defaults: set nostartofline #11135

teto · 2019-10-01T14:21:44Z

Having the cursor change column is surprising.

I haven't found anything that disturbed me running nostartofline in my init.vim.
@justinmk could you precise the annoying scenarios you are confronted with in tpope/vim-sensible#60 ?

justinmk · 2019-10-02T02:29:40Z

Need to update test/functional/helpers.lua (and probably some individual test setup routines)
Needs mention in vim_diff.txt

@justinmk could you precise the annoying scenarios you are confronted with in tpope/vim-sensible#60 ?

@teto my intention there was only to avoid a "drastic" change. I've since been using nostartofline for years without issue, and it is indeed superior to the Vi behavior.

justinmk · 2019-11-25T04:28:33Z

Needs mention in vim_diff.txt
failing test: test/functional/legacy/breakindent_spec.lua

teto · 2019-12-01T20:35:04Z

is that ok if I modernize the test/functional/legacy/breakindent_spec.lua to itemize tests ?

justinmk · 2019-12-01T21:33:28Z

is that ok if I modernize the test/functional/legacy/breakindent_spec.lua to itemize tests ?

the feed_command calls could be one big helpers.source, but otherwise it's best to keep it similar to the original oldtest. I guess we actually don't need it since it might be redundant with src/nvim/testdir/test_breakindent.vim (but then one should check the git history carefully)

justinmk · 2019-12-01T22:59:59Z

test/functional/legacy/breakindent_spec.lua

@@ -129,7 +129,7 @@ describe('breakindent', function()
    feed_command('let g:test="Test 15: breakindent + visual blockwise delete #2"')
    feed_command('%d')
    feed_command('normal! 4a1234567890')
-    feed_command([[exec "normal! >>\<C-V>3f0x"]])
+    feed_command([[exec "normal! >>^\<C-V>3f0x"]])


this shouldn't be necessary. I think what is happening is, the set all& call on line 95 is setting the wrong default. because options.lua should set vim=false

with nostartofline, the cursor remains at the end of the insertion after normal! 4a1234567890 whereas with startofline, it puts it at the beginning of the line so setting the correct default would also require my modification I think.
if I start the binary, :verbose set startofline? returns nostartofline.
I confess I am kinda confused by the values in options.lua, I assume I should have in the end the following behavior:

:set startofline&vi -> startofline :set startofline&vim -> startofline :set startofline& -> nostartofline

?

:set startofline&vim -> startofline

? That would contradict the actual default. Although it's debatable, up to now we've updated "vim defaults" to whatever our defaults are. Otherwise we would need to introduce &nvim variant for :set.

In any case, it would make more senes to set the option after :set all& rather than update various :normal invocations in the test....

Otherwise we would need to introduce &nvim variant for :set.

I've thought about that being useful myself before.

@blueyed yeah, it's probably a good idea. Then we could use :set all&vim as a way of checking "vim defaults" against user reports.

src/nvim/options.lua

blueyed · 2019-12-02T05:49:01Z

I guess we actually don't need it since it might be redundant with src/nvim/testdir/test_breakindent.vim (but then one should check the git history carefully)

👍

justinmk · 2019-12-03T03:14:50Z

src/nvim/testdir/test_edit.vim

@@ -14,6 +14,7 @@ func! Test_edit_00b()
  call setline(1, ['abc '])
  inoreabbr <buffer> h here some more
  call cursor(1, 4)
+  set startofline


wouldn't this make more sense in src/nvim/testdir/setup.vim ?

agreed & done.

justinmk · 2019-12-03T03:15:21Z

src/nvim/testdir/test_breakindent.vim

@@ -261,7 +262,7 @@ function Test_breakindent15()
  call s:test_windows('setl breakindent briopt= ts=8 sw=8')
  vert resize 30
  norm! 4a1234567890
-  exe "normal! >>\<C-V>3f0x"
+  exe "normal! >>^\<C-V>3f0x"


still wondering why we are tweaking random test actions instead of setting the default.

Having the cursor change column can be surprising. Force startofline in functional and old tests. Remove the functional breakindent test, as it's a subset of the oldtest one.

justinmk · 2019-12-04T05:00:42Z

The merge looks weird because of a GitHub bug: "Retry" button after "failed to merge" decided to do a merge-commit with wrong commit message. But it is a merge commit anyway, where @teto is the author of 2nd parent.

teto · 2019-12-04T09:48:30Z

thanks ! I am bit nervous to how people will react about the change as it impacts a few bindings like gg/>> etc. Let's see :D

zhoosch · 2020-02-01T17:44:48Z

I do not like changes in the defaults. Because i use vim and neovim with only one rc file. Even if someone changes the defaults i have to change these too in my configurations too, to have the same result in both editors. So my startup of vim/neovim eats more and more time.

So why do the ppl who like this new behavior do not change this one line in their init? set nostartofline. Too much effort?

justinmk · 2020-02-01T21:31:53Z

Nvim defaults typically will save time since they are intended to be more commonly desirable.

And if you're worried about startup time then you should be worried about Vim's defaults.vim.

neovim-discourse · 2022-01-19T13:16:10Z

This pull request has been mentioned on Neovim Discourse. There might be relevant details there:

https://neovim.discourse.group/t/copying-original-vi-behaviour/1889/9

neovim-discourse · 2022-01-28T20:04:17Z

This pull request has been mentioned on Neovim Discourse. There might be relevant details there:

https://neovim.discourse.group/t/poll-should-startofline-option-be-on-or-off-by-default/1952/3

justinmk added the defaults issues or PRs involving changing the defaults label Oct 2, 2019

teto force-pushed the nostartofline branch from 6bb45d5 to b624f17 Compare November 21, 2019 07:55

teto force-pushed the nostartofline branch from b624f17 to f480cfd Compare December 1, 2019 22:49

justinmk reviewed Dec 1, 2019

View reviewed changes

src/nvim/options.lua Outdated Show resolved Hide resolved

teto force-pushed the nostartofline branch from f480cfd to 0b9c894 Compare December 2, 2019 23:26

justinmk reviewed Dec 3, 2019

View reviewed changes

teto force-pushed the nostartofline branch from 0b9c894 to 3beef8e Compare December 3, 2019 11:32

defaults: set nostartofline

3beef8e

Having the cursor change column can be surprising. Force startofline in functional and old tests. Remove the functional breakindent test, as it's a subset of the oldtest one.

justinmk mentioned this pull request Dec 4, 2019

defaults, part 2 #6289

Closed

20 tasks

justinmk merged commit 3aa95ef into neovim:master Dec 4, 2019

teto deleted the nostartofline branch December 4, 2019 09:47

teto mentioned this pull request Apr 9, 2021

Following HEAD: breaking changes on master #14090

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

defaults: set nostartofline #11135

defaults: set nostartofline #11135

teto commented Oct 1, 2019

justinmk commented Oct 2, 2019 •

edited

justinmk commented Nov 25, 2019 •

edited

teto commented Dec 1, 2019

justinmk commented Dec 1, 2019 •

edited

justinmk Dec 1, 2019

teto Dec 1, 2019

justinmk Dec 1, 2019 •

edited

blueyed Dec 2, 2019

justinmk Dec 2, 2019

blueyed commented Dec 2, 2019

justinmk Dec 3, 2019

teto Dec 3, 2019

justinmk Dec 3, 2019

justinmk commented Dec 4, 2019

teto commented Dec 4, 2019

zhoosch commented Feb 1, 2020

justinmk commented Feb 1, 2020

neovim-discourse commented Jan 19, 2022

neovim-discourse commented Jan 28, 2022

defaults: set nostartofline #11135

defaults: set nostartofline #11135

Conversation

teto commented Oct 1, 2019

justinmk commented Oct 2, 2019 • edited

justinmk commented Nov 25, 2019 • edited

teto commented Dec 1, 2019

justinmk commented Dec 1, 2019 • edited

justinmk Dec 1, 2019

Choose a reason for hiding this comment

teto Dec 1, 2019

Choose a reason for hiding this comment

justinmk Dec 1, 2019 • edited

Choose a reason for hiding this comment

blueyed Dec 2, 2019

Choose a reason for hiding this comment

justinmk Dec 2, 2019

Choose a reason for hiding this comment

blueyed commented Dec 2, 2019

justinmk Dec 3, 2019

Choose a reason for hiding this comment

teto Dec 3, 2019

Choose a reason for hiding this comment

justinmk Dec 3, 2019

Choose a reason for hiding this comment

justinmk commented Dec 4, 2019

teto commented Dec 4, 2019

zhoosch commented Feb 1, 2020

justinmk commented Feb 1, 2020

neovim-discourse commented Jan 19, 2022

neovim-discourse commented Jan 28, 2022

justinmk commented Oct 2, 2019 •

edited

justinmk commented Nov 25, 2019 •

edited

justinmk commented Dec 1, 2019 •

edited

justinmk Dec 1, 2019 •

edited