A Practical Approach to Error Handling
summary
think‑cell describes a 20‑year, practical error strategy: wrap every OS/API call, collect memory dumps and send them to a server that auto‑opens them in a debugger, reproduce a failure before adding handlers, and prefer fixes that silently work for users. A surprising stat: about 5% of handlers fix 95% of errors. A concrete win: register MSXML and retry eliminated many customer failures. Uses p0323/expected mention, MDL compression to correlate DLL versions, and keeps asserts nonfatal to avoid silencing diagnostics.
AI-written summary · about
chapters (15)
- 0:00Introduction
- 1:53Where are errors coming from
- 5:43Output exceptions
- 9:45Exception safe
- 13:01Handling unexpected exceptions
- 15:53Gathering information
- 16:42Carry on
- 22:00Critical Errors
- 24:09Untested Errors
- 28:02Bad User Experience
- 29:21Special Environment
- 31:12Reporting
- 32:52Problem identification
- 37:30How does it work
- 41:22C20 contracts
related talks
44mA Practical Approach to Error Handling
45mTo Err is Human - Robust Error Handling in C++26
43mLearning to Live with Errors
57mYou’re Doing Exceptions Wrong
57mLessons from 20 years of Debugging C++ Code
45mThe design of error
44mA Guide to Debugging C++ Code: Practical and Interactive Examples
04mErrors: forgotten, but not gone