Added a TopLevelBrowsingContextId type. #16916

asajeffrey · 2017-05-17T20:59:47Z

Added a TopLevelBrowsingContextId type, which is a subtype of BrowsingContextId that only refers to top-level browsing contexts.

./mach build -d does not report any errors
./mach test-tidy does not report any errors
These changes do not require tests because refactoring

This change is

highfive · 2017-05-17T20:59:51Z

Heads up! This PR modifies the following files:

@fitzgen: components/script/dom/htmliframeelement.rs, components/script/dom/document.rs, components/script/dom/windowproxy.rs, components/script/dom/node.rs, components/script_traits/script_msg.rs and 6 more
@KiChjang: components/script/dom/htmliframeelement.rs, components/script/dom/document.rs, components/script/dom/windowproxy.rs, components/script/dom/node.rs, components/script_traits/script_msg.rs and 6 more

highfive · 2017-05-17T20:59:53Z

Warning

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

asajeffrey · 2017-05-17T21:00:54Z

This is the next step in the effort to allow more than one root browsing context. cc @cbrewster @paulrouget

asajeffrey · 2017-05-17T21:01:31Z

r? @cbrewster

cbrewster

Looking good! +1 for more type safety

cbrewster · 2017-05-18T17:34:47Z

components/constellation/constellation.rs

@@ -700,8 +685,8 @@ impl<Message, LTF, STF> Constellation<Message, LTF, STF>
    /// Get an iterator for browsing contexts. Specify self.root_browsing context_id to
    /// iterate the entire tree, or a specific browsing context id to iterate only that sub-tree.


Update to state that this can only iterate a sub-tree

cbrewster · 2017-05-18T17:34:47Z

components/constellation/constellation.rs

@@ -710,11 +695,20 @@ impl<Message, LTF, STF> Constellation<Message, LTF, STF>
        }
    }

+    /// Get an iterator for browsing contexts. Specify self.root_browsing context_id to
+    /// iterate the entire tree, or a specific browsing context id to iterate only that sub-tree.


This can only iterate an entire tree.

cbrewster · 2017-05-18T17:34:47Z

components/constellation/constellation.rs

+    {
+        self.fully_active_descendant_browsing_contexts_iter(BrowsingContextId::from(top_level_browsing_context_id))
+    }
+
    /// Get an iterator for browsing contexts. Specify self.root_browsing_context_id to
    /// iterate the entire tree, or a specific browsing context id to iterate only that sub-tree.


Same as above

cbrewster · 2017-05-18T17:34:47Z

components/constellation/constellation.rs

@@ -723,28 +717,37 @@ impl<Message, LTF, STF> Constellation<Message, LTF, STF>
        }
    }

+    /// Get an iterator for browsing contexts. Specify self.root_browsing_context_id to
+    /// iterate the entire tree, or a specific browsing context id to iterate only that sub-tree.


Same as above

cbrewster · 2017-05-18T17:34:47Z

components/constellation/constellation.rs

@@ -1862,17 +1892,20 @@ impl<Message, LTF, STF> Constellation<Message, LTF, STF>
        // and pass the event to that script thread.
        // If the pipeline lookup fails, it is because we have torn down the pipeline,
        // so it is reasonable to silently ignore the event.
-        let browsing_context_id = self.pipelines.get(&pipeline_id).map(|pipeline| pipeline.browsing_context_id);
+        let top_level_browsing_context_id = self.pipelines.get(&pipeline_id)


Nice catch!

Yes, this is one of those odd cases that the extra typechecking helped with. Yay types!

cbrewster · 2017-05-18T17:34:47Z

components/layout_thread/lib.rs

@@ -244,7 +244,6 @@ impl LayoutThreadFactory for LayoutThread {

    /// Spawns a new layout thread.
    fn create(id: PipelineId,
-              top_level_browsing_context_id: Option<BrowsingContextId>,


I remembered why we wanted this, which is for crash reporting, so that if a layout panic happens we know which top-level browsing context to blame.

cbrewster · 2017-05-18T17:34:47Z

components/constellation/pipeline.rs

@@ -104,7 +104,7 @@ pub struct InitialPipelineState {
    pub browsing_context_id: BrowsingContextId,

    /// The ID of the top-level browsing context that contains this Pipeline.
-    pub top_level_browsing_context_id: BrowsingContextId,
+    pub top_level_browsing_context_id: TopLevelBrowsingContextId,


Is this still needed since the layout thread no longer needs it?

asajeffrey

Changes made.

asajeffrey · 2017-05-19T02:04:45Z

components/constellation/constellation.rs

@@ -710,11 +695,20 @@ impl<Message, LTF, STF> Constellation<Message, LTF, STF>
        }
    }

+    /// Get an iterator for browsing contexts. Specify self.root_browsing context_id to
+    /// iterate the entire tree, or a specific browsing context id to iterate only that sub-tree.


asajeffrey · 2017-05-19T02:04:45Z

components/constellation/constellation.rs

+    {
+        self.fully_active_descendant_browsing_contexts_iter(BrowsingContextId::from(top_level_browsing_context_id))
+    }
+
    /// Get an iterator for browsing contexts. Specify self.root_browsing_context_id to
    /// iterate the entire tree, or a specific browsing context id to iterate only that sub-tree.


asajeffrey · 2017-05-19T02:04:45Z

components/constellation/constellation.rs

@@ -723,28 +717,37 @@ impl<Message, LTF, STF> Constellation<Message, LTF, STF>
        }
    }

+    /// Get an iterator for browsing contexts. Specify self.root_browsing_context_id to
+    /// iterate the entire tree, or a specific browsing context id to iterate only that sub-tree.


asajeffrey · 2017-05-19T02:04:45Z

components/constellation/constellation.rs

@@ -1862,17 +1892,20 @@ impl<Message, LTF, STF> Constellation<Message, LTF, STF>
        // and pass the event to that script thread.
        // If the pipeline lookup fails, it is because we have torn down the pipeline,
        // so it is reasonable to silently ignore the event.
-        let browsing_context_id = self.pipelines.get(&pipeline_id).map(|pipeline| pipeline.browsing_context_id);
+        let top_level_browsing_context_id = self.pipelines.get(&pipeline_id)


Yes, this is one of those odd cases that the extra typechecking helped with. Yay types!

asajeffrey · 2017-05-19T02:04:45Z

components/layout_thread/lib.rs

@@ -244,7 +244,6 @@ impl LayoutThreadFactory for LayoutThread {

    /// Spawns a new layout thread.
    fn create(id: PipelineId,
-              top_level_browsing_context_id: Option<BrowsingContextId>,


I remembered why we wanted this, which is for crash reporting, so that if a layout panic happens we know which top-level browsing context to blame.

cbrewster · 2017-05-19T17:17:03Z

components/constellation/constellation.rs

@@ -1783,7 +1803,8 @@ impl<Message, LTF, STF> Constellation<Message, LTF, STF>
        // Send to the explicitly focused pipeline (if it exists), or the root
        // browsing context's current pipeline. If neither exist, fall back to sending to
        // the compositor below.
-        let root_pipeline_id = self.browsing_contexts.get(&self.root_browsing_context_id)
+        let root_browsing_context_id = BrowsingContextId::from(self.root_browsing_context_id);
+        let root_pipeline_id = self.browsing_contexts.get(&root_browsing_context_id)


I wonder if we could replace these occurrences with:

self.browsing_contexts.get(&self.root_browsing_context_id.into())

since From<T> for U implies Into<U> for T

I prefer T::from(x) over x.into() just because I find it easier to read. x.into() is rather mysterious, it converts from one unnamed type to another unnamed type. YMMV.

cbrewster · 2017-05-19T17:38:12Z

You can decide whether to switch to .into() or keep using ::from. Other than that this looks good, so you can squash and r=me

asajeffrey · 2017-05-19T17:47:24Z

Squashed. @bors-servo r=cbrewster

bors-servo · 2017-05-19T17:47:24Z

📌 Commit 273c474 has been approved by cbrewster

bors-servo · 2017-05-20T02:50:17Z

⌛ Testing commit 273c474 with merge 09d18dd...


        Auto merge of #16916 - asajeffrey:constellation-top-level-browsing-co…

…ntexts, r=cbrewster Added a TopLevelBrowsingContextId type.  Added a `TopLevelBrowsingContextId` type, which is a subtype of `BrowsingContextId` that only refers to top-level browsing contexts. ---  - [X] `./mach build -d` does not report any errors - [X] `./mach test-tidy` does not report any errors - [X] These changes do not require tests because refactoring    --- This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/servo/16916)

bors-servo · 2017-05-20T03:29:41Z

💔 Test failed - mac-rel-wpt1

cbrewster · 2017-05-20T17:03:19Z

  ▶ Unexpected subtest result in /html/semantics/embedded-content/the-iframe-element/iframe-synchronously-discard.html:
  │ FAIL [expected PASS] IFrame discards are processed synchronously
  │   → assert_equals: grandchild should be discarded expected null but got object "[object Window]"
  │ 
  │ child.onload</grandchild.onload<@http://web-platform.test:8000/html/semantics/embedded-content/the-iframe-element/iframe-synchronously-discard.html:24:9
  │ Test.prototype.step@http://web-platform.test:8000/resources/testharness.js:1406:20
  └ Test.prototype.step_func/<@http://web-platform.test:8000/resources/testharness.js:1430:20

asajeffrey · 2017-05-22T13:51:13Z

Well the good news is that I can replicate this locally. Not entirely sure how I broke synchronous iframe discarding, off to do some digging...

asajeffrey · 2017-05-22T14:26:22Z

Figured it out...

$ git diff
diff --git a/components/script/dom/htmliframeelement.rs b/components/script/dom/htmliframeelement.rs
index 1c00f3e..0075521 100644
--- a/components/script/dom/htmliframeelement.rs
+++ b/components/script/dom/htmliframeelement.rs
@@ -806,9 +806,11 @@ impl VirtualMethods for HTMLIFrameElement {
         // when the `PipelineExit` message arrives.
         for exited_pipeline_id in exited_pipeline_ids {
             if let Some(exited_document) = ScriptThread::find_document(exited_pipeline_id) {
+                debug!("Discarding browsing context for pipeline {}", exited_pipeline_id);
                 exited_document.window().window_proxy().discard_browsing_context();
                 for exited_iframe in exited_document.iter_iframes() {
-                    exited_iframe.pipeline_id.set(None);
+                    debug!("Discarding nested browsing context");
+                    exited_iframe.destroy_nested_browsing_context();
                 }
             }
         }

When we were synchronously discarding a browsing context, we were calling exited_iframe.pipeline_id.set(None);, which used to do the job, but now should be exited_iframe.destroy_nested_browsing_context();.


        Added a TopLevelBrowsingContextId type.

asajeffrey · 2017-05-22T14:27:37Z

@bors-servo r=cbrewster

bors-servo · 2017-05-22T14:27:38Z

📌 Commit 4257736 has been approved by cbrewster

bors-servo · 2017-05-22T20:13:43Z

⌛ Testing commit 4257736 with merge a0d6d68...


        Auto merge of #16916 - asajeffrey:constellation-top-level-browsing-co…

…ntexts, r=cbrewster Added a TopLevelBrowsingContextId type.  Added a `TopLevelBrowsingContextId` type, which is a subtype of `BrowsingContextId` that only refers to top-level browsing contexts. ---  - [X] `./mach build -d` does not report any errors - [X] `./mach test-tidy` does not report any errors - [X] These changes do not require tests because refactoring    --- This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/servo/16916)

bors-servo · 2017-05-22T21:46:26Z

☀️ Test successful - android, arm32, arm64, linux-dev, linux-rel-css, linux-rel-wpt, mac-dev-unit, mac-rel-css, mac-rel-wpt1, mac-rel-wpt2, windows-msvc-dev
Approved by: cbrewster
Pushing a0d6d68 to master...

highfive assigned pcwalton May 17, 2017

highfive added the S-awaiting-review label May 17, 2017

asajeffrey added the A-constellation label May 17, 2017

highfive assigned cbrewster and unassigned pcwalton May 17, 2017

asajeffrey mentioned this pull request May 18, 2017

[multi window] Constellation should be able to handle multiple root frame #13994

Closed

cbrewster requested changes May 18, 2017

View changes

cbrewster added S-needs-code-changes and removed S-awaiting-review labels May 18, 2017

highfive added S-awaiting-review and removed S-needs-code-changes labels May 19, 2017

asajeffrey reviewed May 19, 2017

View changes

cbrewster reviewed May 19, 2017

View changes

asajeffrey force-pushed the asajeffrey:constellation-top-level-browsing-contexts branch from 53174cb to 273c474 May 19, 2017

highfive added S-awaiting-merge and removed S-awaiting-review labels May 19, 2017

highfive added S-tests-failed and removed S-awaiting-merge labels May 20, 2017

Added a TopLevelBrowsingContextId type.

Loading status checks…

4257736

asajeffrey force-pushed the asajeffrey:constellation-top-level-browsing-contexts branch from 273c474 to 4257736 May 22, 2017

highfive added S-awaiting-review and removed S-tests-failed labels May 22, 2017

highfive added S-awaiting-merge and removed S-awaiting-review labels May 22, 2017

asajeffrey mentioned this pull request May 22, 2017

Webdriver browsing contexts not pipelines #16997

Merged

3 of 3 tasks complete

bors-servo merged commit 4257736 into servo:master May 22, 2017
3 checks passed

3 checks passed

continuous-integration/appveyor/pr AppVeyor build succeeded
Details

continuous-integration/travis-ci/pr The Travis CI build passed
Details

homu Test successful
Details

highfive removed the S-awaiting-merge label May 22, 2017

bors-servo mentioned this pull request May 22, 2017

Replace initial about:blank on navigation #16848

Closed

4 of 5 tasks complete

		@@ -700,8 +685,8 @@ impl<Message, LTF, STF> Constellation<Message, LTF, STF>
		/// Get an iterator for browsing contexts. Specify self.root_browsing context_id to
		/// iterate the entire tree, or a specific browsing context id to iterate only that sub-tree.

servo / servo

Join GitHub today

Added a TopLevelBrowsingContextId type. #16916

Added a TopLevelBrowsingContextId type. #16916

Conversation

asajeffrey commented May 17, 2017 • edited by larsbergstrom

highfive commented May 17, 2017

highfive commented May 17, 2017

asajeffrey commented May 17, 2017

asajeffrey commented May 17, 2017

cbrewster 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.

asajeffrey 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.

cbrewster commented May 19, 2017

asajeffrey commented May 19, 2017

bors-servo commented May 19, 2017

bors-servo commented May 20, 2017

bors-servo commented May 20, 2017

cbrewster commented May 20, 2017

asajeffrey commented May 22, 2017

asajeffrey commented May 22, 2017

asajeffrey commented May 22, 2017

bors-servo commented May 22, 2017

bors-servo commented May 22, 2017

bors-servo commented May 22, 2017

asajeffrey commented May 17, 2017 •

edited by larsbergstrom