:46Z

tests/unit/script/htmlimageelement.rs

+    let mut a = vec![];
+    let length = test_length_for_default_provided(2f32);
+    let size = Size { query: None, length: length };
+    a.push(size);


Nit: If it's just a single size, then let's have let a = vec![size]; (here and everywhere below)


        Parse sizes attribute- Updates


        Parse Sizes attribute


        Parse Sizes attribute

jdm · 2017-10-02T18:33:16Z

I went ahead and squashed these commits and opened a new PR in #18714.


        Auto merge of #18714 - jdm:sizes, r=jdm

Parse sizes attribute values Squashed version of #17808. --- - [x] `./mach build -d` does not report any errors - [x] `./mach test-tidy` does not report any errors - [x] These changes fix (partially) #11416 - [x] There are tests for these changes  --- This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/servo/18714)

highfive assigned asajeffrey Jul 20, 2017

highfive added the S-awaiting-review label Jul 20, 2017

jdm added the S-needs-tests label Jul 20, 2017

highfive added the S-needs-rebase label Jul 21, 2017

jdm removed the S-awaiting-review label Jul 24, 2017

jdm assigned jdm and unassigned asajeffrey Jul 24, 2017

Rakhisharma force-pushed the Rakhisharma:sizes branch from 5fc23a3 to bc95d0a Jul 27, 2017

jdm requested changes Jul 27, 2017

View changes

jdm added S-needs-code-changes and removed S-needs-rebase labels Jul 27, 2017

creativcoder reviewed Jul 30, 2017

View changes

highfive added S-needs-rebase S-awaiting-review and removed S-needs-code-changes labels Aug 23, 2017

Rakhisharma force-pushed the Rakhisharma:sizes branch from ca37ece to b37df95 Aug 29, 2017

jdm requested changes Aug 31, 2017

View changes

jdm added S-needs-code-changes and removed S-awaiting-review S-needs-tests labels Aug 31, 2017

highfive added S-awaiting-review and removed S-needs-code-changes labels Sep 6, 2017

jdm removed the S-awaiting-review label Sep 7, 2017

highfive added the S-awaiting-review label Sep 13, 2017

Parse Sizes Attribute

Loading status checks…

083bb37

Rakhisharma force-pushed the Rakhisharma:sizes branch from ee28de0 to 083bb37 Sep 21, 2017

Parse Sizes Attribute

Loading status checks…

f4321b5

wafflespeanut reviewed Sep 24, 2017

View changes

jdm added S-needs-code-changes and removed S-awaiting-review S-needs-rebase labels Sep 25, 2017

Parse sizes attribute- Updates

Loading status checks…

ce14eab

highfive added S-awaiting-review and removed S-needs-code-changes labels Sep 25, 2017

Rakhisharma added 2 commits Sep 25, 2017

Parse Sizes attribute

Loading status checks…

05185e2

Parse Sizes attribute

Loading status checks…

2f31ff0

jdm mentioned this pull request Oct 2, 2017

Parse sizes attribute values #18714

Merged

4 of 4 tasks complete

jdm closed this Oct 2, 2017

servo / servo

Join GitHub today

WIP #17808

WIP #17808

Conversation

Rakhisharma commented Jul 20, 2017 • edited

Parse sizes attribute implementation.

highfive commented Jul 20, 2017

highfive commented Jul 20, 2017

highfive commented Jul 20, 2017

bors-servo commented Jul 21, 2017

jdm left a comment

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

bors-servo commented Aug 23, 2017

jdm left a comment

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

jdm Aug 31, 2017 • edited by KiChjang

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

KiChjang commented Sep 1, 2017

creativcoder commented Sep 1, 2017

wafflespeanut left a comment

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

jdm commented Oct 2, 2017

Rakhisharma commented Jul 20, 2017 •

edited

jdm Aug 31, 2017 •

edited by KiChjang