server: cancel resumable streams before first token

[NitayRabi]

Summary

• poll the stream-aware stop predicate while waiting for the first streaming result
• accept explicit resumable-stream cancellation in the adjacent assertion
• add a router regression test that stops after session creation but before the first result

Fixes #27481.

Why

Requests with X-Conversation-Id deliberately survive physical socket disconnects. The pre-first-result wait nevertheless used only req.should_stop, so DELETE /v1/stream could not cancel tokenization/prompt processing. Later streaming results already use server_res_spipe::should_stop().

Testing

• python -m pytest -q tools/server/tests/unit/test_stream.py — 4 passed
• Release/Vulkan llama-server build
• live router reproduction with a long prompt: DELETE posts SERVER_TASK_TYPE_CANCEL; no generation follows the prompt batch

[ggml-gh-bot]

Hi @NitayRabi, thanks for your contribution!

Per our contribution guidelines, the automated PR checker found the following issue(s) that need your attention:

• PR Template not respected: Please respect the template when creating a new pull request. Make sure to fill out all required sections.

• AI-generated content: While code is allowed to be generated by AI, please write the PR description and commit messages on your own without the help of AI.

---

Please note that maintainers reserve the right to make final decisions on PRs. If you believe there is a mistake, please comment below.