fix: Don't finish span prematurely on read(0)

Drop the `not rv` check from the read() wrapper's span-finishing
condition. read(0) legitimately returns b"" without consuming the body,
which would falsely trigger span completion. The fp and closed checks
are sufficient to detect when the body is fully consumed.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: sentrivana

sentry_sdk/integrations/stdlib.py

@@ -214,15 +214,13 @@ def read(self: "HTTPResponse", *args: "Any", **kwargs: "Any") -> "Any":
         if span is None:
             return real_read(self, *args, **kwargs)
 
-        rv = None
         try:
-            rv = real_read(self, *args, **kwargs)
-            return rv
+            return real_read(self, *args, **kwargs)
         finally:
             # read() might be called multiple times to consume a single body,
             # so we can't just end the span when read() is done. Instead,
             # try to figure out whether the response body has been fully read.
-            if self.fp is None or self.closed or not rv:
+            if self.fp is None or self.closed:
                 self._sentrysdk_span = None  # type: ignore[attr-defined]
                 _finish_span(span)