Fix NullPointerException in ThrowableStackTraceRenderer for non-idempotent getCause()

[UsmanEjaz10]

What & why

Rendering a Throwable walks its causal chain in two passes: a metadata pre-computation pass and a render pass. The render pass called Throwable#getCause() a second time after the metadata pass had keyed the first result in its IdentityHashMap. For a getCause() override that returns a different instance on every call (legal per the Throwable API contract), the second lookup misses, and rendering dereferences a null Metadata, throwing:

java.lang.NullPointerException: Cannot read field "stackLength" because "metadata" is null
    at org.apache.logging.log4j.core.pattern.ThrowableStackTraceRenderer.renderStackTraceElements(...)

Fix

Capture the cause once in Context.Metadata during the metadata pass and reuse metadata.cause at render time — in the base renderer (%ex, and %xEx which inherits it) and in the inverted renderer (%rEx). This is the same treatment the class already applies to getStackTrace() (#3940) and getSuppressed() (#3929).

ThrowableExtendedStackTraceRenderer is intentionally left alone: its independent JAR-enrichment walk null-checks and degrades gracefully, and it is a legitimate second getCause() caller — so %xEx invokes getCause() twice by design.

Tests

New NonIdempotentGetCauseTest asserts the fix across all three converters:

• %ex / %rEx: getCause() invoked exactly once per render.
• %xEx: invoked exactly twice (its own enrichment walk is the second, legitimate call).
• A non-idempotent getCause() renders its cause without throwing.

Kept as a standalone class rather than added to the shared AbstractStackTraceTest, which mutates a static EXCEPTION concurrently and makes adding test methods there unsafe w.r.t. execution order.

Verification

• ./mvnw verify (scoped to log4j-core, log4j-core-test and upstream modules): passes — RAT, Spotless, japicmp all green.
• Tests: 1,398 across the four affected test classes, 0 failures.
• Changelog entry added under src/changelog/.2.x.x/.

Fixes #4279

[ramanathan1504]

@UsmanEjaz10 Noted!. Give some time i will review it. Thanks for the contribution!