diff options
Diffstat (limited to 'components/script/parse')
| -rw-r--r-- | components/script/parse/html.rs | 109 | ||||
| -rw-r--r-- | components/script/parse/mod.rs | 4 |
2 files changed, 23 insertions, 90 deletions
diff --git a/components/script/parse/html.rs b/components/script/parse/html.rs index 80885acf300..a9e125f7da9 100644 --- a/components/script/parse/html.rs +++ b/components/script/parse/html.rs @@ -4,7 +4,7 @@ #![allow(unsafe_code, unrooted_must_root)] -use document_loader::{DocumentLoader, LoadType}; +use document_loader::DocumentLoader; use dom::attr::AttrHelpers; use dom::bindings::codegen::Bindings::DocumentBinding::DocumentMethods; use dom::bindings::codegen::Bindings::NodeBinding::NodeMethods; @@ -32,13 +32,10 @@ use dom::servohtmlparser::{ServoHTMLParser, FragmentContext}; use dom::text::Text; use parse::Parser; -use encoding::all::UTF_8; -use encoding::types::{Encoding, DecoderTrap}; +use encoding::types::Encoding; -use net_traits::{ProgressMsg, LoadResponse}; +use msg::constellation_msg::PipelineId; use util::str::DOMString; -use util::task_state; -use util::task_state::IN_HTML_PARSER; use std::borrow::Cow; use std::io::{self, Write}; use url::Url; @@ -49,14 +46,6 @@ use html5ever::serialize::TraversalScope::{IncludeNode, ChildrenOnly}; use html5ever::tree_builder::{TreeSink, QuirksMode, NodeOrText, AppendNode, AppendText, NextParserState}; use string_cache::QualName; -use hyper::header::ContentType; -use hyper::mime::{Mime, TopLevel, SubLevel}; - -pub enum HTMLInput { - InputString(String), - InputUrl(LoadResponse), -} - trait SinkHelpers { fn get_or_create(&self, child: NodeOrText<JS<Node>>) -> Temporary<Node>; } @@ -183,7 +172,9 @@ impl<'a> TreeSink for servohtmlparser::Sink { fn complete_script(&mut self, node: JS<Node>) -> NextParserState { let node: Root<Node> = node.root(); let script: Option<JSRef<HTMLScriptElement>> = HTMLScriptElementCast::to_ref(node.r()); - script.map(|script| script.prepare()); + if let Some(script) = script { + return script.prepare(); + } NextParserState::Continue } @@ -272,80 +263,22 @@ impl<'a> Serializable for JSRef<'a, Node> { } } +pub enum ParseContext<'a> { + Fragment(FragmentContext<'a>), + Owner(Option<PipelineId>), +} + pub fn parse_html(document: JSRef<Document>, - input: HTMLInput, + input: String, url: &Url, - fragment_context: Option<FragmentContext>) { - let parser = match fragment_context { - None => ServoHTMLParser::new(Some(url.clone()), document).root(), - Some(fc) => ServoHTMLParser::new_for_fragment(Some(url.clone()), document, fc).root(), - }; - let parser: JSRef<ServoHTMLParser> = parser.r(); - - let nested_parse = task_state::get().contains(task_state::IN_HTML_PARSER); - if !nested_parse { - task_state::enter(IN_HTML_PARSER); - } - - fn parse_progress(document: JSRef<Document>, parser: JSRef<ServoHTMLParser>, url: &Url, load_response: &LoadResponse) { - for msg in load_response.progress_port.iter() { - match msg { - ProgressMsg::Payload(data) => { - // FIXME: use Vec<u8> (html5ever #34) - let data = UTF_8.decode(&data, DecoderTrap::Replace).unwrap(); - parser.parse_chunk(data); - } - ProgressMsg::Done(Err(err)) => { - debug!("Failed to load page URL {}, error: {}", url.serialize(), err); - // TODO(Savago): we should send a notification to callers #5463. - break; - } - ProgressMsg::Done(Ok(())) => { - document.finish_load(LoadType::PageSource(url.clone())); - break; - } - } - } - }; - - match input { - HTMLInput::InputString(s) => { - parser.parse_chunk(s); - } - HTMLInput::InputUrl(load_response) => { - match load_response.metadata.content_type { - Some(ContentType(Mime(TopLevel::Image, _, _))) => { - let page = format!("<html><body><img src='{}' /></body></html>", url.serialize()); - parser.parse_chunk(page); - document.finish_load(LoadType::PageSource(url.clone())); - }, - Some(ContentType(Mime(TopLevel::Text, SubLevel::Plain, _))) => { - // FIXME: When servo/html5ever#109 is fixed remove <plaintext> usage and - // replace with fix from that issue. - - // text/plain documents require setting the tokenizer into PLAINTEXT mode. - // This is done by using a <plaintext> element as the html5ever tokenizer - // provides no other way to change to that state. - // Spec for text/plain handling is: - // https://html.spec.whatwg.org/multipage/#read-text - let page = format!("<pre>\u{000A}<plaintext>"); - parser.parse_chunk(page); - parse_progress(document, parser, url, &load_response); - }, - _ => { - parse_progress(document, parser, url, &load_response); - } - } - } - } - - parser.finish(); - - if !nested_parse { - task_state::exit(IN_HTML_PARSER); - } - - debug!("finished parsing"); + context: ParseContext) { + let parser = match context { + ParseContext::Owner(owner) => + ServoHTMLParser::new(Some(url.clone()), document, owner), + ParseContext::Fragment(fc) => + ServoHTMLParser::new_for_fragment(Some(url.clone()), document, fc), + }.root(); + parser.r().parse_chunk(input); } // https://html.spec.whatwg.org/multipage/#parsing-html-fragments @@ -376,7 +309,7 @@ pub fn parse_html_fragment(context_node: JSRef<Node>, context_elem: context_node, form_elem: form.r(), }; - parse_html(document.r(), HTMLInput::InputString(input), &url, Some(fragment_context)); + parse_html(document.r(), input, &url, ParseContext::Fragment(fragment_context)); // Step 14. let root_element = document.r().GetDocumentElement().expect("no document element").root(); diff --git a/components/script/parse/mod.rs b/components/script/parse/mod.rs index 6111f196053..fa93dbc157f 100644 --- a/components/script/parse/mod.rs +++ b/components/script/parse/mod.rs @@ -5,6 +5,6 @@ pub mod html; pub trait Parser { - fn parse_chunk(&self,input: String); - fn finish(&self); + fn parse_chunk(self, input: String); + fn finish(self); } |