Hi,

On 2019-06-06 13:32:16 -0400, Alvaro Herrera wrote:
> Seawasp (using experimental clang 9.0) has been complaining of late:
> 
> /home/fabien/clgtk/bin/clang -Wno-ignored-attributes -fno-strict-aliasing 
> -fwrapv -O2  -D__STDC_LIMIT_MACROS -D__STDC_FORMAT_MACROS 
> -D__STDC_CONSTANT_MACROS -D_DEBUG -D_GNU_SOURCE -I/home/fabien/clgtk/include  
> -I../../../../src/include  -D_GNU_SOURCE -I/usr/include/libxml2  -flto=thin 
> -emit-llvm -c -o llvmjit_types.bc llvmjit_types.c
> In file included from /home/fabien/clgtk/include/llvm/ADT/DenseMapInfo.h:20:0,
>                  from /home/fabien/clgtk/include/llvm/ADT/DenseMap.h:16,
>                  from /home/fabien/clgtk/include/llvm/ADT/DenseSet.h:16,
>                  from /home/fabien/clgtk/include/llvm/ADT/SetVector.h:23,
>                  from llvmjit_inline.cpp:45:
> /home/fabien/clgtk/include/llvm/Support/ScalableSize.h:27:12: error: macro 
> "Min" requires 2 arguments, but only 1 given
>    : Min(Min), Scalable(Scalable) {}
>             ^
> In file included from /home/fabien/clgtk/include/llvm/ADT/DenseMapInfo.h:20:0,
>                  from /home/fabien/clgtk/include/llvm/ADT/DenseMap.h:16,
>                  from /home/fabien/clgtk/include/llvm/ADT/DenseSet.h:16,
>                  from /home/fabien/clgtk/include/llvm/ADT/SetVector.h:23,
>                  from llvmjit_inline.cpp:45:
> /home/fabien/clgtk/include/llvm/Support/ScalableSize.h: In constructor 
> \xe2\x80\x98llvm::ElementCount::ElementCount(unsigned int, bool)\xe2\x80\x99:
> /home/fabien/clgtk/include/llvm/Support/ScalableSize.h:27:13: error: expected 
> \xe2\x80\x98(\xe2\x80\x99 before \xe2\x80\x98,\xe2\x80\x99 token
>    : Min(Min), Scalable(Scalable) {}
>              ^
> <builtin>: recipe for target 'llvmjit_inline.o' failed
> 
> This was working earlier, and as far as I can tell the cpluspluscheck
> fixes are not the cause (because those happened earlier than the first
> failure).  Apparently clang got upgraded from "trunk 361691" to "trunk
> 362290" ... is the new clang broken?

I think that machine might also update llvm to a trunk checkout. Is that
right Fabien?  If so that's possible "just" a minor API break.

Greetings,

Andres Freund


Reply via email to