Effective Debugging

66 Specific Ways to Debug Software and Systems

[Bob Belderbos @bbelderbos]

Ever wondered how to debug a hanging test in #Python?

pip install pytest-timeout
pytest --timeout=N_SECONDS

To dump the whole stack trace you can add `--timeout_method=thread` as well.

Here is an example:

[sush @2sush]

what's your tactics ?

[Miguel de Icaza ᯅ🍉 @migueldeicaza]

Just had a quite emotional, personal conversation with the debugger. I tried to inspect the stack, the heap and the fault address.

Interestingly I felt heard & warm. Never tried therapy before but this is probably it? Try it especially if you usually just use it to step over.

[Effective Debugging @EffDbg]

The drgn C debugger allows you to express complex data structure traversals and conditions in Python. https://t.co/tMBO28M03n

[Effective Debugging @EffDbg]

This is the Golden Bug, the trophy of the first Competitive Debugging event. Details: https://t.co/OcUdzVqs1C

[🔎Julia Evans🔍 @b0rk]

debugging strategy: rule things out

[Effective Debugging @EffDbg]

The Arduino IDE serial plotter is an excellent debugging tool for real-time embedded applications.

[Diomidis Spinellis @CoolSWEng]

For over a decade the git grep command on macOS handled incorrectly regular expressions with UTF-8 characters. Here is how I debugged and fixed the problem. The fix will be released with Git 2.39. https://t.co/ixkq9YgRq9

[Effective Debugging @EffDbg]

@zkat__ @gesa • Add unit tests
• Add assertions
• Refactor the code
• Consider post-portem debugging
• Consider attaching to the process, once the bug manifests itself

In general, all bugs are tractable given sufficient effort.

[Effective Debugging @EffDbg]

@zkat__ @gesa • Add logging statements.
• If these also hide the bug, log into memory, and then dump the in-memory logs
• If you're debugging a non-optimized debug build (e.g. -O0), debug an optimized production build (e.g. -Ox)
• Exhaustively review/lint the code associated with the bug.

[HN Bot 🤖 @TechTweetBot]

Pleasant Debugging with GDB and DDD
https://t.co/Eeqn5RW8HX

[Effective Debugging @EffDbg]

Comparing the extended logs of a failing implementation (here in Perl) with those of a working one (bash) reveals the reason behind the failure.

[Cory Gabrielsen @corygabrielsen]

Logging always you to see the full state history. Noisy logs + grep is powerful,

Debugging with breakpoints is ephemeral and relies on a lot of short-term memory. Easy to get interrupted or miss something and have to do the whole sequence again.

[Diomidis Spinellis @CoolSWEng]

"Rather than alchemy, methodical troubleshooting", or "Finding a needle in a haystack by running git bisect on synthetic commits" https://t.co/ClRP9Tf6g3

[Effective Debugging @EffDbg]

@sysarmy Web searching is item #2 in the Effective Debugging book. The restart option, often works, but is in most cases a temporary solution. It is therefore unsuitable for production systems, because it leads to more toil.

[🔎Julia Evans🔍 @b0rk]

debugging tips: guesses are often wrong https://t.co/bC7y9Ehehl

[Effective Debugging @EffDbg]

You know you're in trouble when you search for an error message and, rather than obtaining StackOverflow results, all you get are still-open GitHub issues and a Reddit page.
[ Image credit @ThePracticalDev ]

[Effective Debugging @EffDbg]

@coderadda So true! See the "Effective Debugging" book, Chapter 2 (General-Purpose Methods an Practices), Item 9 (Set Yourself Up for Debugging Success), p. 24: "Sleep in a difficult problem".

[Effective Debugging @EffDbg]

Debugging and reverse engineering in one: "How I cut GTA Online loading times by 70%" https://t.co/3WCcH52zUL