test(jump2d): react to Nightly improving stuffed keys handling

[abeldekat]

• I have read CONTRIBUTING.md
• I have read CODE_OF_CONDUCT.md

Details:

• See PR 41529 in neovim/neovim

Test T['start()']['works in Operator-pending mode'] failed. I think it is a timing problem. When I manually follow the test steps everything seems right.
Not sure if adding sleep is the best solution.

[abeldekat]

Not sure if adding sleep is the best solution.

Perhaps it's best to regenerate screenshot tests/screenshots/tests-test_jump2d.lua---start()---works-in-Operator-pending-mode-003

The "dot" on line 5 is similar to the "d" on line 5 in tests/screenshots/tests-test_jump2d.lua---start()---works-in-Operator-pending-mode

[echasnovski]

Not sure if adding sleep is the best solution.

Perhaps it's best to regenerate screenshot tests/screenshots/tests-test_jump2d.lua---start()---works-in-Operator-pending-mode-003

The "dot" on line 5 is similar to the "d" on line 5 in tests/screenshots/tests-test_jump2d.lua---start()---works-in-Operator-pending-mode

Yeah, this does look like some issue with redraw. Which I think is some kind of regression of the linked PR, but not sure if it is consequential enough to create a contained repro and make an issue.

As the core tested stuff is showing the labels after the dot, I think it is safe to ignore non-buffer lines here. There are two approaches here:

• Add a child.cmd('mode') to clear and redraw before problematic screenshot(s). It can also be used instead of child.cmd('redrawstatus') earlier.
• Explicitly ignore text and attributes of non-buffer lines. In this case lines 4 and 5.

Both are fine with me. Let's go with whichever you prefer.

[abeldekat]

Both are fine with me. Let's go with whichever you prefer.

Thanks! I opted to explicitly ignore.

[echasnovski]

Thanks!