Add update_data to Span.

[antonpirker]

This allows users to set multiple span.data attributes at once.
In 3.x this then will become set_attributes (plural).

[szokeasaurusrex]

This API is a bit confusing, imo, as it is unclear what should happen if a user calls set_data with a dict as the first parameter, and something else as the second parameter.

I would prefer we instead introduce a separate function to enable setting data as a dict, rather than supporting two distinct signatures for set_data.

Although if others are okay with having set_data support both ways, we can go with it – just sharing my two cents here. In that case, though, I would define both APIs with @overload.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 85.05%. Comparing base (84adbb7) to head (80fe601).
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #4666      +/-   ##
==========================================
- Coverage   85.09%   85.05%   -0.04%     
==========================================
  Files         156      156              
  Lines       15797    15799       +2     
  Branches     2671     2671              
==========================================
- Hits        13442    13438       -4     
- Misses       1578     1585       +7     
+ Partials      777      776       -1     

Files with missing lines	Coverage Δ
sentry_sdk/tracing.py	86.52% <100.00%> (+0.05%)	⬆️

... and 4 files with indirect coverage changes

[antonpirker]

True @szokeasaurusrex i have now changed it to have a update_data() function instead. Similar to pythons own dict.update().

[cursor]

Bug: NoOpSpan Method Inheritance Issue

The update_data method, newly added to the Span class, is not overridden in NoOpSpan. This allows NoOpSpan.update_data() to inherit the real implementation and modify internal data, breaking its no-operation contract. This is inconsistent with NoOpSpan.set_data() and similar methods, which are correctly overridden as no-ops.

[szokeasaurusrex]

I would also add update_data to the name of the test. And, perhaps add a test which only checks update_data in isolation

[szokeasaurusrex]

I think the Cursor bug bot might have found something. Besides that, looks good