[ISSUE #6752] fix: encode SSE events as UTF-8

[yykaue]

Fixes #6752.

SseEventFormatter previously used the platform-default charset when converting SSE events and comments to bytes. On systems where the default charset is not UTF-8, non-ASCII MCP payloads could be corrupted.

Since Server-Sent Events are UTF-8 text streams, this PR:

• Uses StandardCharsets.UTF_8 when encoding SSE events.
• Uses StandardCharsets.UTF_8 when encoding SSE comments.
• Adds byte-level regression tests covering non-ASCII JSON event data and comments.

Verification

• The regression tests fail before the fix when executed with an ISO-8859-1 default charset.
• The same tests pass after the fix in a non-UTF-8 environment.
• Relevant tests, compilation, and Checkstyle pass.

@Aias00 Could you please review this PR?

[Aias00]

PR #6930 Review — [ISSUE #6752] fix: encode SSE events as UTF-8

Verdict: APPROVE

Scope

SseEventFormatter.formatEvent / formatComment (shenyu-plugin-mcp-server): encode SSE
payloads with StandardCharsets.UTF_8 instead of the platform-default charset.

Verification

• Two production lines changed (getBytes() → getBytes(StandardCharsets.UTF_8)); import java.nio.charset.StandardCharsets added. SSE is a UTF-8 text stream, so this is the correct
  encoding and removes a real corruption risk on ISO-8859-1 default-charset systems.
• Regression test SseEventFormatterTest added: byte-level assertArrayEquals for a non-ASCII
  JSON event (你好，世界) and a non-ASCII comment (保持连接), compared against the expected
  UTF-8 bytes. Directly reproduces the bug (fails under ISO-8859-1) and is fixed by the change.
• No API/signature change; mock-reactive MockServerWebExchange already on classpath.

Notes

• Minimal, correct, well-tested. Approving.