git diff unusually slow

[stefanobaghino]

• Your Windows build number: (Type ver at a Windows Command Prompt)
  Microsoft Windows [Version 10.0.16299.125]

• What you're doing and what's happening: (Copy&paste specific commands and their output, or include screen shots)
  git diff

• What's wrong / what should be happening instead:
  The command takes ~450 ms to execute / The execution should be nearly instantaneous

• Strace of the failing command, if applicable: (If some_command is failing, then run strace -o some_command.strace -f some_command some_args, and link the contents of some_command.strace in a gist here)
  N/A

When working on Ubuntu, I use a fairly complex command prompt that detects if my working directory is part of a git repository. If so, the prompt also adds information about the current branch and its status.

I noticed that there was an unusual latency (~1 second) between the end of a command and the display of the prompt, so I set -x to identify the bottleneck and realized that half of the time was spent running git diff, which I then confirmed by running the command with time (which consistently reports ~450 milliseconds).

If this can be useful, I'm currently running the Docker for Windows (version 17.12.0-ce-win45 (15017), channel edge, build 6bbf198) using the experimental option "Linux containers on Windows" and by running the same git diff on the same repository inside a docker container based on the ubuntu image, the times are slightly better (around ~350 milliseconds).

Furthermore, running git diff on the bash that is installed as part of Git for Windows yields even better times (~200 milliseconds).

On a virtual machine provisioned on OpenStack with comparable resources but running Ubuntu natively, the execution of git diff on the same repository is nearly instantaneous (~1 millisecond).

I can imagine this is due to the many levels of indirections that are involved in the whole WSL, but I was wondering if there's something in particular that causes this slowdown.

[therealkenc]

#981, #1305, #1566 et al

[stefanobaghino]

@therealkenc Sorry, I searched but couldn't find them. Thanks for reporting the original issue, I'll make sure to follow #981.