fix(shared) better inner parameter selection #702
Conversation
The logic of including a point in a range needs `-1` end_col offset only if the point is from the cursor position but not from the end of another range. Fixes nvim-treesitter#700
|
Examples to cover by test: vim.print(1, tonumber('1')) -- on `tonumber`
vim.print({1, '2'}, 3) -- on closing brace
vim.print({1, '2'}) -- on closing braceCurrently the change breaks |
Fixed and added test cases. |
| @@ -25,6 +25,10 @@ describe("command equality Python:", function() | |||
| run:compare_cmds("aligned_indent.py", { row = 1, col = 0, cmds = { "dfn", "d;" } }) | |||
| -- select using move | |||
| run:compare_cmds("aligned_indent.py", { row = 1, col = 0, cmds = { "d]a", "v]ad", "c]a" } }) | |||
| run:compare_cmds("selection_mode.py", { row = 2, col = 4, cmds = { "dam", "dVam", "vamd", "Vamd" } }) | |||
| run:compare_cmds("selection_mode.py", { row = 2, col = 4, cmds = { "dam", "dVam", "vamd", "Vamd", "dG" } }) | |||
| run:compare_cmds("selection_mode.py", { row = 3, col = 8, cmds = { "kdG", "dam", "dVam", "vamd", "Vamd" } }) | |||
There was a problem hiding this comment.
dG, kdG motions break when the reference code changes slightly. Maybe better if you do something like d]M which is more meaningful
There was a problem hiding this comment.
My intention was comparing the plugin action with something with something independent, but still reliable. IIRC the test compares results and expect only equation or only difference. In the former case I got passed test when all keystrokes gave the same bad result. While the test is defined as now, let allow to compare with something reliable (line 25 alraedy contains d;).
|
I don't use neovim nightly so I could only find time now to test it.. Thanks for reporting it. It's definitely odd and in I'd like to accept this PR but I didn't follow much on Can you investigate a little why this is an issue on main and not master? Thank you! |
| ---@return boolean | ||
| local function is_in_range(range, row, col) | ||
| local function is_in_range(range, row, col, end_col_offset) |
There was a problem hiding this comment.
Please add a brief comment on why the end_col_offset is needed (in what situation,) and a link to the issue.
The logic of including a point in a range needs
-1end_col offsetonly if the point is from the cursor position
but not from the end of another range.
Fixes #700