trunk test result inconsistencies

[Andrew MacLeod]

Is anyone else seeing comparison problems on trunk?

I was having problems testing a patch on a 4/16 extraction, so last 
night I checked out a fresh trunk, built it, ran make check... then 
removed the build directory, re-built it from scratch again. make 
check..  and get a bunch of different results.     I even used 
test_summary instead of my own home-grown scripts.

Its not ordering this time, its actual reported failures on 
x86_64-unknown-linux-gnu...  for instance,
run 1 found, among other things:
PASS: g++.dg/cpp0x/rv1n.C  -std=c++14  (test for errors, line 80)
FAIL: g++.dg/cpp0x/rv1n.C  -std=c++14  (test for warnings, line 84)
PASS: g++.dg/cpp0x/rv1n.C  -std=c++14  (test for errors, line 89)
PASS: g++.dg/cpp0x/rv1n.C  -std=c++14  (test for errors, line 91)
FAIL: g++.dg/cpp0x/rv1n.C  -std=c++14  (test for errors, line 92)
FAIL: g++.dg/cpp0x/rv1n.C  -std=c++14  (test for errors, line 93)
FAIL: g++.dg/cpp0x/rv1n.C  -std=c++14  (test for errors, line 94)
FAIL: g++.dg/cpp0x/rv1n.C  -std=c++14  (test for errors, line 95)
<...>

but in run 2, those results were fine, but instead a different set of 
failtures showed up, like:
PASS: g++.dg/cpp0x/udlit-implicit-conv-neg.C  -std=c++11  (test for 
errors, line 47)
PASS: g++.dg/cpp0x/udlit-implicit-conv-neg.C  -std=c++11  (test for 
errors, line 49)
PASS: g++.dg/cpp0x/udlit-implicit-conv-neg.C  -std=c++11  (test for 
errors, line 50)
FAIL: g++.dg/cpp0x/udlit-implicit-conv-neg.C  -std=c++11  (test for 
errors, line 51)
FAIL: g++.dg/cpp0x/udlit-implicit-conv-neg.C  -std=c++11  (test for 
errors, line 53)
FAIL: g++.dg/cpp0x/udlit-implicit-conv-neg.C  -std=c++11  (test for 
errors, line 54)
FAIL: g++.dg/cpp0x/udlit-implicit-conv-neg.C  -std=c++11  (test for 
errors, line 55)
FAIL: g++.dg/cpp0x/udlit-implicit-conv-neg.C  -std=c++11  (test for 
errors, line 57)
<...>


I see in the log, it looks like something got truncated in the output 
from the compiler:
/home/amacleod/gcc/2015-04-21/gcc/gcc/testsuite/g++.dg/cpp0x/rv1n.C:89:27: 
error: use of deleted function 'constexpr A::A(const A&)'^M
/home/amacleod/gcc/2015-04-21/gcc/gcc/testsuite/g++.dg/cpp0x/rv1n.C:91:28: 
error: use of deleted function 'constexpr A::A(const A&)'^M
/home/amacle
PASS: g++.dg/cpp0x/rv1n.C  -std=c++14  (test for warnings, line 34)
PASS: g++.dg/cpp0x/rv1n.C  -std=c++14  (test for errors, line 39)

so the rest of the compiler output for that execution is not in the log, 
so a bunch of FAILS are reported since the messages aren't showing up.

Is there another known issue with mashing together multiple outputs, or 
truncating long output?  I'm using -j16 ... Is there something else I'm 
missing?   its clearly been an issue for almost a week at least

It appears to be mostly the gcc and g++ results.

Andrew