[llvm-dev] RFC Enable -Wimplicit-fallthrough for clang as well as GCC

[Mikael Holmén via llvm-dev]

Hi,

Does this mean that the clang version mentioned on

 
https://llvm.org/docs/GettingStarted.html#host-c-toolchain-both-compiler-and-standard-library

should be updated since it currently says that 3.1 is ok?

I noticed that one of our build-bots that ran clang 3.5.0 failed since 
it ended up in the

  #define LLVM_FALLTHROUGH

part of the ifdefs in include/llvm/Demangle/Compiler.h

#if __cplusplus > 201402L && __has_cpp_attribute(fallthrough)
#define LLVM_FALLTHROUGH [[fallthrough]]
#elif __has_cpp_attribute(gnu::fallthrough)
#define LLVM_FALLTHROUGH [[gnu::fallthrough]]
#elif !__cplusplus
// Workaround for llvm.org/PR23435, since clang 3.6 and below emit a 
spurious
// error when __has_cpp_attribute is given a scoped attribute in C mode.
#define LLVM_FALLTHROUGH
#elif __has_cpp_attribute(clang::fallthrough)
#define LLVM_FALLTHROUGH [[clang::fallthrough]]
#else
#define LLVM_FALLTHROUGH
#endif

and then we of course got tons of fallthrough warnings/errors (with 
-Werror).

Now we've updated the clang version on that build bot too, so it's no 
problem for us, but perhaps someone else will hit the same thing.

/Mikael

On 11/1/18 9:34 PM, Reid Kleckner via llvm-dev wrote:
> I split all the possible behavior changes out of the big cleanup patch, 
> and committed the cleanup as r345882.
> 
> I enabled the warning in r345887, and I'll see if it sticks.
> 
> Thanks for the reviews!
> 
> On Thu, Nov 1, 2018 at 10:25 AM Reid Kleckner <rnk at google.com 
> <mailto:rnk at google.com>> wrote:
> 
>     Great! Thanks everyone for the input, I'm going to start splitting
>     up the patch. I'll send out the non-mechanical parts separately.
> 
>     On Wed, Oct 31, 2018 at 10:03 PM Chris Lattner <sabre at nondot.org
>     <mailto:sabre at nondot.org>> wrote:
> 
>         On Oct 31, 2018, at 2:24 PM, Reid Kleckner <rnk at google.com
>         <mailto:rnk at google.com>> wrote:
>>         Alex Kornienko proposed enabling this warning back in 2012 here:
>>         http://lists.llvm.org/pipermail/llvm-dev/2012-July/051386.html
>>
>>         At the time, Chris Lattner said he didn't feel it was worth
>>         annotating
>>         all of LLVM and Clang with a new macro to enable this warning.
> 
>         FWIW, I formally withdraw my objection, go for it! :-)
> 
>         -Chris
> 
> 
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>