Create a crash report context

We should have some kind of crash report context which will do something like output the callstack (if possible), the logs, etc so that we can more easily debug what is happening.

The issue names no files, tests, or entry points. Start by surveying the C++ engine for existing logging and platform-specific crash handling, then determine where a crash-report context could be integrated. Done should mean an agreed implementation that captures a callstack when possible and includes relevant logs for debugging.

Written by the indexing model from the issue text.