PR summary

PercentFormatter.__call__ always reads self.axis.get_view_interval(), but that display range is only used to choose the number of decimals automatically, i.e. when decimals is None. With decimals set explicitly, calling the formatter while it is not attached to an axis raised AttributeError, even though the axis value is never needed.

• What problem does it solve? A PercentFormatter created with explicit decimals already has everything it needs to format a value, but it still required an axis, so using it standalone (for example to build a label or annotation) crashed.
• Reasoning for the implementation: the axis view interval is only consulted in the auto-decimals branch of format_pct, so __call__ now reads it only when decimals is None. Auto-decimals still requires an axis, unchanged.

Minimal self-contained example:

import matplotlib.ticker as mticker

# Before: AttributeError: 'NoneType' object has no attribute 'get_view_interval'
# After:  '50.0%'
print(mticker.PercentFormatter(xmax=1.0, decimals=1)(0.5))

PR checklist

• [N/A] "closes #0000" is in the body of the PR description - no issue; small self-contained bug fix (happy to open one if preferred)
• new and changed code is tested - added TestPercentFormatter.test_call_without_axis; existing format_pct cases are unchanged and still pass
• [N/A] Plotting related features are demonstrated in an example - bug fix, no new plotting feature
• [N/A] New Features and API Changes are noted with a directive and release note - bug fix, not a new feature or API change
• Documentation complies with general and docstring guidelines - no documentation changes; existing docstrings unaffected