[llvm-dev] Codifying our Brace rules-

[Aaron Ballman via llvm-dev]

On Wed, Jun 24, 2020 at 11:17 AM Fangrui Song via llvm-dev
<llvm-dev at lists.llvm.org> wrote:
>
> On 2020-06-24, David Chisnall via llvm-dev wrote:
> >On 23/06/2020 16:51, Adrian McCarthy via llvm-dev wrote:
> >>Personally, I favor "always use braces" because it helps readability
> >>for me.  The compiler may be good at flagging misleading
> >>indentation, but my visual processing system is terrible at it,
> >>especially since we use a measly two spaces for indentation.  And we
> >>grant indentation exceptions for--among other things--case labels in
> >>switches.
> >
> >I remember fixing one bug in LLVM that was caused by not having braces
> >around an if statement and a modification that didn't notice this.  I
> >have fixed other bugs in a downstream fork where upstream has done
> >this:
> >
> >if (x) {
> >  something();
> >  somethingElse();
> >}
> >
> >We've changed it to:
> >
> >if (x) {
> >  something();
> >  somethingElse();
> >  ohAndOneMoreThing();
> >}
> >
> >And then upstream has changed it to
> >
> >if (x)
> >  something();
> >
> >And, after merge, our code has been:
> >
> >if (x)
> >  something();
> >  ohAndOneMoreThing();
> >
> >In contrast, I have *never*, in any project, had to fix a bug that was
> >caused by someone inserting redundant braces (or extra brackets) and
> >someone else failing to understand the code.
> >
> >As such, I am strongly in favour of requiring extra braces.  Static
> >analysers will now often catch the bugs that they prevent but it's
> >better to make it harder to introduce bugs in the first place than to
> >depend on tooling to find them later.
> >
> >David
>
> First, there is a merge conflict:
>
> --- i/a.c
> +++ w/a.c
> @@@ -1,2 -1,5 +1,8 @@@
>   -if (x) {
>   +if (x)
>       something();
> ++<<<<<<< HEAD
> ++=======
> +    somethingElse();
> +    onemorething();
> + }
> ++>>>>>>> 78750e9... onemore
>
>
> Second, both GCC>=6 and clang (https://reviews.llvm.org/D70638 ) -Wall enable
> -Wmisleading-indentation to catch the issue.

This second point only holds true if you haven't clang-formatted the
code. Once you run the code through clang-format, the misleading
indentation goes away.

~Aaron

> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev