layout: Rewrite anonymous table code, simplify and fix table intrinsic width calculation, and improve safety of flexbox code. #13870

pcwalton · 2016-10-21T04:10:46Z

Closes #13782.

This change is

highfive · 2016-10-21T04:10:49Z

Heads up! This PR modifies the following files:

@bholley: components/style/selector_matching.rs, components/style/animation.rs, components/style/servo_selector_impl.rs, components/style/properties/properties.mako.rs, components/style/matching.rs
@KiChjang: components/script_layout_interface/wrapper_traits.rs
@fitzgen: components/script_layout_interface/wrapper_traits.rs
@emilio: components/style/selector_matching.rs, components/style/animation.rs, components/style/servo_selector_impl.rs, components/layout/multicol.rs, components/layout/list_item.rs, components/style/properties/properties.mako.rs, components/layout/table_caption.rs, components/layout/flex.rs, components/layout/table_cell.rs, components/layout/construct.rs, components/layout/flow.rs, components/layout/fragment.rs, components/layout/table_rowgroup.rs, components/layout/table_wrapper.rs, components/layout/block.rs, components/layout/table_row.rs, components/style/matching.rs, components/layout/flow_list.rs, components/layout/table.rs

highfive · 2016-10-21T04:10:50Z

Warning

This pull request modifies the contents of
tests/wpt/css-tests/, which are overwriten occasionally whenever the
directory is synced from upstream.

pcwalton · 2016-10-21T04:11:05Z

@bors-servo: try

bors-servo · 2016-10-21T04:11:08Z

⌛ Trying commit 392ab81 with merge f930608...


        Auto merge of #13870 - pcwalton:anonymous-table-rewrite, r=<try>

(DO NOT MERGE) layout: Rewrite table stuff. Closes #13782.

bors-servo · 2016-10-21T04:32:15Z

💔 Test failed - mac-dev-unit

bors-servo · 2016-10-21T07:36:05Z

☔ The latest upstream changes (presumably #13848) made this pull request unmergeable. Please resolve the merge conflicts.

Ms2ger · 2016-10-21T08:39:25Z

r? @mbrubeck

All I'd complain about is spec-related :)

emilio

Just a first pass with a few comments. The style parts look good to me. The layout parts make sense to me, though I'd have to do a lot of spec-reading to ensure it's correct.

emilio · 2016-10-21T10:07:20Z

tests/wpt/css-tests/css21_dev/html4/margin-right-applies-to-014.htm

@@ -19,7 +19,7 @@
            #table
            {
                border-right: 10px solid blue;
-                display: inline-table;
+                display: block;


This will need to be upstreamed, or reverted.

emilio · 2016-10-21T10:07:20Z

tests/wpt/css-tests/css21_dev/html4/background-attachment-applies-to-013.htm

I think all the newline changes inside tests/wpt should also be reverted.

emilio · 2016-10-21T10:07:20Z

components/style/servo_selector_impl.rs

+        fun(PseudoElement::ServoAnonymousTableWrapper);
+        fun(PseudoElement::ServoAnonymousTable);
+        fun(PseudoElement::ServoAnonymousTableRow);
+        fun(PseudoElement::ServoAnonymousTableCell);


Probably worth creating a file like gecko's pseudo_element_helper.rs to alleviate this and avoid the duplication (it didn't matter much when they were four pseudos, but...). Also we can specialize a few functions easily. In any case, not your business for this PR, I'll file an issue about it :)

emilio · 2016-10-21T10:07:20Z

components/style/servo_selector_impl.rs

@@ -13,6 +13,7 @@ use selectors::parser::{AttrSelector, ParserContext, SelectorImpl};
 use std::fmt;
 use string_cache::{Atom, Namespace};

+/// NB: If you add to this list, be sure to update `each_pseudo_element` too.


I'm pretty sure I wrote this exact comment when I introduced each_pseudo_element, where did it go? Sorry it bited you :/

emilio · 2016-10-21T10:07:20Z

components/style/selector_matching.rs

@@ -252,19 +253,29 @@ impl Stylist {

    /// Computes the style for a given "precomputed" pseudo-element, taking the
    /// universal rules and applying them.
+    ///
+    /// If `inherit_all` is true, then all properties are inherited from the parent; otherwise,
+    /// non-inherited properties are reset to their initial values.


A comment saying that this is used for anonymous flows would be helpful IMO.

emilio · 2016-10-21T10:07:20Z

components/layout/construct.rs

+        let table_style = node.style(self.style_context());
+        let wrapper_style = self.style_context()
+                                .stylist
+                                .precomputed_values_for_pseudo(&PseudoElement::ServoTableWrapper,


This pattern is becoming increasingly common, maybe add a helper for it in SharedStyleContext? Can be left as a followup, but please fill it if you decide to do so.

Added a helper to Stylist (as not all the places that call precomputed_values_for_pseudo have a handle to the style context).

mbrubeck · 2016-10-21T16:18:40Z

Some test results that weren't auto-linked above:

mac-rel-css

  ▶ PASS [expected FAIL] /mediaqueries-3_dev/html/min-width-tables-001.htm

linux-rel-css

Tests with unexpected results:
  ▶ PASS [expected FAIL] /css21_dev/html4/color-applies-to-014.htm

  ▶ FAIL [expected PASS] /css21_dev/html4/table-anonymous-objects-009.htm
  └   → /css21_dev/html4/table-anonymous-objects-009.htm 9a467359e61bf1cd03e915a871942253e111a547
/css21_dev/html4/reference/no_red_antialiasing_a_bc_d-ref.htm 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1
Testing 9a467359e61bf1cd03e915a871942253e111a547 == 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1

  ▶ FAIL [expected PASS] /css21_dev/html4/table-anonymous-objects-020.htm
  └   → /css21_dev/html4/table-anonymous-objects-020.htm 9a467359e61bf1cd03e915a871942253e111a547
/css21_dev/html4/reference/no_red_antialiasing_a_bc_d-ref.htm 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1
Testing 9a467359e61bf1cd03e915a871942253e111a547 == 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1

  ▶ FAIL [expected PASS] /css21_dev/html4/table-anonymous-objects-019.htm
  └   → /css21_dev/html4/table-anonymous-objects-019.htm 9a467359e61bf1cd03e915a871942253e111a547
/css21_dev/html4/reference/no_red_antialiasing_a_bc_d-ref.htm 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1
Testing 9a467359e61bf1cd03e915a871942253e111a547 == 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1

  ▶ FAIL [expected PASS] /css21_dev/html4/table-anonymous-objects-010.htm
  └   → /css21_dev/html4/table-anonymous-objects-010.htm 9a467359e61bf1cd03e915a871942253e111a547
/css21_dev/html4/reference/no_red_antialiasing_a_bc_d-ref.htm 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1
Testing 9a467359e61bf1cd03e915a871942253e111a547 == 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1

  ▶ FAIL [expected PASS] /css21_dev/html4/table-anonymous-objects-018.htm
  └   → /css21_dev/html4/table-anonymous-objects-018.htm 9a467359e61bf1cd03e915a871942253e111a547
/css21_dev/html4/reference/no_red_antialiasing_a_bc_d-ref.htm 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1
Testing 9a467359e61bf1cd03e915a871942253e111a547 == 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1

  ▶ FAIL [expected PASS] /css21_dev/html4/table-anonymous-objects-012.htm
  └   → /css21_dev/html4/table-anonymous-objects-012.htm 9a467359e61bf1cd03e915a871942253e111a547
/css21_dev/html4/reference/no_red_antialiasing_a_bc_d-ref.htm 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1
Testing 9a467359e61bf1cd03e915a871942253e111a547 == 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1

  ▶ FAIL [expected PASS] /css21_dev/html4/table-anonymous-objects-011.htm
  └   → /css21_dev/html4/table-anonymous-objects-011.htm 9a467359e61bf1cd03e915a871942253e111a547
/css21_dev/html4/reference/no_red_antialiasing_a_bc_d-ref.htm 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1
Testing 9a467359e61bf1cd03e915a871942253e111a547 == 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1

  ▶ FAIL [expected PASS] /css21_dev/html4/table-anonymous-objects-017.htm
  └   → /css21_dev/html4/table-anonymous-objects-017.htm 9a467359e61bf1cd03e915a871942253e111a547
/css21_dev/html4/reference/no_red_antialiasing_a_bc_d-ref.htm 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1
Testing 9a467359e61bf1cd03e915a871942253e111a547 == 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1

  ▶ FAIL [expected PASS] /css21_dev/html4/table-anonymous-objects-197.htm
  └   → /css21_dev/html4/table-anonymous-objects-197.htm 9a467359e61bf1cd03e915a871942253e111a547
/css21_dev/html4/reference/no_red_antialiasing_a_bc_d-ref.htm 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1
Testing 9a467359e61bf1cd03e915a871942253e111a547 == 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1

  ▶ FAIL [expected PASS] /css21_dev/html4/table-anonymous-objects-198.htm
  └   → /css21_dev/html4/table-anonymous-objects-198.htm 9a467359e61bf1cd03e915a871942253e111a547
/css21_dev/html4/reference/no_red_antialiasing_a_bc_d-ref.htm 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1
Testing 9a467359e61bf1cd03e915a871942253e111a547 == 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1

  ▶ FAIL [expected PASS] /css21_dev/html4/table-anonymous-objects-199.htm
  └   → /css21_dev/html4/table-anonymous-objects-199.htm 9a467359e61bf1cd03e915a871942253e111a547
/css21_dev/html4/reference/no_red_antialiasing_a_bc_d-ref.htm 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1
Testing 9a467359e61bf1cd03e915a871942253e111a547 == 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1

  ▶ FAIL [expected PASS] /css21_dev/html4/table-anonymous-objects-200.htm
  └   → /css21_dev/html4/table-anonymous-objects-200.htm 9a467359e61bf1cd03e915a871942253e111a547
/css21_dev/html4/reference/no_red_antialiasing_a_bc_d-ref.htm 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1
Testing 9a467359e61bf1cd03e915a871942253e111a547 == 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1

  ▶ FAIL [expected PASS] /css21_dev/html4/table-anonymous-objects-203.htm
  └   → /css21_dev/html4/table-anonymous-objects-203.htm 9a467359e61bf1cd03e915a871942253e111a547
/css21_dev/html4/reference/no_red_antialiasing_a_bc_d-ref.htm 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1
Testing 9a467359e61bf1cd03e915a871942253e111a547 == 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1

  ▶ FAIL [expected PASS] /css21_dev/html4/table-anonymous-objects-202.htm
  └   → /css21_dev/html4/table-anonymous-objects-202.htm 9a467359e61bf1cd03e915a871942253e111a547
/css21_dev/html4/reference/no_red_antialiasing_a_bc_d-ref.htm 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1
Testing 9a467359e61bf1cd03e915a871942253e111a547 == 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1

  ▶ FAIL [expected PASS] /css21_dev/html4/table-anonymous-objects-201.htm
  └   → /css21_dev/html4/table-anonymous-objects-201.htm 9a467359e61bf1cd03e915a871942253e111a547
/css21_dev/html4/reference/no_red_antialiasing_a_bc_d-ref.htm 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1
Testing 9a467359e61bf1cd03e915a871942253e111a547 == 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1

  ▶ FAIL [expected PASS] /css21_dev/html4/table-anonymous-objects-204.htm
  └   → /css21_dev/html4/table-anonymous-objects-204.htm 9a467359e61bf1cd03e915a871942253e111a547
/css21_dev/html4/reference/no_red_antialiasing_a_bc_d-ref.htm 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1
Testing 9a467359e61bf1cd03e915a871942253e111a547 == 398e25f2937dcbd574d5e6d4aeb2542b5e1752f1

  ▶ PASS [expected FAIL] /mediaqueries-3_dev/html/min-width-tables-001.htm

pcwalton · 2016-10-22T04:58:33Z

I've investigated the failures.

/css21_dev/html4/color-applies-to-014.htm: As far as I can tell, this is an incorrect test because it fails when browsers perform pixel snapping. Specifically, it tests to make sure that (effectively) <div>Filler Text</div> renders identitally to <table><tr><td>Filler</td><td> Text</td></tr></table> (with all borders/padding/margins/spacing on the tables removed). This can fail in browser engines like Servo that perform pixel snapping on table cells, since the "T" glyph can be subpixel aligned in the former case but not the latter.
/css21_dev/html4/table-anonymous-objects-???.htm: These tests rely on summing a set of fractional pixel quantities to overlay the green text "abcd" directly on top of red text "abcd". They then check that the result is precisely equal to the green text "abcd". The display lists come out correct from layout, but WebRender introduces a small amount of rounding error somewhere on Linux (only) and there end up being tiny one-pixel differences around the edges of the letters.

I propose:

Ignoring color-applies-to-014.htm outright, as it's an incorrect test.
Marking table-anonymous-objects-???.htm as failing on Linux, as they trigger minor WebRender bugs on that platform.

Thoughts?

Ms2ger · 2016-10-22T16:03:16Z

As long as you mean "mark as failing" rather than "ignore".

pcwalton · 2016-10-22T18:55:47Z

What's the benefit of marking a test we believe to be invalid as failing instead of ignoring it? Its success depends on the details of the font on the platform.

stshine · 2016-10-24T14:09:14Z

components/layout/construct.rs

+            }
+
+            (FlowClass::Flex, FlowClass::Block) |
+            (FlowClass::Flex, FlowClass::Flex) => {


According to https://drafts.csswg.org/css-flexbox/#flex-items children of a flex container are blockified, and anonymous text is the only kind of elements that needs to be wrapped in a anonymous block.

Flex is OK here because display: flex creates a block-level flex container box.

Oh, I mean IMO the wrapper block should only be needed for FlowClass::Inline since other children are guaranteed to be block-like.

stshine · 2016-10-24T14:09:14Z

components/layout/construct.rs

+                                             parent,
+                                             &[],
+                                             SpecificFragmentInfo::Generic,
+                                             BlockFlow::from_fragment)


I wonder if this is still necessary since all situation with a flex container return true in try_to_add_child().

It shouldn't be. Will remove.

stshine · 2016-10-24T14:09:14Z

components/style/servo_selector_impl.rs

+    ServoAnonymousTable,
+    ServoAnonymousTableRow,
+    ServoAnonymousTableCell,
+    ServoAnonymousFlexBlock,


Anonymous blocks in flexbox are just general blocks (yes, in servo and gecko inline elements are wrapped in an InlineFlow rather than a anonymous block), thus I suggest to call this ServoAnonymousBlock since it can be used in other places.


        layout: Add a new cascading mode that inherits all properties, even

non-inheritable ones. This works like the `modify_style_for_*` functions and will allow us to easily migrate from them to real cascading.


        resources: Style anonymous blocks and anonymous table flows

appropriately in the UA stylesheet.


        layout: Don't use `BlockFlow::assign_block_size()` when assigning flex

flows' block sizes. This ensures that we never collapse margins for flex flows.


        layout: Rewrite the anonymous table object generation code.

This patch introduces a "legalizer", which encapsulates all the anonymous flow generation logic in one place. The legalizer knows how to, for example, place adjacent blocks with `display: table-cell` in the same table row. Improves etsy.com. Closes #13782.


        layout: Remove the old anonymous table object generation code.


        style: Remove `modify_style_for_anonymous_table_object`, now that the

cascade takes care of it.


        layout: Implement `TableWrapper::positioning()`.

The base `Flow` implementation returns the incorrect value.


        layout: Use the new `style_for_anonymous_box()` function for the

anonymous text child of `<input>`.


        layout: Assign border and padding to tables instead of table wrappers.

This is more straightforward when anonymous table object generation is involved.


        layout: Make intrinsic inline sizes take the value of `box-sizing` into

account.


        tests: Update test expectations.


        layout: Minor style cleanup.

pcwalton · 2016-10-26T21:15:13Z

Both issues should be fixed now.

mbrubeck · 2016-10-27T00:47:47Z

@bors-servo r+

bors-servo · 2016-10-27T00:47:48Z

📌 Commit ba005a9 has been approved by mbrubeck

bors-servo · 2016-10-27T01:14:05Z

⌛ Testing commit ba005a9 with merge 6b40f97...


        Auto merge of #13870 - pcwalton:anonymous-table-rewrite, r=mbrubeck

layout: Rewrite anonymous table code, simplify and fix table intrinsic width calculation, and improve safety of flexbox code. Closes #13782.  --- This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/servo/13870)

bors-servo · 2016-10-27T02:49:22Z

☀️ Test successful - arm32, arm64, linux-dev, linux-rel-css, linux-rel-wpt, mac-dev-unit, mac-rel-css, mac-rel-wpt1, mac-rel-wpt2, windows-dev

pcwalton added I-wrong A-layout/block A-layout/tree-builder C-has-manual-testcase A-layout/table A-layout/flex labels Oct 21, 2016

highfive assigned Ms2ger Oct 21, 2016

highfive added the S-awaiting-review label Oct 21, 2016

bors-servo added a commit that referenced this pull request Oct 21, 2016

Auto merge of #13870 - pcwalton:anonymous-table-rewrite, r=<try>

Loading status checks…

f930608

(DO NOT MERGE) layout: Rewrite table stuff. Closes #13782.

highfive added the S-tests-failed label Oct 21, 2016

highfive added the S-needs-rebase label Oct 21, 2016

highfive assigned mbrubeck and unassigned Ms2ger Oct 21, 2016

emilio reviewed Oct 21, 2016

View changes

emilio mentioned this pull request Oct 24, 2016

Update WR - includes a number of optimizations, bsd build fixes, tidy ups. #13900

Merged

stshine reviewed Oct 24, 2016

View changes

pcwalton force-pushed the pcwalton:anonymous-table-rewrite branch from 392ab81 to 2e00b85 Oct 24, 2016

highfive removed the S-tests-failed label Oct 24, 2016

pcwalton force-pushed the pcwalton:anonymous-table-rewrite branch from 2e00b85 to ee3aebd Oct 24, 2016

highfive removed the S-needs-code-changes label Oct 26, 2016

pcwalton added 12 commits Oct 25, 2016

layout: Add a new cascading mode that inherits all properties, even

fb2d1e1

non-inheritable ones. This works like the `modify_style_for_*` functions and will allow us to easily migrate from them to real cascading.

resources: Style anonymous blocks and anonymous table flows

1ab56c0

appropriately in the UA stylesheet.

layout: Don't use `BlockFlow::assign_block_size()` when assigning flex

3d645a9

flows' block sizes. This ensures that we never collapse margins for flex flows.

layout: Remove the old anonymous table object generation code.

a7f64f3

style: Remove `modify_style_for_anonymous_table_object`, now that the

8b138e5

cascade takes care of it.

layout: Implement `TableWrapper::positioning()`.

2ffdfbd

The base `Flow` implementation returns the incorrect value.

layout: Use the new `style_for_anonymous_box()` function for the

a1bbc5d

anonymous text child of `<input>`.

layout: Assign border and padding to tables instead of table wrappers.

d222e3e

This is more straightforward when anonymous table object generation is involved.

layout: Make intrinsic inline sizes take the value of `box-sizing` into

ce2e5dc

account.

tests: Update test expectations.

824a5e2

layout: Minor style cleanup.

Loading status checks…

ba005a9

pcwalton force-pushed the pcwalton:anonymous-table-rewrite branch from 511362f to ba005a9 Oct 26, 2016

highfive added S-awaiting-merge and removed S-awaiting-review labels Oct 27, 2016

bors-servo merged commit ba005a9 into servo:master Oct 27, 2016
2 of 3 checks passed

2 of 3 checks passed

continuous-integration/appveyor/pr AppVeyor build failed
Details

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

homu Test successful
Details

highfive removed the S-awaiting-merge label Oct 27, 2016

bors-servo mentioned this pull request Oct 27, 2016

Migrated -Z trace-layout to serde_json #13740

Merged

4 of 5 tasks complete

stshine mentioned this pull request Nov 8, 2016

Implement column flexbox #14123

Open

This was referenced Jan 3, 2017

Double border appears on elements with display: table #11728

Closed

Hide border and padding on anonymous table objects #11783

Closed

schuster mentioned this pull request Mar 3, 2017

Consider using fake pseudo-elements for anonymous box styling instead of modify_style_for_foo functions #8570

Closed

stshine mentioned this pull request May 10, 2017

Adjacent elements with "display: table-cell" don't become part of the same table #13806

Closed

servo / servo

layout: Rewrite anonymous table code, simplify and fix table intrinsic width calculation, and improve safety of flexbox code. #13870

layout: Rewrite anonymous table code, simplify and fix table intrinsic width calculation, and improve safety of flexbox code. #13870

pcwalton commented Oct 21, 2016 •

edited by larsbergstrom

highfive commented Oct 21, 2016

highfive commented Oct 21, 2016

pcwalton commented Oct 21, 2016

bors-servo commented Oct 21, 2016

bors-servo commented Oct 21, 2016

bors-servo commented Oct 21, 2016

Ms2ger commented Oct 21, 2016

emilio 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.

mbrubeck commented Oct 21, 2016

pcwalton commented Oct 22, 2016

Ms2ger commented Oct 22, 2016

pcwalton commented Oct 22, 2016

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.

pcwalton commented Oct 26, 2016

mbrubeck commented Oct 27, 2016

bors-servo commented Oct 27, 2016

bors-servo commented Oct 27, 2016

bors-servo commented Oct 27, 2016

servo / servo

Join GitHub today

layout: Rewrite anonymous table code, simplify and fix table intrinsic width calculation, and improve safety of flexbox code. #13870

layout: Rewrite anonymous table code, simplify and fix table intrinsic width calculation, and improve safety of flexbox code. #13870

Conversation

pcwalton commented Oct 21, 2016 • edited by larsbergstrom

highfive commented Oct 21, 2016

highfive commented Oct 21, 2016

pcwalton commented Oct 21, 2016

bors-servo commented Oct 21, 2016

bors-servo commented Oct 21, 2016

bors-servo commented Oct 21, 2016

Ms2ger commented Oct 21, 2016

emilio 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.

mbrubeck commented Oct 21, 2016

pcwalton commented Oct 22, 2016

Ms2ger commented Oct 22, 2016

pcwalton commented Oct 22, 2016

This comment has been minimized.

This comment has been minimized.

pcwalton Oct 24, 2016 • edited

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

pcwalton commented Oct 26, 2016

mbrubeck commented Oct 27, 2016

bors-servo commented Oct 27, 2016

bors-servo commented Oct 27, 2016

bors-servo commented Oct 27, 2016

pcwalton commented Oct 21, 2016 •

edited by larsbergstrom

pcwalton Oct 24, 2016 •

edited