1371150 - stylo: Support negative opacity for SMIL

Status: RESOLVED FIXED

(Core :: CSS Parsing and Computation, enhancement, P2)

Description

[Brian Birtles (:birtles, Feb 26~ parental leave)]

Attachment: Test case

The attached test case is a reduced test case from a failing test in dom/smil/test/test_smilCSSFromBy.xhtml

The square should fade from opacity 1 to 0 because we have:

  <rect x="20" y="20" width="200" height="200" fill="purple">
    <animate attributeName="opacity" from="1" by="-1" dur="3s"
      fill="freeze" />
  </rect>

However, it doesn't.

When I debug it I see:

  Parsed opacity property value '1' => 1
  Parsed opacity property value '-1' => 0

So it would appear that we are either failing to parse the negative value, or clamping it too early.

As a result we effectively do:

  <animate attributeName="opacity" from="1" by="0" dur="3s"
    fill="freeze" />

Comment 1

[Jeremy Chen ([e-mail])]

I think we're doing the clamping stuff during the computing... 

http://searchfox.org/mozilla-central/rev/1a054419976437d0778a2b89be1b00207a744e94/servo/components/style/values/specified/mod.rs#707

Comment 2

[Hiroyuki Ikezoe (:hiro)]

I am leaving comments for someone who are going to fix this bug. 

What we need to do here is:

1) Add a flag that represents that current mode allows all numeric values for converting specified value to computed value
2) Set the flag true in Servo_GetAnimationValues (that is used only for SMIL)
3) Check the flag in Opacity.to_computed_value() and clamp the value only if the flag is false.

Once we allow negative color component (bug 1364279), we also need to change Color.to_computed_value, I guess.

Comment 3

[Cameron McCormack (:heycam)]

Thanks for the helpful instructions, Hiro. :-)

https://treeherder.mozilla.org/#/jobs?repo=try&revision=88c2da6603021b38d166e1232149d51372991167

Comment 4

[Cameron McCormack (:heycam)]

Attachment: Bug 1371150 - Update test expectations.

Review commit: https://reviewboard.mozilla.org/r/160406/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/160406/

Comment 5

[Cameron McCormack (:heycam)]

Attachment: Bug 1371150 - style: Don't clamp Opacity values when computing them for SMIL.

Review commit: https://reviewboard.mozilla.org/r/160408/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/160408/

Comment 6

[Cameron McCormack (:heycam)]

I didn't add a separate test case, but I checked that the sub-test in dom/smil/test/test_smilCSSFromBy.xhtml started passing.

Comment 7

[Hiroyuki Ikezoe (:hiro)]

Comment on attachment 8889331 [details]
Bug 1371150 - Update test expectations.

https://reviewboard.mozilla.org/r/160406/#review165666

Thanks for fixing this!

Comment 8

[Hiroyuki Ikezoe (:hiro)]

Comment on attachment 8889332 [details]
Bug 1371150 - style: Don't clamp Opacity values when computing them for SMIL.

https://reviewboard.mozilla.org/r/160408/#review165660

Comment 9

[Cameron McCormack (:heycam)]

Comment on attachment 8889331 [details]
Bug 1371150 - Update test expectations.

Review request updated; see interdiff: https://reviewboard.mozilla.org/r/160406/diff/1-2/

Comment 10

[Cameron McCormack (:heycam)]

Comment on attachment 8889331 [details]
Bug 1371150 - Update test expectations.

https://reviewboard.mozilla.org/r/160406/#review165688

Comment 11

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/26d556a0f07c
Update test expectations. r=hiro

Comment 12

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/26d556a0f07c