POST request falls back to GET after first successful render, causing 400/414 errors

[thausercelonis]

Environment:

• Extension version: 2.18.1
• OS: Windows 11
• PlantUML server: plantuml/plantuml-server:tomcat (Docker)

Steps to reproduce:

1. Configure extension with plantuml.render: "PlantUMLServer" and plantuml.server: "http://localhost:8080"
2. Open a large .puml file
3. Preview diagram - works correctly (first render)
4. Close preview, reopen preview - fails with 400 error
5. Server logs show "Request header is too large" indicating GET is being used instead of POST
6. Restart VS Code - first render works again, subsequent renders fail

Expected: Extension should consistently use POST for all renders
Actual: First render uses POST (works), subsequent renders use GET (fails for large diagrams)