fix(cli): read kernel metadata and source files using UTF-8 in kernels_push

[sridipbasu]

Summary

This PR updates kernels_push() to explicitly read the kernel metadata and source files using UTF-8 instead of relying on the platform's default encoding.

Problem

On Windows systems where the default text encoding is not UTF-8, running:

kaggle kernels push -p .

can fail with an error similar to:

'charmap' codec can't decode byte 0x81...

This happens because kernel-metadata.json and the kernel source file are opened without specifying an encoding. If either file contains Unicode characters (for example emojis, accented characters, or non-English text), reading the file may fail depending on the system's default encoding.

Solution

This PR adds encoding="utf-8" to the two file reads in kernels_push():

• kernel-metadata.json
• Kernel source file (.py, .ipynb, etc.)

This makes file reading consistent across platforms while preserving the existing behavior.

Tests

Added a regression test that:

• Creates a temporary kernel directory.
• Writes UTF-8 encoded metadata and source files containing Unicode characters.
• Verifies that kernels_push() reads them successfully.
• Confirms the test fails with the original implementation and passes with this change.

Why this change is safe

• JSON metadata files are expected to be UTF-8.
• Python source files and Jupyter notebooks are also expected to use UTF-8.
• The change only affects two open() calls inside kernels_push().
• No other commands or file operations are modified.

[stevemessick]

/gcbrun