[Nfd-dev] code-style: switch with compound statement

[Wentao Shang]

On Wednesday, August 12, 2015, Junxiao Shi <shijunxiao at email.arizona.edu>
wrote:

> Dear folks
>
>
>
> It seems that I misunderstood Wentao’s opinion about not to mix compound
> and non-compound ‘case’ labels.
>
> It indeed meant that when a compound statement is used, the ‘break’ or ‘//
> Fallthrough’ should be inside the compound.
>
> So:
>
>
>
> // option F
>
> switch (cond) {
>
>   case 1:
>
>   {
>
>     int i = 1;
>
>     doSomething(i);
>
>     // Fallthrough
>
>   }
>
>
>
>   case 2:
>
>     doSomething(2);
>
>     break;
>
>
>
>   case 3:
>
>   {
>
>     int j = 3;
>
>     doSomething(j);
>
>     break;
>
>   }
>
>
>
>   default:
>
>     doSomething(0);
>
>     break;
>
> }
>
>
>
> ·         When a ‘case’ or ‘default’ contains a declaration and therefore
> needs a compound statement, all statements under this label should be
> placed inside the compound statement, including ‘break’ or ‘// Fallthrough’.
>
> ·         When any ‘case’ or ‘default’ label is using a compound
> statement, all ‘case’ and ‘default’ labels should be indented two spaces
> after ‘switch’, ‘{‘ and ‘}’ of the compound statements are indented at the
> same level as ‘case’ and ‘default’, and statements within the compound
> statements are indented two spaces after ‘case’ and ‘default’.
>
>
>
> A ‘switch’ statement without compound statements may still be formatted in
> the current way, i.e. no extra indentation for ‘case’ and ‘default’.
>
>
>
> However, if we allow a mix of compound and non-compound labels, we should
> also permit (but not require) 2-space indented ‘case’ and ‘default’ labels
> even if none of the labels contains a compound statement. Otherwise, the
> addition or removal of a label with compound statement would require
> indenting or unindenting the entire switch statement, which is unnecessary.
>
Agree. It definitely should not be forbidden. Actually I'm leaning towards
having a uniform format for both cases, i.e., always indent inner
statements of switch.

Wentao

>
>
> Yours, Junxiao
>
>
>
>
> *From: *Wentao Shang
> *Sent: *Wednesday, August 12, 2015 16:01
> *To: *Junxiao Shi
> *Cc: *;Eric Newberry
> *Subject: *Re: [Nfd-dev] code-style: switch with compound statement
>
>
>
>
>
>
>
> On Wednesday, August 12, 2015, Junxiao Shi <shijunxiao at email.arizona.edu
> <javascript:_e(%7B%7D,'cvml','shijunxiao at email.arizona.edu');>> wrote:
>
> ·         When any ‘case’ or ‘default’ contains a declaration and
> therefore needs a compound statement, all ‘case’s and ‘default’ must use
> compound statements.
>
> I wouldn't go that far as to require all cases to have the same format. I
> think it's ok to mix blocked case and non-blocked case, as long as the
> block covers the entire single case branch. But the current specification
> (option E by Junxiao) also looks fine to me.
>
> ·         ‘break’ and ‘// Fallthrough’ should be placed inside compound
> statements.
>
> I would suggest one more readability requirement: any intentional
> fall-through MUST be explicitly annotated in the code. (Do we have this
> rule already? I can't remember.)
>


-- 
PhD @ IRL, CSD, UCLA
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/nfd-dev/attachments/20150812/51c1081e/attachment.html>