Update CanvasLayer 'Follow Viewport' documentation to match its behavior

[kthang55]

Fixes #98463 by updating the documentation for the 'Follow Viewport' property to accurately describe its current behavior.

[AThousandShips]

This should be fixed by adjusting the documentation IMO, this breaks compatibility in a major way

[kthang55]

Fair enough, how should I best proceed? Reverting the changes and making another commit with changes in the documentation? Or should I make a new PR altogether?

[AThousandShips]

You can update this PR

[Mickeon]

I don't understand why the documentation change has been done in Parallax2D. It looks like the original issue's confusion stem from CanvasLayer. No mention of Parallax2D.

[kthang55]

My mistake, this should be fixed now with the latest push :D

[Mickeon]

This is not the first time users have gotten confused by this setting. I've seen it crop up sporadically in conversation that it should be the opposite.
So maybe this makes more sense, but it would be nice to factually check this.

[Meorge]

For reference, this is what it currently says (in case it gets updated in the future):

remains anchored in a fixed position on the screen and does not move with the camera. If disabled, the [CanvasLayer] will move along with the camera's transform

This wording still feels ambiguous to me, like both descriptions for true/false are saying the same thing.

"Remains anchored in a fixed position on the screen and does not move with the camera" sound contradictory - if it remains in a fixed position on the screen (not the world), but the camera moves, then it should look like it is moving with the camera within the world space, because the camera is our view into the world.

"Moving along with the camera's transform" sounds to me like, if you scroll the camera to the right for example, the CanvasLayer will move to the right as well ("along with the camera's transform"), and thus appear to be in a fixed position on the screen.

[kthang55]

What about something like this?
"If enabled, the [CanvasLayer] stays in a fixed position relative to the screen, ignoring camera movement.
If disabled, the [CanvasLayer] moves along with the camera, maintaining its position in world space."

[Meorge]

I'm thinking that sounds good, but perhaps with "ignoring camera movement" and "moves along with the camera" removed? Simply stating "if enabled, it stays in a fixed position on the screen; if disabled, it maintains its position in world space" seems clearest to me.

[kthang55]

That sounds good, this would be the full documentation for it then:
"If enabled, the [CanvasLayer] stays in a fixed position on the screen.
If disabled, the [CanvasLayer] maintains its position in world space.
Together with [member follow_viewport_scale], this can be used for a pseudo-3D effect."
Does that look good?

[Meorge]

Sounds good to me personally! It may be good to check with others and ensure it sounds good to them too, though.

[kthang55]

Pushed the new change and awaiting further feedback!

[Mickeon]

The updated description is still fine by me, assuming accuracy. I would suggest updating your PR's own description as it's outdated.

[kthang55]

Would that be the comment that is at the top? The one starting with "Fixes #98463..."?

Edit: It has been changed

[akien-mga]

Rebased to squash the commits, should be ready to merge.

[akien-mga]

Thanks! And congrats for your first merged Godot contribution 🎉

[kthang55]

Thanks!

[rt9391]

The description does not match the actual behavior, in fact it does the opposite.

I made a project here.
"If enabled, the CanvasLayer stays in a fixed position on the screen. If disabled, the CanvasLayer maintains its position in world space."
The GUI nodes (those labels) is in a canvaslayer that DISABLED follow_viewport_enabled, whitch would be "maintains its position in world space" according to the doc.
But apparently these GUI only "fixed position on the screen", not maintains in wolrd space.

CanvasLayerFviewTest.zip

MRP file:

QQ.20250219174931.mp4