Re: [libvirt PATCH 2/4] docs: coding-style: Rewrite section on shortening comparisons

The code style showed `bool hasFoos; if (hasFoos == true)` as a good example in one place, only to warn against comparisons with `true` a couple of paragraphs further down. Merge this advice on comparing with `true` into the "Conditional expressions" section and split the example up for readability. Signed-off-by: Tim Wiederhake <twiederh(a)redhat.com&gt; --- docs/coding-style.rst | 60 +++++++++++++++++++++++++++---------------- 1 file changed, 38 insertions(+), 22 deletions(-) diff --git a/docs/coding-style.rst b/docs/coding-style.rst index dca9de1915..3dedb032f4 100644 --- a/docs/coding-style.rst +++ b/docs/coding-style.rst @@ -427,25 +427,47 @@ Conditional expressions ----------------------- For readability reasons new code should avoid shortening -comparisons to 0 for numeric types. Boolean and pointer -comparisons may be shortened. All long forms are okay: +comparisons to 0 for numeric types: :: - virFoo *foos = NULL; size nfoos = 0; - bool hasFoos = false; GOOD: - if (!foos) - if (!hasFoos) + if (nfoos != 0) if (nfoos == 0) - if (foos == NULL) - if (hasFoos == true) BAD: - if (!nfoos) if (nfoos) + if (!nfoos) + +Prefer the shortened version for boolean values. Boolean values +should never be compared against the literal ``true``, as a +logical non-false value need not be ``1``. + +:: + + bool hasFoos = false; + + GOOD: + if (hasFoos) + if (!hasFoos) + + BAD: + if (hasFoos == true) + if (hasFoos != false) + if (hasFoos == false) + if (hasFoos != true)

+ +Pointer comparisons may be shortened. All long forms are okay. +