"Error: No lowest priority node found (path: iie) at lxt" when having too little "maxInputTokens"

- VS Code Version: 1.128.0

Steps to Reproduce:

1. Configure your own chat language model in `chatLanguageModels.json` with a `maxInputTokens` value that is too small to fit the entire context (in this case it seems the list of the tools VSCode adds already is too long). Example:
```json
[
{
"name": "LM Studio",
"vendor": "customendpoint",
"apiType": "chat-completions",
"apiKey": "lm-studio",
"models": [
{
"id": "google/gemma-4-e4b",
"name": "google/gemma-4-e4b",
"url": "http://localhost:1234/v1/chat/completions",
"toolCalling": true,
"maxInputTokens": 8192,
"maxOutputTokens": 8192
}
]
}
]
```
2. Send a chat message.

An unhelpful error message is displayed:

> Sorry, your request failed. Please try again.
>
> Client Request Id: 8d45ea42-935b-46ba-8ea4-9fd6d404a5b3
>
> Reason: No lowest priority node found (path: iie): Error: No lowest priority node found (path: iie) at lxt

From the Dev Tools:
```
ERR [CHAT] extension request ERRORED in STREAM No lowest priority node found (path: iie): Error: No lowest priority node found (path: iie)
at lxt (c:\Users\xxluke\AppData\Local\Programs\Microsoft VS Code\fc3def6774\resources\app\extensions\copilot\dist\extension.js:173:4513)
at n.removeLowestPriorityChild (c:\Users\xxluke\AppData\Local\Programs\Microsoft VS Code\fc3def6774\resources\app\extensions\copilot\dist\extension.js:171:10945)
at cn._getFinalElementTree (c:\Users\xxluke\AppData\Local\Programs\Microsoft VS Code\fc3def6774\resources\app\extensions\copilot\dist\extension.js:174:30965)
at async cn.renderRaw (c:\Users\xxluke\AppData\Local\Programs\Microsoft VS Code\fc3def6774\resources\app\extensions\copilot\dist\extension.js:174:29509)
at async cn.render (c:\Users\xxluke\AppData\Local\Programs\Microsoft VS Code\fc3def6774\resources\app\extensions\copilot\dist\extension.js:174:29172)
at async cn.render (c:\Users\xxluke\AppData\Local\Programs\Microsoft VS Code\fc3def6774\resources\app\extensions\copilot\dist\extension.js:1666:17413)
at async $G._provideLanguageModelResponse (c:\Users\xxluke\AppData\Local\Programs\Microsoft VS Code\fc3def6774\resources\app\extensions\copilot\dist\extension.js:1690:11476)
at async $G.provideLanguageModelResponse (c:\Users\xxluke\AppData\Local\Programs\Microsoft VS Code\fc3def6774\resources\app\extensions\copilot\dist\extension.js:1690:15079) GitHub.copilot-chat 310932
logToConsole @ log.ts:117
error @ log.ts:564
error @ log.ts:669
error @ logService.ts:51
(anonymous) @ mainThreadLanguageModels.ts:249
await in (anonymous)
$tryStartChatRequest @ mainThreadLanguageModels.ts:252
await in $tryStartChatRequest
_doInvokeHandler @ rpcProtocol.ts:458
_invokeHandler @ rpcProtocol.ts:443
_receiveRequest @ rpcProtocol.ts:369
_receiveOneMessage @ rpcProtocol.ts:308
(anonymous) @ rpcProtocol.ts:161
_deliver @ event.ts:1364
fire @ event.ts:1395
fire @ ipc.net.ts:658
(anonymous) @ localProcessExtensionHost.ts:436
```

Expected behaviour:

Show an error message that there are more tokens than the model can process as input.