Mirrors/servo
1
0
Fork 0
mirror of https://github.com/servo/servo.git synced 2025-06-25 01:24:37 +01:00
Code Issues Projects Releases Packages Wiki Activity

Implement encoding determination for external scripts.

Browse source
This commit is contained in:
askalski 2016-01-09 21:40:45 +01:00 • committed by Ms2ger
parent b1cd28e9c6
commit 89b8499df8
7 changed files with 175 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

17
components/script/dom/htmlscriptelement.rs
View file

@ -25,7 +25,6 @@ use dom::node::{ChildrenMutation, CloneChildrenFlag, Node};
use dom::node::{document_from_node, window_from_node};
use dom::virtualmethods::VirtualMethods;
use dom::window::ScriptHelpers;
use encoding::all::UTF_8;
use encoding::label::encoding_from_whatwg_label;
use encoding::types::{DecoderTrap, Encoding, EncodingRef};
use html5ever::tree_builder::NextParserState;
@ -71,7 +70,7 @@ pub struct HTMLScriptElement {
#[ignore_heap_size_of = "Defined in rust-encoding"]
/// https://html.spec.whatwg.org/multipage/#concept-script-encoding
block_character_encoding: DOMRefCell<EncodingRef>,
block_character_encoding: DOMRefCell<Option<EncodingRef>>,
}
impl HTMLScriptElement {
@ -86,7 +85,7 @@ impl HTMLScriptElement {
ready_to_be_parser_executed: Cell::new(false),
parser_document: JS::from_ref(document),
load: DOMRefCell::new(None),
block_character_encoding: DOMRefCell::new(UTF_8 as EncodingRef),
block_character_encoding: DOMRefCell::new(None),
}
}
@ -248,7 +247,7 @@ impl HTMLScriptElement {
// Step 13.
if let Some(ref charset) = element.get_attribute(&ns!(), &atom!("charset")) {
if let Some(encodingRef) = encoding_from_whatwg_label(&charset.Value()) {
*self.block_character_encoding.borrow_mut() = encodingRef;
*self.block_character_encoding.borrow_mut() = Some(encodingRef);
}
}
@ -391,8 +390,14 @@ impl HTMLScriptElement {
// Step 2.b.1.a.
ScriptOrigin::External(Ok((metadata, bytes))) => {
// TODO(#9185): implement encoding determination.
(DOMString::from(UTF_8.decode(&*bytes, DecoderTrap::Replace).unwrap()),
debug!("loading external script, url = {}", metadata.final_url);
let encoding = metadata.charset
.and_then(|encoding| encoding_from_whatwg_label(&encoding))
.or_else(|| *self.block_character_encoding.borrow())
.unwrap_or_else(|| self.parser_document.encoding());
(DOMString::from(encoding.decode(&*bytes, DecoderTrap::Replace).unwrap()),
true,
metadata.final_url)
},

8
tests/wpt/metadata/MANIFEST.json
View file

@ -19241,6 +19241,14 @@
"path": "html/semantics/scripting-1/the-script-element/script-before-after-events.html",
"url": "/html/semantics/scripting-1/the-script-element/script-before-after-events.html"
},
{
"path": "html/semantics/scripting-1/the-script-element/script-charset-01.html",
"url": "/html/semantics/scripting-1/the-script-element/script-charset-01.html"
},
{
"path": "html/semantics/scripting-1/the-script-element/script-charset-02.html",
"url": "/html/semantics/scripting-1/the-script-element/script-charset-02.html"
},
{
"path": "html/semantics/scripting-1/the-script-element/script-for-event-xhtml.xhtml",
"url": "/html/semantics/scripting-1/the-script-element/script-for-event-xhtml.xhtml"

5
tests/wpt/web-platform-tests/html/semantics/scripting-1/the-script-element/external-script-utf8.js Normal file
View file

@ -0,0 +1,5 @@
(function() {
window.getSomeString = function() {
return "śćążź"; //<- these are five Polish letters, similar to scazz. It can be read correctly only with windows 1250 encoding.
};
})();

5
tests/wpt/web-platform-tests/html/semantics/scripting-1/the-script-element/external-script-windows1250.js Normal file
View file

@ -0,0 +1,5 @@
(function() {
window.getSomeString = function() {
return "œæ¹¿Ÿ"; //<- these are five Polish letters, similar to scazz. It can be read correctly only with windows 1250 encoding.
};
})();

89
tests/wpt/web-platform-tests/html/semantics/scripting-1/the-script-element/script-charset-01.html Normal file
View file

@ -0,0 +1,89 @@
<!DOCTYPE html>
<head>
<meta charset="utf-8">
<title>Script @type: unknown parameters</title>
<link rel="author" title="askalski" href="github.com/askalski">
<link rel="help" href="https://html.spec.whatwg.org/multipage/#scriptingLanguages">
<script src="/resources/testharness.js"></script>
<script src="/resources/testharnessreport.js"></script>
<div id="log"></div>
<!-- "Step1" tests -->
<!-- charset is set incorrectly via Content Type "text/javascript;charset=utf-8" in response
which has priority before a correct setting in "charset" attribute of script tag.
-->
<script type="text/javascript"
src="serve-with-content-type.py?fn=external-script-windows1250.js&ct=text/javascript%3Bcharset=utf-8" charset="windows-1250">
</script>
<script>
test(function() {
//these strings should not match, since the file charset is set incorrectly
assert_not_equals(window.getSomeString(), "śćążź");
});
</script>
<!-- charset is set correctly via Content Type "text/javascript;charset=utf-8" in response
which has priority before a incorrect setting in "charset" attribute of script tag.
-->
<script type="text/javascript"
src="serve-with-content-type.py?fn=external-script-windows1250.js&ct=text/javascript%3Bcharset=windows-1250" charset="utf-8">
</script>
<script>
//the charset is set correctly via Content Type "text/javascript;charset=windows-1250" in respones
test(function() {
assert_equals(window.getSomeString(), "śćążź");
});
</script>
<!-- end of step1 tests, now step2 tests -->
<!-- in this case, the response's Content Type does not bring charset information.
Second step takes block character encoding if available.-->
<script type="text/javascript"
src="serve-with-content-type.py?fn=external-script-windows1250.js&ct=text/javascript" charset="utf-8">
</script>
<script>
test(function() {
//these strings should not match, since the file charset is set incorrectly in "charset" tag of <script> above
assert_not_equals(window.getSomeString(), "śćążź");
});
</script>
<!-- charset is set correctly via Content Type "text/javascript;charset=utf-8" in response
which has priority before a incorrect setting in "charset" attribute of script tag.
-->
<script type="text/javascript"
src="serve-with-content-type.py?fn=external-script-windows1250.js&ct=text/javascript" charset="windows-1250">
</script>
<script>
//the charset is set correctly via content attribute in <script> above
test(function() {
assert_equals(window.getSomeString(), "śćążź");
});
</script>
<!-- end of step2 tests, now step3 tests -->
<!-- in this case, neither response's Content Type nor charset attribute bring correct charset information.
Third step takes this document's character encoding (declared correctly as UTF-8).-->
<script type="text/javascript"
src="serve-with-content-type.py?fn=external-script-windows1250.js&ct=text/javascript">
</script>
<script>
test(function() {
//these strings should not match, since the tested file is in windows-1250, and document is utf-8
assert_not_equals(window.getSomeString(), "śćążź");
});
</script>
<script type="text/javascript"
src="serve-with-content-type.py?fn=external-script-utf8.js&ct=text/javascript">
</script>
<script>
//these strings should match, both document and tested file are utf-8
test(function() {
assert_equals(window.getSomeString(), "śćążź");
});
</script>
<!-- the last portion of tests (step4) are in file script-charset-02.html
</head>

40
tests/wpt/web-platform-tests/html/semantics/scripting-1/the-script-element/script-charset-02.html Normal file
View file

@ -0,0 +1,40 @@
<!DOCTYPE html>
<head>
<!-- TODO:
askalski: while this test pass, it does not test anything now.
It should test, whether with no document.charset set in any way, the
external scripts will get decoded using utf-8 as fallback character encoding.
It seems like utf-8 is also a fallback encoding to html (my guess), so
the part of the code I was attempting to test is never reached.
-->
<title>Script @type: unknown parameters</title>
<link rel="author" title="askalski" href="github.com/askalski">
<link rel="help" href="https://html.spec.whatwg.org/multipage/#scriptingLanguages">
<script src="/resources/testharness.js"></script>
<script src="/resources/testharnessreport.js"></script>
<div id="log"></div>
<!-- test of step4, which is taking utf-8 as fallback -->
<!-- in this case, neither response's Content Type nor charset attribute bring correct charset information.
Furthermore, document's encoding is not set.-->
<script type="text/javascript"
src="serve-with-content-type.py?fn=external-script-windows1250.js&ct=text/javascript">
</script>
<script>
test(function() {
//these strings should not match, since the tested file is in windows-1250, and fallback is defined as utf-8
assert_not_equals(window.getSomeString().length, 5);
});
</script>
<script type="text/javascript"
src="serve-with-content-type.py?fn=external-script-utf8.js&ct=text/javascript">
</script>
<script>
//these strings should match, since fallback utf-8 is the correct setting.
test(function() {
assert_equals(window.getSomeString().length, 5);
});
</script>
</head>

15
tests/wpt/web-platform-tests/html/semantics/scripting-1/the-script-element/serve-with-content-type.py Normal file
View file

@ -0,0 +1,15 @@
import os
def main(request, response):
directory = os.path.dirname(__file__)
try:
file_name = request.GET.first("fn")
content_type = request.GET.first("ct")
with open(os.path.join(directory, file_name), "rb") as fh:
content = fh.read()
response.headers.set("Content-Type", content_type)
response.content = content
except:
response.set_error(400, "Not enough parameters or file not found")