Protractor test run does not abort after first failure when failFast is enabled with directConnect

When using Protractor’s failFast option to halt execution after the first spec failure, teams expect the runner to stop immediately and skip remaining specs.

However, the interaction between failFast (set via configuration or the --failFast CLI flag) and the directConnect mode is not fully documented for recent releases, leaving uncertainty about whether the flag is honored when directConnect:true is used.

The goal is to verify if the failFast behavior is reliable under directConnect in current Protractor versions, considering version‑specific fixes and flag precedence rules.

• Does the --failFast flag cause Protractor to abort after the first failure when directConnect:true is set in versions ≥5.4.0?
• If the flag is overridden at runtime, does it affect only specs that have not yet started, or does it also interrupt a currently executing spec?
• Are there any version‑specific conditions (e.g., using multiCapabilities or shardTestFiles) that change this behavior?