chrisgrieser/nvim-various-textobjs

[Bug]: Outer doesn't work for subword

musjj opened this issue · 6 comments

musjj commented

Bug Description

require("various-textobjs").subword always performs action on inner word regardless if it's called with true or false.

Relevant Screenshot

No response

To Reproduce

  1. Run daS on a word
  2. Spaces will be left

neovim version

v0.10.0-dev-355+g3ae62c470

Make sure you have done the following

  • I have updated to the latest version of the plugin.
musjj commented

Thanks for the speedy update, but it doesn't seem to work:

the quick brown fox jumped over the lazy dog
    ^
    daS

Expected:

the brown fox jumped over the lazy dog

Got:

the  brown fox jumped over the lazy dog
chrisgrieser commented

ah right, forgot to move the dash to the end of the set. fixed.

musjj commented

Thank you again!

One small nitpick: there's an edge case with ending words

the quick brown fox
                ^
                daS

Expected:

the quick brown
              ^

Got (it leaves a space in the end):

the quick brown
               ^
chrisgrieser commented

yeah, outer only affects trailing characters, so this result is somewhat intended.

A clean implementation of affecting the preceding character when on the last word of a line would require some extra implementation since the current mechanism does not pay attention to that.

Also, it would add quite a bit of ambiguity in some edge cases (e.g. not visible trailing whitespace affecting how trailing underscores are treated), so I am not even sure that something like that would be desirable. Feel free to open a FR for that, where this can be discussed if other people are interested in it.

musjj commented

The Expected: results is how the native word object works with nvim --clean, but I guess I can live with that. Thanks for the quick fixes!

chrisgrieser commented

yeah, I know and I can see where you are coming from, but the fact that the subword textobj also consideres _ and - as potential trailing characters complicates this and leads to some weird edge cases.

Not saying that I'm against it, just that a proper implementation would be quite tricky