#1960 v0.1.42: serve/server.py crash

Description

I updated using UPDATE.bat from 0.1.41 to 0.1.42 and started with START-HERE.bat, but am greeted with an exception:

```
PS C:\Tools\Strata\Strata> .\START-HERE.bat
Strata - Qwen3.8-Flash-Next on a normal PC (a GPU + system RAM + CPU)
  [ok] GPU: AMD Radeon RX 9070 XT (16 GB, AMD)

  ----------------------------------------------------------------------------------------------------
  Starting qwen3.8-flash-next-iq3_s: it loads about 55 GB into RAM and locks part of it for the GPU.
  While it does, YOUR PC CAN BE SLOW OR STOP RESPONDING FOR 1-3 MINUTES (longer the first time after a
  restart). That is normal: please wait and don't close this window - the browser opens when it is ready.
  Later, closing this window stops the model.
  ----------------------------------------------------------------------------------------------------
  Settings (strata-iq3_s.json): --expert-cache auto --prefill auto --spec 4 --spec-min-p 0.5
    --max-context 262144 --kv int8 --kv-resident 32768; server 127.0.0.1:8080, gpu 1
loading the model (the first start takes a minute or two) ...
[strata] starting the engine: reading the model's weights ...
Traceback (most recent call last):
  File "C:\Tools\Strata\Strata\serve\server.py", line 6329, in <module>
    sys.exit(main())
             ~~~~^^
  File "C:\Tools\Strata\Strata\serve\server.py", line 6139, in main
    engine = StrataEngine(exe, engine_args(cfg) + (effort_end or []), cwd=cfg.get("cwd"), log=cfg.get("log"),
                          env=env, lazy=lazy)
  File "C:\Tools\Strata\Strata\serve\server.py", line 712, in __init__
    raise EngineDied(why + (f" (see {log})" if log else "") +
                     start_failure_hint(log, log_start) + start_log_tail(log, log_start))
EngineDied: the engine exited before it was ready (see C:\Tools\Strata\Strata\strata-iq3_s.log)
the engine log's last lines:
  [strata] 2026-10-11 16:26:28 engine started: C:\Tools\Strata\Strata\engine\strata.exe --serve --pack C:\Tools\Strata\Strata-data\packs\iq3_s --native T:\LLMs\strata\models\IQ3_S\Qwen3.8-Flash-Next-GSQ-RCO-IQ3_S-00001-of-00002.gguf --ple-gguf T:\LLMS\strata\models\IQ3_S\Qwen3.8-Flash-Next-GSQ-RCO-IQ3_S-00002-of-00002.gguf --expert-profile C:\Tools\Strata\Strata\data\expert-profile.bin --expert-cache auto --prefill auto --spec 4 --spec-min-p 0.5 --mtp C:\Tools\Strata\Strata-data\mtp\rt --max-context 262144 --kv int8 --kv-resident 32768
  strata: gfx12 (RDNA4): 5 exact speed switches on by default (STRATA_GFX12_DEFAULTS=0 turns them off; a switch you set is kept): GDN_HEAD GDN_PP GDN_CONVL2 GDN_NOY CVEC_FUSE
  strata generate: memory guard: limit 125.5 GiB (source: meminfo), current 32.9 GiB, reclaimable cache credited 0.0 GiB; available 92.7 GiB (MemAvailable 92.7 GiB), floor 2560 MiB
  strata generate: PCIe probe: 54.2 GB/s host->device (best of 53.9 54.2 54.0 54.2) -> pcie_frac 0.55 (default 0.55)
  strata generate: native pack: C:\Tools\Strata\Strata-data\packs\iq3_s experts (largest blob 2.66 MB), token embedding IQ4_XS in mapped host memory (322 MiB, 0.3 s)
  strata generate: 1466 MiB of weights loaded from C:\Tools\Strata\Strata-data\packs\iq3_s in 0.3 s (302 canonical tensors skipped: served natively)
  strata generate: native dense projections: native dense: duplicate split shard number
```

My PC was just resumed from standby and no other LLM was running. 2,3/16GB vram usage. 33/126GB system ram usage.

I did move my model files to another drive and manually updated the json config file, but that is it for the changes I did (besides updating to 0.1.42).