I use git bash for git related activities on windows.
I created a branch, did some changes and committed. When I tried to push, it gives me:
error "refpath does not exist".
I then checkout to other branch and retried to checkout to my branch but it says
error: pathspec 'feature/my-branch' did not match any file(s) known to git.
on running git branch, it lists my branch name as-
<U+0085><U+0085><U+0085><U+0086>feature/my-branch
I tried renaming this branch but that also didn't work.
git branch -m '<U+0085><U+0085><U+0085><U+0086>feature/my-branch' feature/new-branch
what is the reason behind this and what could be the possible solution?
The unwanted unicode chracters are converted to sequences such as <U+0085> or <U+0086> when they are printed on your terminal.
faulty=$(git branch | grep -a feature/my-branch) should hold the "correct" value for that branch name, so this should work:
# The '-a' option to grep is there to skip grep's auto detection of binary content
# it could kick in if there was a '\x00' byte in the input
#
# Your specific issue (with <U+0085> characters) shouldn't trigger it, I'm just
# mentioning that option for a more general scope
faulty=$(git branch | grep -a feature/my-branch)
# to avoid shenanigans with leading spaces and a possible '*' in the output:
faulty=$(git branch --format="%(refname:short)" | grep -a feature/my-branch)
git branch -m "$faulty" feature/my-branch
Otherwise:
use the $'...' syntax to have a string with the escape sequences interpreted:
git branch -m $'\u0085\u0085\u0085\u0086feature/my-branch' feature/my-branch
printf knows about \uXXXX sequencesyou can try to run :
faulty=$(printf "\u0085\u0085\u0085\u0086feature/my-branch")
# you can check if 'echo "$faulty"' gives you the same output as 'git branch'
git branch -m "$faulty" feature/my-branch