Igor Pechtchanski <pechtcha <at> cs.nyu.edu> writes: > > Because it's in a for loop, and when the first file fails but second > > succeeds, you still want the overall command to exit with failure. > > That's the correct intent, but shouldn't it be &&= instead of &=, > technically?
There's no such thing as &&=. And even if there was, you wouldn't want to use it, because it would short-circuit running cygcheck(). The whole point of the boolean collector is to run the test on every file, but to remember if any of the tests failed. Maybe thinking of a short-circuit in the reverse direction will help you understand: ok = cygcheck (argv[i]) && ok; But since ok is a simple boolean, short-circuiting doesn't save us any side effects, so we can use: ok = cygcheck (argv[i]) & ok; And since & is commutative, it has the same outcome as: ok = ok & cygcheck (argv[i]); Hence my shorthand (coreutils uses this idiom a lot, too): ok &= cygcheck (argv[i]); [By deMorgan's law, I could have also reversed the sense of the collector: bool failed = false; for (int i; ...) failed |= test_that_returns_true_on_failure(); return failed ? EXIT_FAILURE : EXIT_SUCCESS; But I hate thinking in negative logic, hence my definition of cygcheck to return true on success.] -- Eric Blake
