Bug#822783: eztrace-contrib: FTBFS with libc 2.23: 'memcpy' was not declared in this scope

[Samuel Thibault]

Control: reassign -1 nvidia-cuda-dev

Samuel Thibault, on Tue 31 May 2016 11:29:40 +0200, wrote:
> Aurelien Jarno, on Tue 31 May 2016 11:27:33 +0200, wrote:
> > On 2016-04-28 15:09, Samuel Thibault wrote:
> > > Control: reassign -1 libc6-dev
> > > Control: tags -1 + experimental
> > > 
> > > Graham Inggs, on Wed 27 Apr 2016 15:07:52 +0200, wrote:
> > > > Eztrace-contrib FTBFS with glibc 2.23 available in Experimental and
> > > > Ubuntu Xenial.
> > > > 
> > > > > /usr/include/string.h: In function ‘void* __mempcpy_inline(void*, const void*, size_t)’:
> > > > > /usr/include/string.h:652:42: error: ‘memcpy’ was not declared in this scope
> > > > >    return (char *) memcpy (__dest, __src, __n) + __n;
> > > > >                                           ^
> > > > > /usr/include/string.h: In function ‘void* __mempcpy_inline(void*, const void*, size_t)’:
> > > > > /usr/include/string.h:652:42: error: ‘memcpy’ was not declared in this scope
> > > > >    return (char *) memcpy (__dest, __src, __n) + __n;
> > > > >                                           ^
> > > > > /usr/include/string.h: In function ‘void* __mempcpy_inline(void*, const void*, size_t)’:
> > > > > /usr/include/string.h:652:42: error: ‘memcpy’ was not declared in this scope
> > > > >    return (char *) memcpy (__dest, __src, __n) + __n;
> > > > 
> > > > I found a similar issue had been reported for TensorFlow [1], and the solution:
> > > > 
> > > > > @e14159 can you try with -D_FORCE_INLINES? I just had the same issue with pcl,
> > > > > I checked the string.h header, and using that preprocessor directive skips the block
> > > > > where the memcpy error appears. There's probably a cleaner workaround though...
> > > 
> > > Well, a workaround remains a workaround. There is no reason why we
> > > shouldn't just fix the bug at its source, glibc. Otherwise we'd keep
> > > chasing packages which would need a workaround, that's really not the
> > > way to go.
> > > 
> > > So I'm reassigning to glibc.
> > 
> > I am not fully sure the bug is actually in on the libc sid. It looks
> > like that nvcc redefines part of this header in a way that is not
> > compatible with the new version of the header.

So reassigning to cuda, perhaps a proper fix can be done in the .h files
there without having to ask upstream :)

Samuel