make check failures

[Toebs Douglass]

On 14/11/16 14:41, Carlos O'Donell wrote:
> On Sun, Nov 13, 2016 at 11:29 AM, Toebs Douglass <toby@winterflaw.net> wrote:
>> Comparing the make check results for the system compiler and my
>> hand-built compiler, I see;
> 
>> Which is obviously interesting ("conform" for all failures) but I'm not
>> quite sure what to make of it and concomitantly unsure of what action to
>> take.  It implies (perhaps) the other failures can since they affect
>> both compilers be set aside (or, that there's something wrong with my
>> system as a whole, which is affecting the glibc in general, regardless
>> of compiler).
> 
> Each of the glibc releases has common failure lists for all supported arches.
> 
> You can see the ones for glibc 2.24 here:
> 
> https://sourceware.org/glibc/wiki/Release/2.24

Wonderful!

> What you have looks like an environmental issue when building.
> 
> If you build with a non-/usr --prefix then you must assemble a
> complete sysroot at that location for the build process.

Yaaahhhhhkkkkk!

> This includes having the right kernel headers, and a dynamically
> loadable compiler helper routine library (libgcc_s.so) for
> cancellation (the threading cancellation tests are failing in your
> example).

The hand-built GCC has a libgcc_s.so, but I didn't set LIBRARY_PATH to
include it - I was concerned it would mess up the BUILD_CC compiler.  I
would guess then the wrong libgcc is being used.

> I suggest you start here:
> https://sourceware.org/glibc/wiki/Testing/Builds

Thankyou -so- much, Carlos.  That should see me a good long way further
towards a successful build.