Optimize script interpreter testing by bypassing state root computation

[liuchengxu]

Currently, I run the Subcoin node with the import-blocks command to test our script interpreter in #89. However, block execution time significantly slows down (~1 block per second when the height exceeds 500,000 or earlier) due to state root computation during block finalization. This makes interpreter testing excessively time-consuming. To speed up the process, we could introduce an option to bypass state root computation specifically for testing purposes.