How to fix error: invalid path 'CON' (reserved Windows name) in Git

First, look at the error message to identify which files are causing the issue:

error: invalid path 'path/to/con.h'
fatal: unable to checkout working tree
warning: Clone succeeded, but checkout failed.

Common reserved names that cause this error:
- CON - Console device
- PRN - Printer device
- AUX - Auxiliary device
- NUL - Null device
- COM1 through COM9 - Serial ports
- LPT1 through LPT9 - Parallel ports

Note that con.h, aux.txt, nul.conf etc. are also invalid because the base name is reserved.

The cleanest solution is to rename the offending files from a Linux, macOS, or WSL environment:

# Clone on Linux/macOS/WSL where the files are valid
git clone https://github.com/user/repo.git
cd repo

# Find reserved filenames
find . -iname "con.*" -o -iname "prn.*" -o -iname "aux.*" -o -iname "nul.*" \
  -o -iname "com[1-9].*" -o -iname "lpt[1-9].*"

# Rename the problematic files
git mv path/to/con.h path/to/console.h
git mv path/to/aux.txt path/to/auxiliary.txt

# Commit the changes
git commit -m "Rename Windows-reserved filenames for cross-platform compatibility"
git push

After pushing, Windows users can clone successfully.

If you're on Windows, use Windows Subsystem for Linux (WSL) to clone and work with the repository:

# Open WSL terminal
wsl

# Clone inside the WSL filesystem (not /mnt/c/)
cd ~
git clone https://github.com/user/repo.git

# Work with the files normally in WSL
cd repo

Important: Clone into the native WSL filesystem (~/), not into /mnt/c/ which maps to Windows and has the same restrictions.

You can disable Git's NTFS protection to allow checkout, but the problematic files will be skipped or created with zero bytes:

# Clone with NTFS protection disabled
git clone -c core.protectNTFS=false https://github.com/user/repo.git

# Or for an existing repo
git config core.protectNTFS false
git checkout

Security Warning: This setting was enabled by default to protect against CVE-2019-1353. Only use this temporarily and avoid setting it globally (--global). Files with reserved names will still be problematic - they may appear as zero-byte files or be skipped entirely.

If you don't need the files with reserved names, use sparse checkout to exclude them:

# Clone without checkout
git clone --no-checkout https://github.com/user/repo.git
cd repo

# Initialize sparse checkout
git sparse-checkout init --cone

# Set directories to include (exclude the problematic path)
git sparse-checkout set src/main src/tests docs

# Or exclude specific patterns
git sparse-checkout set '/*' '!path/to/problematic/'

# Now checkout
git checkout

This lets you work with the repository while avoiding the incompatible files.

If you only need to read the files and don't need Git history, download the repository as a ZIP:

1. Go to the repository on GitHub/GitLab/Bitbucket
2. Click Code > Download ZIP
3. Extract the archive (Windows may skip or warn about reserved filenames)

This bypasses Git entirely but loses version control capabilities. Use this only for quick access to code you need to read but not modify.

If files were created with zero bytes due to the NTFS protection workaround, recover the content:

# Check if the file in Git has content
git show HEAD:path/to/con.h

# Create a patch with the content
git diff --binary HEAD~1 HEAD -- path/to/con.h > fix.patch

# Manually extract the content (on a non-Windows system)
# Then rename and commit as shown in step 2

The original content is safely stored in Git's object database; only the working directory checkout fails.