Three settings were only half-wired:
- autoPickByLanguage was saved but never read anywhere: delete it.
- showUnstableTail had a checkbox and a setter, but nothing ever called
the setter, so the checkbox did nothing: apply it in applyAppearance
and re-render from the cached text instead of blanking the caption.
- the language dropdown was hard-coded to auto/zh/en/ja while the registry
declares up to 16 languages per model: build it from the selected
model's own language list, with an ISO-code to display-name mapping
(plus a unit test covering every code in the registry). The caption
window's language pill now follows the setting too, instead of keeping
whatever it showed at startup.
windowPos is not dead -- it stores the dragged window position -- so it stays.
三个「半死」设置：autoPickByLanguage 存了但从没被读取，直接删掉；showUnstableTail
有复选框也有 setter，但从来没人调用 setter（复选框点了没用），改为在 applyAppearance
里生效并用缓存文本重画；「识别语言」下拉原来硬编码 自动/中/英/日，而注册表里每个模型
声明了最多 16 种语言，改为按所选模型的语言动态生成，并补 ISO 码→显示名映射与单元测试。
字幕窗的语言标签也跟着设置走了（原来只在启动时读一次）。windowPos 不是死设置（它存的
是拖动后的位置），保留。
Add an append-only streaming path that feeds each audio chunk as a new
conversation turn, so nothing is ever recomputed: the KV grows by
append, no prefix is re-decoded and no audio is re-encoded. Per-step
cost becomes flat instead of growing with the segment (RTF 1.06 -> 0.26
on 62.6s of continuous speech), but accuracy requires ~8s chunks because
the audio encoder needs its 8s attention window, so text only appears
every 8s. The incremental mode stays the default.
新增追加式流式：每块音频作为一个新的对话轮追加，前缀不重算、旧音频不重
编码，单步成本不再随段落增长（连续 62.6s 语音 RTF 从 1.06 降到 0.26）。
但准确率要求 chunk≈8s，否则会被截断，所以出字粒度是 8s；增量式仍为默认。
Move all business logic into liblivecaption_core (no Qt dependency) and
the widgets into liblivecaption_ui, leaving two executables: the caption
window and the CLI, which no longer links Qt Widgets. Abstract the ASR
backend behind IAsr so the streaming policy and the live pipeline can be
tested with a scripted fake backend instead of a real model.
把全部业务逻辑下沉到不依赖 Qt 的 liblivecaption_core，界面独立成
liblivecaption_ui，拆出悬浮窗与命令行两个可执行程序（CLI 不再链接 Qt
Widgets）；并用 IAsr 抽象推理后端，使流式策略与实时管线可脱离真模型
用假实现做单元测试。测试还抓到一个真 bug：段恰好结束在 chunk 边界时
只提交了稳定部分，最后几个词被丢掉，已修。
blumia