[ISSUE #8239] Fix the issue of potential message loss after a crash under synchronous disk flushing configuration.

[RongtongJin]

Which Issue(s) This PR Fixes

Fixes #8239

Brief Description

Under the condition of synchronous disk flushing, unflushed messages should be unacknowledged messages. Such messages should not be consumed prematurely, as this would lead to the consumption offset moving forward, resulting in potential phantom reads and message loss on the consumer side.

Therefore, in this PR, the dispatch offset of the ConsumeQueue is restricted to be before the disk flush offset to prevent consumers from experiencing phantom reads.

How Did You Test This Change?

Add a UT

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 40.00000% with 6 lines in your changes missing coverage. Please review.

Project coverage is 42.89%. Comparing base (7a5ea90) to head (4dfcd53).
Report is 9 commits behind head on develop.

Files	Patch %	Lines
...org/apache/rocketmq/store/DefaultMessageStore.java	20.00%	1 Missing and 3 partials ⚠️
...ache/rocketmq/store/config/MessageStoreConfig.java	50.00%	2 Missing ⚠️

Additional details and impacted files

@@              Coverage Diff              @@
##             develop    #8240      +/-   ##
=============================================
- Coverage      42.89%   42.89%   -0.01%     
- Complexity     10368    10381      +13     
=============================================
  Files           1270     1271       +1     
  Lines          88724    88827     +103     
  Branches       11406    11420      +14     
=============================================
+ Hits           38060    38103      +43     
- Misses         45956    46015      +59     
- Partials        4708     4709       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[guyinyou]

lgtm