Troubleshooting

This page covers the most common failures and how to fix them.

Snapshot file is missing

You are in replay mode (default) and the snapshot file does not exist.

Fix:

• Record snapshots locally and commit them:

  pytest --r-snapshot=record
  

• Double-check name= and ext=... match the recorded snapshot filename.

• If you configured r_snapshot_dir, confirm the snapshot root is correct.

R chunk '...' not found

The snapshot name you requested does not match any labelled R chunk in the current test file.

Fix:

• Ensure the chunk is labelled and the label matches name=....
• If you use @pytest.mark.r_snapshot("name"), the plugin will fail early with a list of available labels.

Missing R chunk label

Every fenced R chunk must have a label.

Accepted examples:

```{r, my_label}
...
```

```{r label}
...
```

Fix: Add a label to the chunk header.

Duplicate R chunk label

Labels must be unique per Python file.

Fix: Rename one of the conflicting chunks.

Rscript executable not found

Recording requires Rscript to be available.

Fix:

• Install R, or configure the Rscript path:

  pytest --r-snapshot=record --r-snapshot-rscript=/path/to/Rscript
  

Or set r_snapshot_rscript in pyproject.toml.

R execution failed

The R process returned a non-zero exit code. The failure message includes stdout/stderr.

Fix:

• Run in record mode locally to reproduce and inspect the full error:

  pytest --r-snapshot=record -q
  

• Verify required R packages are installed and discoverable (for example, set R_LIBS_USER via --r-snapshot-env).

Encoding issues

Snapshots are stored as UTF-8 by default.

Fix: If you must use a different encoding, set r_snapshot_encoding / --r-snapshot-encoding.