bugfix: fix throw NPE when get the state list

[wangliang181230]

Ⅰ. Describe what this PR did

optimize: opt SAGA CompensationTriggerStateHandler
此处的stateMachineInstance必定存在，无需判断空值。
而且else处代码逻辑也有问题，即使进入else了，也是必定NPE。

Ⅱ. Does this pull request fix one issue?

fixes #2925

Ⅲ. Why don't you add test cases (unit test/integration test)?

Ⅳ. Describe how to verify it

Ⅴ. Special notes for reviews

[wangliang181230]

@long187 ，麻烦review一下。

[codecov-commenter]

Codecov Report

Merging #2949 into develop will increase coverage by 0.01%.
The diff coverage is n/a.

@@              Coverage Diff              @@
##             develop    #2949      +/-   ##
=============================================
+ Coverage      50.57%   50.58%   +0.01%     
+ Complexity      3105     3104       -1     
=============================================
  Files            599      599              
  Lines          19510    19510              
  Branches        2406     2407       +1     
=============================================
+ Hits            9868     9870       +2     
+ Misses          8650     8649       -1     
+ Partials         992      991       -1     

Impacted Files	Coverage Δ	Complexity Δ
...torage/file/store/FileTransactionStoreManager.java	57.41% <0.00%> (+0.64%)	29.00% <0.00%> (+1.00%)

[codecov-io]

Codecov Report

Merging #2949 into develop will decrease coverage by 0.36%.
The diff coverage is n/a.

@@              Coverage Diff              @@
##             develop    #2949      +/-   ##
=============================================
- Coverage      50.93%   50.57%   -0.37%     
+ Complexity      3221     3162      -59     
=============================================
  Files            605      598       -7     
  Lines          19887    19702     -185     
  Branches        2481     2442      -39     
=============================================
- Hits           10130     9964     -166     
- Misses          8743     8749       +6     
+ Partials        1014      989      -25     

Impacted Files	Coverage Δ	Complexity Δ
...java/io/seata/server/session/SessionCondition.java	46.66% <0.00%> (-36.67%)	7.00% <0.00%> (-6.00%)
...rage/redis/store/RedisTransactionStoreManager.java	36.36% <0.00%> (-28.02%)	20.00% <0.00%> (-18.00%)
.../src/main/java/io/seata/core/rpc/ShutdownHook.java	0.00% <0.00%> (ø)	0.00% <0.00%> (ø%)
.../java/io/seata/core/store/GlobalTransactionDO.java	0.00% <0.00%> (ø)	0.00% <0.00%> (ø%)
...ver/storage/redis/session/RedisSessionManager.java	29.31% <0.00%> (ø)	8.00% <0.00%> (-2.00%)
...ain/java/io/seata/core/rpc/hook/StatusRpcHook.java
...java/io/seata/server/storage/SessionConverter.java
...n/src/main/java/io/seata/common/rpc/RpcStatus.java
...java/io/seata/common/exception/RedisException.java
.../discovery/loadbalance/LeastActiveLoadBalance.java
... and 6 more

[long187]

这段逻辑主要是怕从statemachineInstance里拿到的stateList为空，因为要恢复时可能只查询了statemachineInstance对象没有级联查询stateInstance列表，所以这里做了一个双保险，如果stateList为空，则查询一次。所以这里我建议在你改的基础上再判断一下stateList是不是为空，如果为空则用之前else里那段代码查询一次。

[wangliang181230]

这段逻辑主要是怕从statemachineInstance里拿到的stateList为空，因为要恢复时可能只查询了statemachineInstance对象没有级联查询stateInstance列表，所以这里做了一个双保险，如果stateList为空，则查询一次。所以这里我建议在你改的基础上再判断一下stateList是不是为空，如果为空则用之前else里那段代码查询一次。

done

[long187]

LGTM

[jsbxyyx]

LGTM