Test results from building LFS git 20210609 on a Skylake i3. This replaces the earlier results from May, I had missed some of the numbers of PASSes in packages such as gettext. This is with a patch for gcc-11.1.0 to allow it to build firefox-89 and later firefoxes. This build uses the packages' own default CFLAGS. For tests, I'm only really concerned about test failures. glibc-2.33 only 1 failure FAIL: io/tst-lchmod On this run, misc/tst-ttyname was flagged as UNSUPPORTED: it starts to look as if elf/tst-cpu-features-cpuinfo does not fail on semi-recent intel machines. zlib-1.2.11 all OK bzip2-1.0.8 all 6 tests passed xz-5.2.5 all 9 tests passed zstd-1.5.0 No mention of any 'FAIL'. file-5.40 seems ok m4-1.4.18 All checks successful, then 157 PASS, 0 FAIL. bc-4.0.1 All tests (dc, bc) passed flex-2.6.4 All 114 tests passed tcl-8.6.10 Passed 30527, Failed 0 Unlike my build on the haswell, I was correctly using tcl-8.6.11. expect5.45.4 Passed 29, Failed 0 dejagnu-1.6.2 No failures (5 and 77 expected passes) binutils-2.36.1 7 failures in gold FAIL: debug_msg.sh FAIL: missing_key_func.sh FAIL: gdb_index_test_1.sh FAIL: gdb_index_test_2.sh FAIL: gdb_index_test_2_gabi.sh FAIL: gdb_index_test_3.sh FAIL: gdb_index_test_4.sh 4 failures in ld FAIL: Run property 4 FAIL: Run property 4 (PIE) FAIL: Run property 5 FAIL: Run property 5 (PIE) gmp-6.2.1 197 tests passed mpfr-4.1.0 181 PASS, 0 FAIL mpc-1.2.1 all 69 PASS attr-2.5.1 both tests PASS libcap-2.50 both PASSED gcc-11.1.0 Variation: I added a patch from upstream to fix breakage when trying to use g++ to compile firefox-89.0 (that is a regression, g++-10.2.0 is apparently able to build it, as is clang). g++ 3 unexpected failures FAIL: g++.dg/cpp0x/constexpr-52830.C -std=c++14 (test for excess errors) FAIL: g++.dg/cpp0x/constexpr-52830.C -std=c++17 (test for excess errors) FAIL: g++.dg/cpp0x/constexpr-52830.C -std=c++2a (test for excess errors) libstdc++ 8 unexpected failures, 1 associated new unresolved FAIL: 22_locale/numpunct/members/char/3.cc execution test FAIL: 22_locale/time_get/get_time/char/2.cc execution test 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/wchar_t/wrapped_env.cc execution test FAIL: 22_locale/time_get/get_time/wchar_t/wrapped_locale.cc execution test FAIL: 30_threads/this_thread/95989.cc (test for excess errors) UNRESOLVED: 30_threads/this_thread/95989.cc compilation failed to produce executable pkg-config-0.29.2 0 tests were run in the prelimary parts. In the main part All 30 tests passed sed-4.8 59, 157 PASS, 0 FAIL gettext-0.21 0, 2, 422, 14, 252 PASS, 0 FAIL bison-3.7.6 11 PASS, 0 FAIL in initial part, 620 successful in main part grep-3.6 94, 184 PASS, 0 FAIL bash-5.1 No failures nor any diff output. libtool-2.4.6 5 unexpected failures, as noted in the book gdbm-1.19 All 30 tests were successful gdbmtool: FAIL: version gperf-3.1 no failures reported expat-2.4.1 both tests PASS inetutils-2.0 10 PASS, 0 FAIL perl-5.34.0 Failed 1 test out of 2542, 99.96% okay. ../cpan/Pod-Usage/t/pod/pod2usage2.t Looks like this maybe always fails with perl-5.34 Possibly related to https://github.com/Perl/perl5/issues/18855 XML-Parser-2.46 PASS intltool-0.51.0 1 PASS, 0 FAIL autoconf-2.71 543 tests behaved as expected automake-1.16.3 2721 PASS, 2 FAIL - FAIL: t/deprecated-acinit.sh FAIL: t/init.sh Note that t/subobj.sh PASSed libtool testing after automake - 138 tests behaved as expected. elfutils-0.185 222 PASS, 0 FAIL libffi-3.3 --with-gcc-arch=x86-64 2284 expected passes openssl-1.1.1k All tests successful. PASS ninja-1.10.2 passed coreutils-8.32 491 and 328 PASS, 1 FAIL in first part: FAIL: tests/chown/separator.sh This has failed since (at least) 10.1 on this machine acl-2.3.1 after coreutils 13 PASS, 0 FAIL check-0.15.2 1 and 9 PASS, 0 FAIL diffutils-3.7 20 and 160 PASS, 0 FAIL gawk-5.1.0 ALL TESTS PASSED findutils-4.8.0 2, 238, 14 PASS, 0 FAIL gzip-1.10 22 PASS, 0 FAIL kbd-2.4.0 36 tests were successful libpipeline-1.5.3 7 PASS, 0 FAIL make-4.3 690 Tests in 125 Categories Complete ... No Failures :-) patch-2.7.6 41 PASS, 0 FAIL man-db-2.9.4 29, 9, 12 PASS, 0 FAIL tar-1.34 216 tests were successful 227: capabilities: binary store/restore is skipped on this machine I suspect this particular test is perhaps reliant on something in the kernel config which I have not enabled on my intel machines. texinfo-6.7 85, 57, 0, 83, 0, 1, 1, 1 PASS, 0 FAIL vim-8.2.2890 Executed 3822, Failed 0 eudev-3.2.10 PASS 2, FAIL 0 in initial part, then 0 errors occurred procps-ng-3.3.17 5 FAIL on pkill tests, as noted util-linux-2.36.2 1 tests of 207 FAILED column: invalid multibyte ... FAILED (this seems to always fail for me) e2fsprogs-1.46.2 372 tests succeeded, 0 failed Python-3.9.5 tests were run after sqlite3 had been installed: == Tests result: SUCCESS == 406 tests OK.