[Emryswu]

Fixes #7452

Problem

CachePluginDataHandler.handlerRule() skips a rule whose handle is null, so no CacheRuleHandle is cached for it. When such a rule matches, CachePlugin still wraps the response in CacheHttpResponse with a null handle, and cacheResponse() throws a NullPointerException on cacheRuleHandle.getTimeoutSeconds() while writing an otherwise successful downstream response.

Fix

CachePlugin.doExecute() now looks up the rule handle first. If none is cached, it passes the request to the rest of the chain without reading or writing the cache. This is how the Hystrix, Resilience4J, Sentinel and RateLimiter plugins handle a missing handle since #6877. No default timeout is applied. Rules with a handle behave as before. The handle is now looked up once instead of in both branches.

Tests

Added two tests to CachePluginTest. In both, the downstream chain writes a real response body:

• testDoExecuteSkipsCacheWhenRuleHandleIsNull: a rule with no cached handle completes without error, returns the downstream body, and writes nothing to the cache. Without the fix, this test fails with the NPE from the issue.
• testDoExecuteCachesResponseWhenRuleHandleIsPresent: a rule with a handle still returns the body and caches it.

mvn -pl shenyu-plugin/shenyu-plugin-cache/shenyu-plugin-cache-handler -am test passes locally, checkstyle included.

Make sure that:

• You have read the contribution guidelines.
• You submit test cases (unit or integration tests) that back your changes.
• Your local test passed ./mvnw clean install -Dmaven.javadoc.skip=true.

[Aias00]

Review: skip caching when cache rule handle is null

Clean, minimal fix. CachePlugin.doExecute now builds the rule handle once up front and short-circuits to chain.execute(exchange) when the cache rule handle is null, instead of passing a null CacheRuleHandle into CacheHttpResponse (which would NPE / misbehave). It also removes the now-duplicated buildRuleHandle(rule) calls that previously appeared in two branches.

Verification

• buildRuleHandle returns CACHED_HANDLE.get().obtainHandle(CacheKeyUtils.INST.getKey(rule)), which is null when the rule has no configured handle — so the original code could pass null into CacheHttpResponse. The new null guard is the correct fix.
• java.util.Objects was already imported (used elsewhere in the file), so the added Objects.isNull(...) compiles.
• Tests added for both paths: testDoExecuteSkipsCacheWhenRuleHandleIsNull (asserts body flows through unchanged and nothing is written to the cache) and testDoExecuteCachesResponseWhenRuleHandleIsPresent (asserts the response is cached).

CI note
No checks are reported on the fork branch fix/issue-7452 — the fork hasn't triggered CI, so there is no CI evidence for this change. The modification is single-module (shenyu-plugin-cache-handler), self-contained, and well-tested, so risk is low, but please confirm pr_build is green before merging.

Approving — the fix is correct and the added tests cover the regression.

[Aias00]

if you are using wechat, pls add mine: aias00