fix(tokenizer): accept the OpenAI developer role (map to system for templates without it)

[gdevenyi]

What this fixes

A request whose system prompt uses the OpenAI developer role (the current spelling of system; pi sends it by default for openai-completions providers, compat.supportsDeveloperRole: true) fails with 400 could not encode request: Unexpected message role. on every model whose chat template does not spell the role, which is most of them (Qwen3.8-Flash-Next, Qwen3.x, GLM, ...). The template raises, the server relays it.

TokenizeManager._render now maps developer to system before apply_chat_template, unless the template mentions developer itself (gpt-oss renders the two roles differently and keeps its input). The mapping copies the message; the request is not mutated.

Testing

• tests/tokenizer/test_tokenize.py::test_developer_role_maps_to_system_unless_the_template_knows_it.
• Live on a Qwen3.8-Flash-Next TP=2 server: {"role": "developer", ...} went from the 400 above to the same prompt_tokens as the system spelling.

🤖 Generated with Claude Code

https://claude.ai/code/session_0173pf9k9fSVtwbm3f898HDt

[gdevenyi]

Re-tested against current main (fb7f732, i.e. after the #418 / #427 / #426 quantization refactor) on 2 x RTX 6000 Ada, TP=2 box.

Method. This PR's head merged onto main, then the full pytest tests suite. The run is CUDA-hidden (CUDA_VISIBLE_DEVICES="") on purpose: this box is serving a model on both GPUs, and with them visible 65-95 GPU tests fail on main itself with AcceleratorError: out of memory, with the count swinging ~10 between identical runs. Hiding CUDA makes the result deterministic, so a failure-set difference against main means something. Baseline: main = 1205 passed, 350 skipped, 0 failed.

Result: 1206 passed, 350 skipped, no new failures.

The +1 over main are this PR's own tests, and they ran (not skipped).

🤖 Generated with Claude Code

https://claude.ai/code/session_0173pf9k9fSVtwbm3f898HDt