ManimCE 0.19.0 - caching issue

[uwezi]

Description of bug / unexpected behavior

Rendering the scene below twice with caching enabled results in a false placement of one object at the end.
(follow-up of https://discord.com/channels/581738731934056449/1337379695628521513/1337379695628521513 )

Expected behavior

The second (and subsequent) rendering of the scene should result in the same output as the first rendering.

How to reproduce the issue

Code for reproducing the problem

class Sin2(Scene):
    def construct(self):
        ax = Axes(x_range=(-2, 2, 1), y_range=(-2, 2, 1), x_length=6)
        trigo_circle = Circle(radius=1.5)
        self.add(ax, trigo_circle, Dot(color=RED).shift(1.5*RIGHT))

        theta = ValueTracker(0)

        neg_point = always_redraw(
            lambda: Dot(
                trigo_circle.point_at_angle(-theta.get_value()),
                color=YELLOW
            )
        )

        neg_arc = always_redraw(
            lambda: Arc(radius=1.4, angle=-(theta.get_value() % TAU), stroke_width=2,
                color=YELLOW
            )
        )

        pos_neg = always_redraw(
            lambda: Tex(
                "$e^{-i \\theta}$",
                color=YELLOW
            ).move_to(0.6*neg_point.get_center())
        )

        self.add(pos_neg, neg_point, neg_arc)

        self.play(theta.animate(run_time=5).set_value(TAU))

        self.play(theta.animate(run_time=2).set_value(TAU+2.4))

        self.wait()

Additional media files

Images/GIFs

Result of the first rendering

Sin2.mp4

Result of the second rendering

Sin2.mp4

System specifications

System Details

• OS (with version, e.g., Windows 10 v2004 or macOS 10.15 (Catalina)): Windows 10
• Python version (python/py/python3 --version): 3.13.0

[huguesdevimeux]

Hi,

I don't have much time to investigate but that typically looks like a variation of #550.
Upon using a cached animation, Manim artificially sets up the global t (current timestamp) to be at the end of the animation. The issue might arise from there.