COMP: Remove UTF-8 BOM from source files

[Punzo]

#1213
https://discourse.slicer.org/t/failed-to-build-slicer-cause-the-ctk-build-exited-with-code-1/37872

@lassoan

[jcfr]

For now, the easiest is to keep using ASCII in all CTK source files. If there is a good reason to change this in the future then it should be changed for all source files.

To enforce this, is there a check we could add to https://github.com/commontk/CTK/blob/master/.pre-commit-config.yaml ?

[Punzo]

For now, the easiest is to keep using ASCII in all CTK source files. If there is a good reason to change this in the future then it should be changed for all source files.

To enforce this, is there a check we could add to https://github.com/commontk/CTK/blob/master/.pre-commit-config.yaml ?

I am not familiar with the pre-commit check infrastructure. But if possible we could run a script that checks the first 3 characters of a file, i.e. something like this:

# Read the first three bytes of the file
BOM=$(head -c 3 "$FILE" | xxd -p -c 3)

# Check if the BOM matches the UTF-8 BOM (EFBBBF)
if [ "$BOM" = "efbbbf" ]; then
  echo "The file '$FILE' contains a UTF-8 BOM."
else
  echo "The file '$FILE' does not contain a UTF-8 BOM."
fi

[lassoan]

Any non-ASCII character (code > 127) might cause trouble, such as accented characters. Also, we should only check source code files.

I've got Copilot generate this code, does this look good?

# Get the list of staged source files
source_files=$(git diff --cached --name-only -- '*.cpp' '*.h' '*.in' '*.cmake' 'CMakeLists.txt' '*.py')

# Check for non-ASCII characters in the staged source files
non_ascii_files=$(echo "$source_files" | xargs grep -P -n "[^\x00-\x7F]")

if [ -n "$non_ascii_files" ]; then
  echo "Error: The following source files contain non-ASCII characters:"
  echo "$non_ascii_files"
  exit 1
fi