Improve L1 message reorg testing

[jonastheis]

This PR mainly does the following:

• refactors can_handle_reorgs_while_sequencing into can_handle_l1_message_reorg to test not only sequencing in the presence of L1 reorg but also behavior of a follower node. Made the test much more explicit to assert certain node states
• adds a lot of new convenience functionality for e2e tests for waiting for specific events, assert node state, etc
• adjusts default_sequencer_test_scroll_rollup_node_config so that by default the sequencer in tests does not issue blocks automatically anymore. This is so that tests become more predictable.
• fixes issue where an L1 reorg was not properly handled within the engine driver, leading to no L2 reorg taking place even if an L1 message was reorged. (this was discovered by refactoring the can_handle_reorgs_while_sequencing into can_handle_l1_message_reorg and making sure that node states are properly tested).