Some commands still don't support long paths under Windows

[grv87]

• I was not able to find an open or closed issue matching what I'm seeing

• Which version of Git for Windows are you using? Is it 32-bit or 64-bit?

git version 2.17.1.windows.2
cpu: x86_64
built from commit: a60968cf435951d9411fc0f980a2e362d5cccea2
sizeof-long: 4

• Which version of Windows are you running? Vista, 7, 8, 10? Is it 32-bit or 64-bit?

Microsoft Windows [Version 6.1.7601]

• What options did you set as part of the installation? Or did you choose the
  defaults?

Editor Option: VIM
Path Option: Cmd
SSH Option: OpenSSH
CURL Option: OpenSSL
CRLF Option: CRLFAlways
Bash Terminal Option: MinTTY
Performance Tweaks FSCache: Enabled
Use Credential Manager: Enabled
Enable Symlinks: Enabled

• Any other interesting things about your environment that might be related
  to the issue you're seeing?

core.longpaths config is set to true

Details

While there was msysgit#122, some git commands still don't work correctly with long paths under Windows.
I found problems with 2 commands:

1. git init - error, when current directory name + filename became longer than MAX_PATH
2. git push to (bare) repository located in local directory

Working examples:
https://gist.github.com/grv87/90e72ac5d6b5964f16c3de6f1ec4fb05
(Written in Go, since it is one of few languages that correctly works with Windows long paths, since Go 1.8).
Run with:

go run test-init.go
go run test-push.go

• What actually happened instead?

For init:

fatal: unable to access 'q:/Temp/Basil/AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA/config': Filename too long
2018/06/07 00:54:02 exit status 128

For push:

fatal: The remote end hung up unexpectedly
fatal: The remote end hung up unexpectedly
2018/06/07 00:47:24 exit status 128

[dscho]

Unfortunately, I do not have a working go, so this example is anything but minimal for me.

[dscho]

@grv87 If you could spend some quality time with https://github.com/git-for-windows/git/blob/v2.18.0.windows.1/t/t7416-submodule-long-path.sh to teach it the init and push reproducers in the form of test_expect_failure, that would go a looong way to help accelerating the inner edit-debug cycle that I would then use to fix this bug.

Every question about these test cases that is not covered in https://github.com/git-for-windows/git/wiki/Adding-regression-tests: I am happy to answer.