This is the first time to import patches into Neovim for me and almost all hunks were not automatically applied by patch command. So something may be wrong. Please let me know in the case. I'll fix it.

@rhysd
Thanks!
You seem to have used scripts/vim-patch.sh, which is good. Patches often do not apply before other (earlier) patches are missing, or due to previous linting, and/or things not having been moved to new files (yet, or on purpose).

@blueyed Thank you for the advice. Yeah, I also fixed some lint errors.

@rhysd
See also https://github.com/neovim/nvimdev.nvim - it includes a Neomake maker, so you can do :Neomake lint also. /cc @janlazo

CI only failed on macOS with GCC. I'm investigating the failure.

[  FAILED  ]1test, listed below:
[  FAILED  ]...is/build/neovim/neovim/test/functional/core/job_spec.lua @ 599: jobs jobwait will run callbacks while waiting
...is/build/neovim/neovim/test/functional/core/job_spec.lua:622: Expected objects to be the same.

Passed in:

(table) {
  [1] = 'notification'
  [2] = 'wait'
 *[3] = {
   *[1] = 3 } }

Expected:

(table) {
  [1] = 'notification'
  [2] = 'wait'
 *[3] = {
   *[1] = 4 } }

stack traceback:
	...is/build/neovim/neovim/test/functional/core/job_spec.lua:622: in function <...is/build/neovim/neovim/test/functional/core/job_spec.lua:599>

I'm not sure why it failed, but as far as I read test/functional/core/job_spec.lua, it seems not related to this patch.

That's normal, especially for the macOS. Job and timer tests are flaky on macOS in the spite of the numerous adjustments.

My only concern is if these patches depend on older patches that haven't been (fully) ported from Vim.

This one is a bug in jobwait(), or rather jobwait() never implemented what the test claims it should do. Now we at least have a datapoint that #10021 didn't fix/hide the flakiness. I hope to look into fully fixing it soon.

@janlazo @bfredl

I understood the test case is flaky. Actually the test case passed in next CI job.. Let me ignore the failure in this PR.

@justinmk

Thanks for the review. I removed updates in version.c from the commits and did git push -f.

My only concern is if these patches depend on older patches that haven't been (fully) ported from Vim.

I basically agree with the concern. I took care about applying these patches not breaking something, but I'm not 100% sure. The codebase related to these patches seem almost the same as original Vim. The only non-trivial change between Neovim and Vim while applying these patch was Neovim does not implement here document for now.

const foo =<< EOS
hello,
world
EOS

Above code is valid for Vim but isn't for Neovim.

@rhysd 8.1.1539 is partially merged because Neovim does not support heredoc yet.
I don't know which patch supports the following (also omitted from ported 8.1.1539 patch):

:cons[t] [{name1}, {name2}, ...] .= {expr1}

@janlazo that doc was later corrected in vim/vim@1c196e7

Yeah, the line was my mistake when I stole the description from let. And I fixed it later. :const cannot modify existing variables. I'm sorry for confusion.