Fixes #102

ProcessJsonRpcClient::readProcessOutput() read newly produced bytes from the Node bridge process with Process::getIncrementalOutput() / getIncrementalErrorOutput(), but never called the matching Process::clearOutput() / clearErrorOutput(). Symfony's Process keeps the entire stdout/stderr history in an internal buffer (a php://temp stream that spills to a disk-backed temp file past a small in-memory threshold) for as long as the process object is alive — the incremental getters only track a read cursor, they do not discard what was already consumed.

On a long-lived session (a multi-hour Behat regression run, for example) every JSON-RPC response — page HTML, base64 screenshots, ARIA snapshots — accumulated forever in that buffer. Once the temp file grew large enough to fill the available disk space, Process::addOutput() started silently failing to write further bytes, and every subsequent request in waitForResponse() spun until its deadline and failed with the generic JSON-RPC request N timed out — indistinguishable from a slow browser action or a genuine hang.

Fixes it by calling clearOutput()/clearErrorOutput() right after each incremental read, so the Process buffers stay empty; nothing else reads the accumulated (non-incremental) output.

Testing

• Added two unit tests on ProcessJsonRpcClient (testReadProcessOutputClearsBothProcessBuffers, testReadProcessOutputClearsBuffersAcrossMultiplePolls) asserting clearOutput()/clearErrorOutput() are called once per poll — verified both fail against the pre-fix code (0 calls observed) and pass with the fix.
• Full unit suite: 798/798 passing (2 pre-existing, unrelated deprecation warnings elsewhere).
• php-cs-fixer --dry-run clean; phpstan analyse (level 10) reports 0 errors on the changed files.

Discovered while running the OroCommerce Behat regression suite on this driver: several nodes started failing late in the run with unexplained JSON-RPC request N timed out errors, traced back to full CI-node disks.