fix(#5626): handle multi-text content better for openai format models

[alexhancock]

We weren't properly handling when messages have multiple text chunks in providers/formats/openai.rs

This surfaced in #5626 when encountering a message with content like this:

"content": [
  {
    "type": "text",
    "text": "--- Resource: file:///Users/pjv/development/projects/test%20python%20project/test2.py ---\nfrom typing import NamedTuple\n\n\nclass Person(NamedTuple):\n  name: str\n  age: int | None = None\n\n\nperson = Person(\"Alice\", 25)\n\n---\n"
  },
  {
    "type": "text",
    "text": " what does this code do?"
  }
],

Previously when a message contained multiple text content chunks, we would overwrite the earlier ones and replace them with the last one. So this message what does this code do? would lead to an answer like what code?

Example

Before:

After:

[Copilot]

Pull request overview

This PR fixes a bug where messages containing multiple text content blocks were incorrectly handled in the OpenAI format provider. Previously, when multiple text chunks were present in a single message, only the last chunk would be retained. The fix accumulates all text parts and joins them with newlines, ensuring all content is preserved.

Key Changes:

• Accumulate text blocks in a text_parts vector instead of overwriting converted["content"]
• Join accumulated text blocks with newlines after processing all content
• Add test coverage for multiple text block scenarios

[Copilot]

When an image path is detected and successfully loaded (lines 72-77), this code directly assigns to converted["content"], which will overwrite any previously accumulated text_parts. This breaks the fix for multiple text content chunks.

If a message has multiple text blocks and one contains an image path, only the text with the image will be included, losing the other text blocks. Consider accumulating the text in text_parts and handling the image separately, or restructuring to build the content array incrementally.

[alexhancock]

Closing in favor of #5839 which had some similar code to fix images in acp