gcc-9.1.0

[Bruce Dubbs]

New major version.

Release notes are at ​https://gcc.gnu.org/gcc-9/changes.html

[Bruce Dubbs]

In the LFS Chapter 6 environment, I get the following failures:

FAIL: 22_locale/time_get/get_time/char/wrapped_env.cc execution test
FAIL: 22_locale/time_get/get_time/char/wrapped_locale.cc execution test
FAIL: 22_locale/time_get/get_time/wchar_t/2.cc execution test
FAIL: 22_locale/time_get/get_time/char/2.cc execution test
FAIL: 22_locale/time_get/get_time/wchar_t/wrapped_env.cc execution test
FAIL: 22_locale/time_get/get_time/wchar_t/wrapped_locale.cc execution test

FAIL: gcc.target/i386/pr57193.c scan-assembler-times movdqa 2
FAIL: experimental/net/internet/resolver/ops/lookup.cc execution test
FAIL: experimental/net/internet/resolver/ops/reverse.cc execution test
FAIL: gcc.target/i386/pr90178.c scan-assembler-times xorl[\\t ]*\\%eax,[\\t ]*%eax 1


      === gcc Summary ===

# of expected passes    139167
# of unexpected failures   2
# of expected failures     527
# of unsupported tests     2151

      === g++ Summary ===

# of expected passes    134426
# of expected failures     527
# of unsupported tests     5875

      === libatomic Summary ===

# of expected passes    54

      === libgomp Summary ===

# of expected passes    2312
# of expected failures     2
# of unsupported tests     210

      === libitm Summary ===

# of expected passes    42
# of expected failures     3
# of unsupported tests     1

      === libstdc++ Summary ===

# of expected passes    12850
# of unexpected failures   8
# of expected failures     77
# of unsupported tests     379

I guess that 10 failures out of about 300,000+ tests is not too bad. The six locale tests are already documented.

I'll do a build in the BLFS environment and compare.

[Bruce Dubbs]

Fixed at revision 11589. The update to gcc9 was pretty clean.

[Bruce Dubbs]

I did a full test in BLFS. The results were not great.

      === gcc Summary ===

# of expected passes      143750
# of unexpected failures     100
# of unexpected successes     28
# of expected failures       593
# of unsupported tests      2299

      === gfortran Summary ===

# of expected passes       50099
# of expected failures       178
# of unsupported tests        75

      === g++ Summary ===

# of expected passes      134691
# of unexpected failures       1
# of expected failures       527
# of unsupported tests      5923

      === go Summary ===

# of expected passes        7349
# of unexpected failures       1
# of expected failures         1
# of untested testcases        6
# of unsupported tests         1

      === obj-c++ Summary ===

# of expected passes        1458
# of expected failures        10
# of unsupported tests        77

      === objc Summary ===

# of expected passes        2797
# of expected failures         6
# of unsupported tests        68

      === gotools Summary ===
# of expected passes         333
# of unexpected failures       1
# of untested testcases      183

      === libatomic Summary ===

# of expected passes          54

      === libffi Summary ===

# of expected passes        2214

      === libgo Summary ===

# of expected passes         182
# of unexpected failures       2


      === libgomp Summary ===

# of expected passes        6454
# of expected failures         2
# of unsupported tests       349

      === libstdc++ Summary ===

# of expected passes       13563
# of unexpected failures       8
# of expected failures        77
# of unsupported tests       377

----------------

Totals:

# of expected passes       362986
# of unexpected failures      113

That's a failure rate of 0.03%, but I would be a lot happier without the 100 gcc failures. I looked for what was happening upstream and found a couple of referenceds that indicate they know about this two weeks ago and made fixes, but they didn't get into the release.

Similar results:

​https://gcc.gnu.org/ml/gcc-testresults/2019-05/msg00382.html

Almost all of the failures are caught by gdb so that is what they are not run in LFS Chapter 6.

[Douglas R. Reno]

If it helps, here's my test results from 8.3.0 when I updated it in BLFS last cycle. This is on an x86_64 system (Skylake).

renodr [ ~/Documents/Update Logs ]$ cat gcc-8.3.0-no_ada-update-tests2.log  | grep -A7 Summ
                === gcc Summary ===

# of expected passes            135425
# of unexpected failures        76
# of unexpected successes       20
# of expected failures          435
# of unsupported tests          2188
/sources/gcc-8.3.0/gcc-8.3.0/build/gcc/xgcc  version 8.3.0 (GCC) 
--
                === gfortran Summary ===

# of expected passes            47760
# of expected failures          83
# of unsupported tests          74
/sources/gcc-8.3.0/gcc-8.3.0/build/gcc/gfortran  version 8.3.0 (GCC) 

                === g++ tests ===
--
                === g++ Summary ===

# of expected passes            126210
# of unexpected failures        1
# of expected failures          507
# of unsupported tests          5050
/sources/gcc-8.3.0/gcc-8.3.0/build/gcc/xg++  version 8.3.0 (GCC) 

--
                === go Summary ===

# of expected passes            7285
# of unexpected failures        1
# of expected failures          1
# of untested testcases         6
# of unsupported tests          1
/sources/gcc-8.3.0/gcc-8.3.0/build/gcc/gccgo  version 8.3.0 (GCC) 
--
                === obj-c++ Summary ===

# of expected passes            1456
# of expected failures          10
# of unsupported tests          77
/sources/gcc-8.3.0/gcc-8.3.0/build/gcc/xg++  version 8.3.0 (GCC) 

                === objc tests ===
--
                === objc Summary ===

# of expected passes            2797
# of expected failures          6
# of unsupported tests          68
/sources/gcc-8.3.0/gcc-8.3.0/build/gcc/xgcc  version 8.3.0 (GCC) 

                === gotools tests ===
--
                === gotools Summary ===
# of expected passes            384
# of untested testcases         195
/sources/gcc-8.3.0/gcc-8.3.0/build/./gcc/gccgo version 8.3.0 (GCC)

                === libatomic tests ===


--
                === libatomic Summary ===

# of expected passes            54
                === libffi tests ===


Running target unix

                === libffi Summary ===

# of expected passes            2214
                === libgo tests ===


Running target unix
FAIL: runtime/pprof
--
                === libgo Summary ===

# of expected passes            162
# of unexpected failures        1
/sources/gcc-8.3.0/gcc-8.3.0/build/./gcc/gccgo version 8.3.0 (GCC)

                === libgomp tests ===

--
                === libgomp Summary ===

# of expected passes            5554
# of unsupported tests          291
                === libitm tests ===


Running target unix
--
                === libitm Summary ===

# of expected passes            42
# of expected failures          3
# of unsupported tests          1
                === libstdc++ tests ===


--
                === libstdc++ Summary ===

# of expected passes            13068
# of unexpected failures        6
# of expected failures          71
# of unsupported tests          257

Compiler version: 8.3.0 (GCC) 

For 8.3.0, the results were also a lot cleaner on i686, with 70 unexpected failures in GCC, 1 in gfortran, none in libstdc++, and no failures in Go, obj-c, obj-c++, gotools, etc.

[Bruce Dubbs]

Milestone renamed