On Tue, Feb 9, 2016 at 4:16 PM, David Sommerseth
<[email protected]> wrote:
> On 08/02/16 17:03, Arne Schwabe wrote:
>> Am 07.02.16 um 20:47 schrieb Steffan Karger:
>>> This keeps naming consistent. For example, instead of id-aes128-GCM use
>>> AES-128-GCM, which is more like AES-128-CBC.
>>>
>>> +
>>> + if (NULL == pair)
>>> + return cipher_name;
>>> +
>> I like the pair == NULL style more.
>
> I "kind of" agree, as that's what we've grown up with. However, from a
> defensive programming style, this is a far better approach. A mistyping using
> '=' will in 'NULL = pair' case cause a compile failure. The other way around
> will compile just fine with a nice ugly bug attached with it ... So I'm trying
> to get used to his "flipped around" style.
This indeed the motivation is.
But, modern compilers do detect this (given the correct flags):
gcc 5.2.1:
$ clang -o no-yoda no-yoda.c
no-yoda.c:5:9: warning: using the result of an assignment as a condition without
parentheses [-Wparentheses]
if (a = NULL) {
~~^~~~~~
no-yoda.c:5:9: note: place parentheses around the assignment to silence this
warning
if (a = NULL) {
^
( )
no-yoda.c:5:9: note: use '==' to turn this assignment into an equality
comparison
if (a = NULL) {
^
==
clang 3.6.2:
$ clang -o no-yoda no-yoda.c
no-yoda.c:5:9: warning: using the result of an assignment as a condition without
parentheses [-Wparentheses]
if (a = NULL) {
~~^~~~~~
So, I do not have any strong preference.
-Steffan
