"If you listen carefully you can hear dozens of Linux kernel developers collectively holding their breath and thinking 'Maybe Linus will finally merge kgdb'. Yes, user bug reports are important. Developer efficiency is important too."
"If you listen carefully you can hear dozens of little kids at the kindergarden collectively holding their breath and thinking 'Maybe Linus will finally recruit kernel developers from kindergarden'.
Debuggers are for those who cant code right.
When Linux compiles something that he have coded, if it generates warnings, its a bug in the compiler.
If hes code would ever segfault, he formats the disk.
I'm sure Andrew Morton really enjoys his days in kindergarten, too...
Sure, if you have to single-step your for loops because they iterate 1 too many times, kernel programming may not be your thing. But debuggers are useful for much more than that.
There. I did it. I fed the troll. :-P
Program Intellivision and play Space Patrol!
I take your point, but are you suggesting that there is never an off-by-one error in the kernel ?
Of course not. There've been fencepost errors in the past and there will be again. There will always be brown-paper-bag bugs. If you need to single step in a debugger to find it, though, I suggest you might be in over your head if you're trying to do kernel development.
I can't remember the last time I single-stepped a C program, or even fired up a debugger for that matter. I'm sure I've had plenty of trivial coding mistakes in the last year though, and I somehow managed to find and fix them all nonetheless. I don't think I'm particularly exceptional in this regard, either. I look in awe at the stuff the kernel developers manage to pull off—quickly, no less!—and I realize I'm nowhere near their league.
You can suggest anything you want, it does not make it right.
The last two paragraphs should have told you that GP was a (feeble) attempt at a joke.