This repository has been archived by the owner. It is now read-only.
Permalink
Commits on Dec 5, 2017
Commits on Dec 4, 2017
  1. Fix a minor typo

    birtles committed Dec 4, 2017
  2. Update changelog

    birtles committed Dec 4, 2017
  3. Fix a couple of minor typos

    birtles committed Dec 4, 2017
  4. Drop the section of the global clock

    birtles committed Dec 4, 2017
    This is no longer needed. The times provided to the update animations
    procedure are generated from a monotonic clock. Furthermore, there's no
    requirement for the clock to be global and referring to this too much
    makes hooking in scroll timelines awkward.
  5. Introduce 'update animations and send events' procedure

    birtles committed Dec 4, 2017
    And make the whole timing model drive from this rather than some
    ill-defined sampling concept.
Commits on Nov 21, 2017
  1. Fix updates to play state definition.

    birtles committed Nov 21, 2017
    See issue #196.
    
    Commit 9e2053f updated the definition
    of the play state but it failed to check for pending tasks. As a result,
    when we are play-pending, we'd report 'paused' since we don't have
    a resolved start time.
    
    I've gone through all the possible states and revised this based on the
    following chart. Hopefully it's right this time.
    
    +---------+-------+------+-------+--------------------------------------+
    | Current | Start | Play | Pause | State / Notes                        |
    |  time   |  time | task |  task |                                      |
    +---------+-------+------+-------+--------------------------------------+
    |    o    |   o   |  o   |   o   | 'idle'                               |
    +---------+-------+------+-------+--------------------------------------+
    |    o    |   o   |  o   |   x   | [1] Normally pausing sets the hold   |
    |         |       |      |       | time (current time) but if the       |
    |         |       |      |       | timeline is cleared while pending    |
    |         |       |      |       | this could happen.                   |
    |         |       |      |       | We should report 'paused' because    |
    |         |       |      |       | we won't be ready until we have a    |
    |         |       |      |       | current time to use and at that      |
    |         |       |      |       | point we will end up being in the    |
    |         |       |      |       | regular paused state: resolved       |
    |         |       |      |       | current time, unresolved start time. |
    +---------+-------+------+-------+--------------------------------------+
    |    o    |   o   |  x   |   o   | [2] Not expected to happen because   |
    |         |       |      |       | we don't expect to play without      |
    |         |       |      |       | either a start time or hold time     |
    |         |       |      |       | (unless we had an aborted pause and  |
    |         |       |      |       | the timelineline was cleared?).      |
    |         |       |      |       | If this were to happen then          |
    |         |       |      |       | presumably we believe we can run the |
    |         |       |      |       | pending play task (or else we would  |
    |         |       |      |       | have canceled it) so we should       |
    |         |       |      |       | 'running' or 'finished'.             |
    +---------+-------+------+-------+--------------------------------------+
    |    o    |   o   |  x   |   x   | [3] Cannot have two pending tasks at |
    |         |       |      |       | once.                                |
    +---------+-------+------+-------+--------------------------------------+
    |    o    |   x   |  o   |   o   | 'idle'                               |
    +---------+-------+------+-------+--------------------------------------+
    |    o    |   x   |  o   |   x   | As with [1], 'paused'.               |
    +---------+-------+------+-------+--------------------------------------+
    |    o    |   x   |  x   |   o   | As with [2], 'running' / 'finished'. |
    +---------+-------+------+-------+--------------------------------------+
    |    o    |   x   |  x   |   x   | As with [3], n/a.                    |
    +---------+-------+------+-------+--------------------------------------+
    |    x    |   o   |  o   |   o   | 'paused'                             |
    +---------+-------+------+-------+--------------------------------------+
    |    x    |   o   |  o   |   x   | 'paused'                             |
    +---------+-------+------+-------+--------------------------------------+
    |    x    |   o   |  x   |   o   | 'running' / 'finished'               |
    +---------+-------+------+-------+--------------------------------------+
    |    x    |   o   |  x   |   x   | As with [3], n/a.                    |
    +---------+-------+------+-------+--------------------------------------+
    |    x    |   x   |  o   |   o   | 'running' / 'finished'               |
    +---------+-------+------+-------+--------------------------------------+
    |    x    |   x   |  o   |   x   | 'paused'                             |
    +---------+-------+------+-------+--------------------------------------+
    |    x    |   x   |  x   |   o   | 'running' / 'finished'               |
    +---------+-------+------+-------+--------------------------------------+
    |    x    |   x   |  x   |   x   | As with [3], n/a.                    |
    +---------+-------+------+-------+--------------------------------------+
Commits on Nov 17, 2017
  1. Further fix indentation in 'play an animation' procedure

    birtles committed Nov 17, 2017
  2. Merge pull request #206 from w3c/pending-member

    birtles committed Nov 17, 2017
    Drop "pending" state and add separate `pending` member to Animation
Commits on Nov 6, 2017
  1. Drop "pending" state and add separate `pending` member to Animation

    birtles committed Nov 6, 2017
    See issue #196.
Commits on Oct 22, 2017
  1. Update editors list

    birtles committed Oct 22, 2017
Commits on Oct 19, 2017
  1. Various fixes to 'processing a keyframes argument' procedure

    birtles committed Oct 19, 2017
    These fixes address issues introduced in changeset
    8efd180 and changeset
    f43ecdf (issue #148) and identified
    during [review of the Gecko
    implementation](https://bugzilla.mozilla.org/show_bug.cgi?id=1398038).
Commits on Oct 18, 2017
  1. Allow specifying composite/offset/easing when using property-indexed …

    birtles committed Sep 8, 2017
    …keyframes
    
    This closes issue #148.
Commits on Oct 16, 2017
  1. Allow specifying composite/offset/easing when using property-indexed …

    birtles committed Sep 8, 2017
    …keyframes
    
    See issue #148.
Commits on Oct 13, 2017
  1. Fix cut off label in diagram

    birtles committed Oct 13, 2017
  2. Merge pull request #202 from w3c/sticky-interval-end-changes

    birtles committed Oct 13, 2017
    Revise simple iteration progress when an interval is clipped
  3. Revise simple iteration progress when an interval is clipped

    birtles committed Oct 12, 2017
    This closes #201.
Commits on Oct 6, 2017
  1. Drop issue about using missing offsets instead of null offsets

    birtles committed Oct 6, 2017
    As discovered in #198, this is a non-starter for compatibility reasons.
Commits on Oct 5, 2017
  1. Fix a couple of references

    birtles committed Oct 5, 2017
  2. Reference issue #200

    birtles committed Oct 5, 2017
Commits on Sep 21, 2017
  1. Clarify that invalid easings throw when processing a keyframes argument

    birtles committed Sep 21, 2017
    This fixes #197.
  2. Update timing and animation model graphic

    birtles committed Sep 21, 2017
Commits on Sep 20, 2017
  1. Try specifying the full path to grunt

    birtles committed Sep 20, 2017
  2. Drop SharedKeyframeList

    birtles committed Sep 20, 2017
    This closes #199.
Commits on Aug 25, 2017
  1. Fix calculation of hold time at end of async pause task

    birtles committed Aug 25, 2017
    The calculation used to read:
    
      hold time = ready time - (start time / playback rate)
    
    But this is clearly wrong since the relationship between the hold time
    and other times is based on the calculation of the current time which
    is:
    
      current time = (timeline time - start time) x playback rate
    
    Accordingly, this patch updates the calculation as follows:
    
      hold time = (ready time - start time) x playback rate
    
    I have confirmed that the current two implementations of this behavior
    both agree on this updated calculation of the hold time:
    
      https://hg.mozilla.org/mozilla-central/file/2306e153fba9/dom/animation/Animation.cpp#l1219
      https://cs.chromium.org/chromium/src/third_party/WebKit/Source/core/animation/Animation.cpp?l=360&rcl=f16e90904fb2df1866eb8580543c32528e557a60
Commits on Aug 23, 2017
  1. Drop issues about the possibility of cancelable promises

    birtles committed Aug 23, 2017
    According to #141 these are not going to eventuate after all.
    
    This closes #141.
  2. Use [Exposed] consistently

    birtles committed Aug 23, 2017
    Apparently this is needed for heycam/webidl#365.