2c93da9a70
docs: Add comments documenting subtle points of AS3 loop & tag queueing behavior
2022-08-25 19:24:56 -06:00
074f2ff76f
core: Replace `Vec` of queued tags with a more restrictive enum that enforces our queueing invariants
2022-08-25 19:24:56 -06:00
59623a8452
core: Fix a potential stack overflow caused by gotos triggering frame scripts on one another.
2022-08-25 19:24:56 -06:00
4510d51e18
tests: Fix regression in `as3_loaderinfo_events` caused by the frame lifecycle fixes
2022-08-25 19:24:56 -06:00
05caa0fceb
core: Explicit gotos drain the tag queue of any prior operations.
...
This is in keeping with the whole idea of a "recursive frame": gotos run the entire frame lifecycle on the target clip, including broadcasts for `frameConstructed` and `exitFrame`.
2022-08-25 19:24:56 -06:00
6f0de246ed
core: Clear queued tags if AS3 attempts a goto in the middle of a loop.
2022-08-25 19:24:56 -06:00
ed28578c12
core: Fast-forwarding gotos should run frame scripts on all removed children
2022-08-25 19:24:56 -06:00
bcb1030883
core: Only `Construct` and `Enter` phases want forced script frames on gotos
2022-08-25 19:24:56 -06:00
3677a0107b
core: Only force-queue frame scripts if the goto is not a no-op.
2022-08-25 19:24:56 -06:00
ac2b82f26e
core: Gotos *always* queue the target frame's scripts on the target clip, even if we already executed that script beforehand.
2022-08-25 19:24:56 -06:00
9bff6406d1
core: Allow queueing both a remove and a place tag at the same depth, in that order only.
2022-08-25 19:24:56 -06:00
2929681a93
core: No-op gotos should be treated as rewinding.
2022-08-25 19:24:56 -06:00
7a09bbfaab
core: In AVM2, run removals before frame advance.
...
We still retain the queue system as events are fired at removal time, and those events can trigger more gotos. If such a goto happens, AS3 code will hit a clip still in the old state rather than an inconsistent one. I don't have test coverage for this exact scenario just yet.
2022-08-25 19:24:56 -06:00
d691543c4c
core: Child removals appear to resolve in reverse render order.
2022-08-25 19:24:56 -06:00
8feb3fc7b0
core: Fast-forward removals as the result of a goto happen before the frame number advances.
2022-08-25 19:24:56 -06:00
66c24ff188
core: Queue new object placement at loop time on the tag queue.
2022-08-25 19:24:56 -06:00
0353d176e4
core: Movie clip rewinds should not remove any children until after the new timeline position has been set.
2022-08-25 19:24:56 -06:00
99003643ac
core: Same-frame gotos should only proceed on AVM2.
2022-08-25 19:24:56 -06:00
cdfc6f5eae
core: Explicit gotos must run even if they are no-ops.
2022-08-25 19:24:56 -06:00
cb86340fb1
core: `frameConstructed` during a goto fires after the goto commands have fully completed.
2022-08-25 19:24:56 -06:00
51401ae660
core: Remove AVM2 compatibility hacks surrounding place frame.
...
They are no longer needed, frame advance happens at enterframe time.
2022-08-25 19:24:56 -06:00
1eef8a47c9
core: Movie clips run all their tags at `enter_frame`, and place/remove actions are queued until later.
...
This prevents repeated place/remove actions at the same depth from resulting in repeated event dispatches.
2022-08-25 19:24:56 -06:00
963144e59c
core: Centralize all the various calls to `enter_frame`/`construct_frame` into a single method, `catchup_display_object_to_frame`.
...
The rationale for the catch-up logic is as follows:
* We must always enter-frame and construct objects, even if those respective display events haven't happened yet.
* Display objects created in event handlers still need to run catchup phases, otherwise they will tag-stream desync
* Frame scripts are never triggered by catchup phases
* `exit_frame` is not a catchup phase as it is *only* an event broadcast currently
2022-08-25 19:24:56 -06:00
0400d4dee4
core: `Player` should start in the `Idle` phase
2022-08-25 19:24:56 -06:00
f0ef8adb42
core: Add a new frame phase to represent non-frame work such as input event handlers.
2022-08-25 19:24:56 -06:00
ae530b5d6b
core: Track what part of the frame processing loop we're in
2022-08-25 19:24:56 -06:00
ee2454e09f
core: Frame actions should start with `enterFrame`; not `exitFrame`.
2022-08-25 19:24:56 -06:00
1f0b8bdda4
core: In AVM2, movie clips process `RemoveObject` before any other frame actions are done.
2022-08-25 19:24:56 -06:00
318b018da8
core: `enter_frame` should be an event that all display objects get and can respond to
2022-08-25 19:24:56 -06:00
beb0545b11
tests: Add tests for gotos interacting with frame scripts.
2022-08-25 19:24:56 -06:00
6569edda88
tests: Add a goto-and-stop test that inspects the children after the goto.
2022-08-25 19:24:56 -06:00
c3add3db93
tests: Ignore the `as3_simplebutton_constr_childevents` test as it is broken by fixes I moved to another PR
2022-08-25 19:24:56 -06:00
38f28ce80c
tests: Add a second, more thorough `clickgoto2` test
2022-08-25 19:24:56 -06:00
f16b76866a
tests: Add timeline and stage manipulation tests at user event time.
2022-08-25 19:24:56 -06:00
81eb5ed8ac
tests: Add tests for symbol instantiation in `enterFrame`, `exitFrame`, and `frameConstructed`
2022-08-25 19:24:56 -06:00
27d509f169
tests: Extend the goto tests with more non-looping jumps
2022-08-25 19:24:56 -06:00
4d3788bfe1
tests: Add tests for AS3 gotos during `enterFrame`, `frameConstructed`, and/or `exitFrame`
2022-08-25 19:24:56 -06:00
c10417ddb8
tests: `as3_movieclip_displayevents_timeline` should log events on the root movie clip too
2022-08-25 19:24:56 -06:00
f1d203311f
tests: `as3_movieclip_displayevents_looping` should have a `RemoveObject` in it
2022-08-25 19:24:56 -06:00
ee1f39bba7
tests: `as3_movieclip_displayevents_xyzframeplay` tests should also list what children are on the timeline
2022-08-25 19:24:56 -06:00
b3793f9b4c
tests: Oops, forgot to recompile the exitframe play test.
2022-08-25 19:24:56 -06:00
10955c3a03
tests: Add tests for playing and pausing a movie clip in `enterFrame`, `frameConstructed`, and `exitFrame`
2022-08-25 19:24:56 -06:00
795dafa46a
tests: Fix the test output for `as3_movieclip_displayevents_dblhandler`.
...
The test output for this test is sensitive to where we cut off each frame, because it doesn't stop all of it's handlers at the end of the test. Flash Player will just print lines forever so the end of the test is entirely arbitrary.
2022-08-25 19:24:56 -06:00
1a4b977325
tests: `as3_movieclip_displayevents_timeline` now explicitly winds a clip back and forth to better demonstrate a bug I ran into with frame numbering
2022-08-25 19:24:56 -06:00
0c18e8e99a
tests: Log `this.currentFrame` on every event watcher we can.
2022-08-25 19:24:56 -06:00
72f878bb9d
tests: Add a test for the order of events surrounding `gotoAndStop`
2022-08-25 19:24:56 -06:00
d152cc3e64
avm2: Port `Math` to ActionScript
2022-08-25 16:03:08 -07:00
4e9cc919a1
avm2: Fix bindings of empty namespace
...
Previously native functions defined inside a `package {}` block
(without a package name) would cause a syntax error by two consecutive
`::`.
2022-08-25 16:03:08 -07:00
91247f5779
tests: Only require `ruffle_render_wgpu` dependency with 'imgtests'
...
This should speed up developer workflows when working on non-image
related changes. CI will still run all image tests.
2022-08-25 15:11:18 -07:00
211a64b402
webgl: Replace Box<dyn Error> with actual Error enum
2022-08-25 23:43:21 +02:00
David Wendt