core: Run frames for AVM2 movies loaded into AVM1 #21959
Conversation
Lord-McSweeney
added
mixed-avm
Lord-McSweeney
added
the
waiting-on-review
Lord-McSweeney
force-pushed
the
avm2-avm1-orphan
branch
from
a78c62b to
196baac
Compare
core/src/orphan_manager.rs
Outdated
| // Also, we consider AVM2 MovieClips loaded into an AVM1 parent to always | ||
| // be orphans, since we know they were placed by AVM1 code and not by | ||
| // the timeline. | ||
| dobj.placed_by_script() || has_avm1_parent |
There was a problem hiding this comment.
Shouldn't such movies be placed by script anyway?
There was a problem hiding this comment.
Hmm, from what I understand, placed_by_script is only used by AVM2? As in, AVM1 code won't set_placed_by_script on DOs that it creates?
There was a problem hiding this comment.
Oh right, but in this case you have placed_by_avm1_script. As a side note, I think we should aim at merging these two.
There was a problem hiding this comment.
Hmm, but actually we don't really care about placed_by_avm1_script? As the comment says, we consider AVM2 MovieClips loaded into an AVM1 parent to be orphans no matter what, even if the AVM1 parent was placed by the timeline, because we need that AVM2 movie to tick its frames no matter what.
There was a problem hiding this comment.
Maybe I'm misunderstanding something, but the comment says that we do that exactly because the object couldn't have been placed by the timeline. In which case it makes sense to check placed_by_avm1_script instead of using this assumption, right?
I'm looking into merging placed_by_avm1_script and placed_by_avm2_script into one
Lord-McSweeney
force-pushed
the
avm2-avm1-orphan
branch
2 times, most recently
from
2316a64 to
6cec69f
Compare
Lord-McSweeney
added
waiting-on-author
Lord-McSweeney
force-pushed
the
avm2-avm1-orphan
branch
from
6cec69f to
7271017
Compare
Lord-McSweeney
added
waiting-on-review
Lord-McSweeney
force-pushed
the
avm2-avm1-orphan
branch
from
7271017 to
76e859d
Compare
Instead, do it explicitly everywhere it would've been done before. This allows us to make `TDisplayObject::object2` take a `Mutation` instead of an `UpdateContext`
Lord-McSweeney
force-pushed
the
avm2-avm1-orphan
branch
from
76e859d to
68d9350
Compare
|
|
||
| /// The next MovieClip in the AVM1 execution list. | ||
| /// | ||
| /// `None` in an AVM2 movie. |
There was a problem hiding this comment.
Could you elaborate here when it's None vs Some? IIUC it's Some when either this movie clip or the root MC is from AVM1. Because without a comment I'd assume it's for AVM1 only.
| if !self.movie().is_action_script_3() { | ||
| // Run my load/enterFrame clip event. | ||
| // Run my load/enterFrame clip event. | ||
| if !self.movie().is_action_script_3() || !context.root_swf.is_action_script_3() { |
There was a problem hiding this comment.
Could you also add a comment here explaining this situation? I.e. why it is being run for an AVM2 movie
| // To detect this, we check 'placed_by_avm2_script'. This flag get set to 'true' | ||
| // To detect this, we check 'placed_by_script'. This flag get set to 'true' |
There was a problem hiding this comment.
These 2 changes should be reverted
|
This is the wrong approach- when AVM1 loads an AVM2 movie, we probably want to interpret the loaded movie as actually being AVM1, except for the running of DoAction tags |
Lord-McSweeney
removed
the
waiting-on-review
2 participants
Fixes #21007