author Mike Hommey <>
Fri, 02 Jun 2017 15:07:58 +0900
changeset 410640 f360dd9d411e7ba61531bed2c32a0981cc6c418a
parent 272011 bd079aadd3feeee3f9b9f73c5e0bc4bd6a870722
child 505383 6f3709b3878117466168c40affa7bca0b60cf75b
permissions -rw-r--r--
Bug 1369622 - Restore static asserts for lack of arguments on some macros. r=froydnj Bug 1368932 removed MOZ_STATIC_ASSERT_VALID_ARG_COUNT because it didn't actually work for large numbers of arguments. But it was kind of useful for macros that expand to something broken when they are given no variadic argument at all. Now that we have a macro that doesn't require tricks to count empty arguments lists, however, we can use straightforward static_asserts instead of a generic macro, which has the side effect of providing more context in the error message.

/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* vim: set ts=8 sts=2 et sw=2 tw=80: */
/* This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * file, You can obtain one at */


 * A significant moment in an nsSMILTimedElement's lifetime where a sample is
 * required.
 * Animations register the next milestone in their lifetime with the time
 * container that they belong to. When the animation controller goes to run
 * a sample it first visits all the animations that have a registered milestone
 * in order of their milestone times. This allows interdependencies between
 * animations to be correctly resolved and events to fire in the proper order.
 * A distinction is made between a milestone representing the end of an interval
 * and any other milestone. This is because if animation A ends at time t, and
 * animation B begins at the same time t (or has some other significant moment
 * such as firing a repeat event), SMIL's endpoint-exclusive timing model
 * implies that the interval end occurs first. In fact, interval ends can be
 * thought of as ending an infinitesimally small time before t. Therefore,
 * A should be sampled before B.
 * Furthermore, this distinction between sampling the end of an interval and
 * a regular sample is used within the timing model (specifically in
 * nsSMILTimedElement) to ensure that all intervals ending at time t are sampled
 * before any new intervals are entered so that we have a fully up-to-date set
 * of instance times available before committing to a new interval. Once an
 * interval is entered, the begin time is fixed.
class nsSMILMilestone
  nsSMILMilestone(nsSMILTime aTime, bool aIsEnd)
    : mTime(aTime), mIsEnd(aIsEnd)
  { }

    : mTime(0), mIsEnd(false)
  { }

  bool operator==(const nsSMILMilestone& aOther) const
    return mTime == aOther.mTime && mIsEnd == aOther.mIsEnd;

  bool operator!=(const nsSMILMilestone& aOther) const
    return !(*this == aOther);

  bool operator<(const nsSMILMilestone& aOther) const
    // Earlier times sort first, and for equal times end milestones sort first
    return mTime < aOther.mTime ||
          (mTime == aOther.mTime && mIsEnd && !aOther.mIsEnd);

  bool operator<=(const nsSMILMilestone& aOther) const
    return *this == aOther || *this < aOther;

  bool operator>=(const nsSMILMilestone& aOther) const
    return !(*this < aOther);

  nsSMILTime   mTime;  // The milestone time. This may be in container time or
                       // parent container time depending on where it is used.
  bool mIsEnd; // true if this milestone corresponds to an interval
                       // end, false otherwise.