Remove vertical scrollbar padding from line width calc

[bronsonholden]

Hello 👋

Fix for text edits occasionally rendering a horizontal scrollbar when not necessary, causing what I can only describe as a jitter: sometimes it appears, sometimes not—depending on the text edit's width.

This appears to be caused by the calculation to determine if the max line width exceeds the visible area also accounting for the presence of the vertical scrollbar. As I understand it, this isn't necessary since if the text area wraps, the _update_wrap_at_column function will ensure the line stays within the visible area (which includes the vertical scrollbar, if visible).

And for text fields with no wrapping (specifically I'm thinking of the script editor), ignoring the vertical scrollbar width in this calculation doesn't cause clipped text because lines already are given some extra space on L7283

Before

hscroll-before.mov

After

hscroll-after.mov

Closes #83186

[bronsonholden]

@KoBeWi a test for expected horizontal scroll position is failing, likely since the width of the text is reduced with this change. Is there a way to run the tests visually to verify changing the test expectation to what it now receives (306) doesn't gloss over visual or functional issues with the text edit field itself?

[bronsonholden]

@KoBeWi I added a check that "justifies" the update to 306 as the expected horizontal scroll position since the change just removes the v-scroll bar's width from the total width of rendered text. Let me know if you think there's a better approach. We can remove the newly added test if approved, though. It's really just a sanity check for changing 314 → 306

[KoBeWi]

Is there a way to run the tests visually to verify changing the test expectation to what it now receives (306) doesn't gloss over visual or functional issues with the text edit field itself?

No, you'd have to copy the code e.g. into some part of the editor, or translate it to GDScript. In this case it's easier to use print_line() to see the expected value.

[KoBeWi]

The vmin.x was added in 26a7152, but removing it does not seem to reintroduce the original issue nor cause other adverse effects.

[akien-mga]

Thanks! And congrats for your first merged Godot contribution 🎉

For future contributions, please keep them as a single commit, especially when the follow-up commit is a fixup to a problem in the first one.