Application ↔ OpenGL Driver: Core-Profile Minimum vs Runtime Capability Detection

The goal is to achieve predictable rendering behavior when the same OpenGL code runs on a local developer workstation and in a production environment that may use a different driver stack (e.g., Mesa/llvmpipe or a headless EGL context).

Constraints arise because local GPUs often provide a compatibility profile and a permissive shader compiler, while production drivers may expose only a core profile, report different extension sets, or enforce stricter GLSL rules. These variations can cause context creation failures, missing function pointers, or silent rendering differences without clear error signals.

What minimum OpenGL core‑profile version should be required to guarantee baseline functionality across all target drivers? How should the application detect and handle missing extensions or function pointers when a stricter core profile is used? Is it preferable to fail fast when the requested profile is unavailable, or to implement fallback paths that attempt lower‑profile contexts with degraded features?