Debugging C++ Code: A Practical and Interactive Guide - Sebastian Theophil - CppNow 2023

[@FalcoGer]

56:08 if your test with random data fails, then you should definitely print out the seed used to create the random data. A good idea would be to use special cases (0, 1, -1, int_min, int_max, int_min + 1, int_max - 1, and whatever might be special for your case) and an assortment of random inputs.

[@about2mount]

Write a Python program to read through your files to check and or count for mistakes in your C++ code. In fact many are doing this already including myself. I use one to simply read and count every open tag as ( and { and [ and count them then count all the closing tags as ) and } and ] then check for even numbers as pairs. If i get an odd number I throw a line number and what type of tag is missing.

[@alanwest6949]

The bugs we write are reflections of unresolved emotional issue in the soul. Those who have (debugged) healed their soul of all unresolved emotional issues, would write all correct code.

[@StefaNoneD]

I'm pretty senstive to noise. There is a lot of background noise to hear. I think the microphone was not near enough to the mouth and the gain too high. I hope, presenters are more sensitive to that in future. Another thing is: I don't see where the presenter is pointing to with his finger. So maybe he should just use the mouse cursor.

[@retropaganda8442]

26:57 You also need to be aware that the same physical shared memory may be mapped at different virtual addresses in different processes (or in the same process if you map it several times), so you can't store any pointer in there.

[@victotronics]

It seems that lectures about debugging use 80 percent examples from threading, and the remainder examples from using "star pointers". So if you don't thread, and write modern C++ you will not have bugs?