Julia: Building from source with system dependencies has numerous failures
Attempting to switch on USE_SYSTEM_* in turn for various components produces the following build or testsuite failures:
pcre and mpfr both fail the stress test:
pcre:
Test (Worker) | Time (s) | GC (s) | GC % | Alloc (MB) | RSS (MB)
compiler/validation (4) | 3.52 | 0.19 | 5.4 | 154.15 | 215.89
strings/search (7) | 5.34 | 0.24 | 4.4 | 246.64 | 215.89
strings/io (9) | 6.26 | 0.23 | 3.8 | 257.39 | 221.52
unicode/utf8 (7) | 0.67 | 0.00 | 0.0 | 2.25 | 215.89
From worker 2: Skipping Base.Filesystem.JL_O_RANDOM
From worker 2: Skipping Base.Filesystem.JL_O_SEQUENTIAL
From worker 2: Skipping Base.Filesystem.JL_O_SHORT_LIVED
From worker 2: Skipping Base.Filesystem.JL_O_TEMPORARY
From worker 2: Skipping Base.active_repl
From worker 2: Skipping Base.active_repl_backend
strings/util (8) | 9.20 | 0.19 | 2.0 | 309.89 | 222.77
worlds (7) | 4.18 | 0.07 | 1.6 | 131.74 | 215.89
strings/types (4) | 8.30 | 0.26 | 3.1 | 1183.92 | 216.64
From worker 2: Skipping Base.active_repl
From worker 2: Skipping Base.active_repl_backend
compiler/compiler (3) | 15.84 | 0.37 | 2.3 | 690.13 | 240.75
From worker 2: Skipping Base.Filesystem.JL_O_RANDOM
From worker 2: Skipping Base.Filesystem.JL_O_SEQUENTIAL
From worker 2: Skipping Base.Filesystem.JL_O_SHORT_LIVED
From worker 2: Skipping Base.Filesystem.JL_O_TEMPORARY
From worker 2: Skipping Base.active_repl
From worker 2: Skipping Base.active_repl_backend
keywordargs (8) | 8.55 | 0.06 | 0.7 | 111.67 | 225.67
triplequote (8) | 0.04 | 0.00 | 0.0 | 1.66 | 225.67
ambiguous (2) | 18.73 | 0.40 | 2.1 | 1153.26 | 241.27
intrinsics (8) | 0.42 | 0.01 | 1.3 | 26.16 | 225.77
char (3) | 3.83 | 0.16 | 4.3 | 227.22 | 254.04
strings/basic (6) | 25.26 | 0.40 | 1.6 | 666.98 | 242.88
iobuffer (3) | 4.97 | 0.07 | 1.5 | 117.25 | 269.21
staged (6) | 6.80 | 0.17 | 2.4 | 266.15 | 244.02
hashing (8) | 35.58 | 0.70 | 2.0 | 1475.74 | 261.31
subtype (4) | 46.69 | 1.15 | 2.5 | 1797.38 | 694.52
tuple (8) | 13.15 | 0.19 | 1.4 | 264.37 | 278.21
dict (2) | 59.54 | 1.55 | 2.6 | 1971.63 | 300.71
reduce (4) | 25.98 | 0.99 | 3.8 | 1308.81 | 694.52
intfuncs (4) | 4.33 | 0.04 | 1.0 | 139.69 | 694.52
simdloop (4) | 3.30 | 0.11 | 3.2 | 266.22 | 694.52
vecelement (4) | 20.97 | 0.63 | 3.0 | 1326.66 | 694.52
core (9) | 107.83 | 10.95 | 10.2 | 10730.47 | 621.39
offsetarray (3) | 95.76 | 2.78 | 2.9 | 3357.09 | 349.64
copy (3) | 3.22 | 0.09 | 2.7 | 158.77 | 354.75
rational (4) | 20.51 | 0.56 | 2.7 | 854.60 | 694.52
reducedim (8) | 67.57 | 2.20 | 3.3 | 3260.58 | 332.75
fastmath (4) | 13.93 | 0.44 | 3.1 | 534.03 | 694.52
functional (8) | 13.78 | 0.46 | 3.4 | 604.60 | 347.16
numbers (7) | 146.10 | 2.80 | 1.9 | 5016.75 | 388.21
operators (8) | 8.58 | 0.28 | 3.3 | 503.21 | 360.43
path (7) | 3.61 | 0.02 | 0.6 | 37.99 | 413.22
parse (7) | 7.35 | 0.29 | 4.0 | 363.09 | 414.13
abstractarray (2) | 119.79 | 4.72 | 3.9 | 5288.67 | 406.29
math (3) | 83.55 | 1.75 | 2.1 | 3010.11 | 414.48
bigfloat (3) | 0.95 | 0.02 | 2.5 | 40.31 | 414.48
From worker 7: WARNING: replacing module Foo.
loading (7) | 48.87 | 2.20 | 4.5 | 3985.50 | 427.63
ccall (8) | 60.06 | 41.82 | 69.6 | 783.99 | 390.25
From worker 7: [stdio passthrough ok]
iterators (4) | 70.42 | 2.64 | 3.8 | 3827.68 | 719.09
bigint (2) | 10.08 | 0.26 | 2.6 | 310.46 | 410.27
backtrace (8) | 2.11 | 0.03 | 1.3 | 36.32 | 398.46
version (8) | 1.94 | 0.05 | 2.6 | 99.16 | 399.62
file (4) | 5.89 | 0.18 | 3.1 | 264.31 | 733.49
namedtuple (8) | 4.74 | 0.17 | 3.5 | 197.66 | 401.43
arrayops (6) | 196.56 | 7.18 | 3.7 | 9208.54 | 535.19
spawn (7) | 25.69 | 0.39 | 1.5 | 577.72 | 436.00
floatapprox (7) | 2.53 | 0.08 | 3.1 | 152.77 | 436.92
mpfr (4) | 23.94 | 0.57 | 2.4 | 1193.60 | 778.85
regex (4) | 0.76 | 0.02 | 3.3 | 36.53 | 778.85
reflection (7) | 7.59 | 0.32 | 4.2 | 469.47 | 445.36
float16 (4) | 3.16 | 0.09 | 2.7 | 159.65 | 778.85
sysinfo (4) | 0.23 | 0.01 | 5.2 | 15.33 | 778.85
combinatorics (7) | 1.30 | 0.05 | 3.5 | 53.01 | 448.13
env (4) | 0.76 | 0.01 | 1.5 | 19.67 | 778.85
complex (6) | 22.57 | 0.66 | 2.9 | 853.61 | 535.19
mod2pi (6) | 0.73 | 0.01 | 2.0 | 26.55 | 535.19
sorting (3) | 49.32 | 1.45 | 2.9 | 2377.99 | 450.04
euler (6) | 1.75 | 0.09 | 5.0 | 130.47 | 535.19
rounding (7) | 8.80 | 0.32 | 3.6 | 515.73 | 466.38
read (2) | 43.48 | 1.97 | 4.5 | 3298.00 | 627.56
goto (2) | 0.12 | 0.00 | 0.0 | 3.82 | 627.73
llvmcall (2) | 0.85 | 0.02 | 2.9 | 55.60 | 627.73
llvmcall2 (2) | 0.07 | 0.00 | 0.0 | 2.00 | 627.73
grisu (2) | 4.89 | 0.12 | 2.5 | 212.25 | 627.73
errorshow (6) | 13.41 | 0.42 | 3.2 | 568.20 | 538.46
some (2) | 2.31 | 0.06 | 2.4 | 99.19 | 627.73
meta (6) | 2.38 | 0.09 | 3.8 | 89.59 | 539.43
stacktraces (2) | 5.09 | 0.16 | 3.2 | 203.13 | 627.73
sets (7) | 24.21 | 0.70 | 2.9 | 763.24 | 477.86
From worker 7: Warning: threaded loop executed in order
docs (6) | 14.32 | 0.65 | 4.5 | 546.64 | 563.70
enums (6) | 10.61 | 1.02 | 9.6 | 384.46 | 564.53
broadcast (8) | 82.81 | 5.03 | 6.1 | 4726.82 | 527.19
threads (7) | 33.13 | 0.38 | 1.1 | 236.46 | 486.01
From worker 2: Test Summary: | Pass Total
From worker 2: @deprecate | 6 6
From worker 2: Test Summary: | Pass Total
From worker 2: firstcaller | 1 1
From worker 2: Test Summary: | Broken Total
From worker 2: parser syntax deprecations | 1 1
misc (2) | 53.78 | 3.02 | 5.6 | 1908.49 | 656.34
int (8) | 25.85 | 0.45 | 1.7 | 273.94 | 548.45
bitarray (9) | 229.27 | 14.02 | 6.1 | 12585.36 | 629.53
inline (9) | 1.57 | 0.10 | 6.3 | 13.80 | 630.17
bitset (2) | 11.25 | 0.09 | 0.8 | 70.89 | 656.79
show (3) | 90.40 | 3.55 | 3.9 | 2788.23 | 513.42
cartesian (3) | 0.42 | 0.00 | 0.0 | 1.16 | 513.52
checked (7) | 31.07 | 0.10 | 0.3 | 48.95 | 518.24
osutils (3) | 1.27 | 0.00 | 0.0 | 0.61 | 514.31
error (2) | 6.56 | 0.24 | 3.6 | 81.89 | 657.27
iostream (3) | 3.57 | 0.12 | 3.3 | 71.10 | 514.36
specificity (3) | 0.43 | 0.03 | 6.0 | 4.20 | 514.42
ranges (4) | 106.62 | 8.70 | 8.2 | 8564.29 | 830.43
boundscheck (9) | 16.66 | 0.13 | 0.8 | 20.74 | 630.99
secretbuffer (2) | 3.51 | 1.04 | 29.5 | 40.49 | 657.41
syntax (9) | 6.34 | 0.12 | 1.9 | 99.07 | 632.53
codegen (3) | 11.15 | 0.98 | 8.7 | 26.95 | 533.74
floatfuncs (8) | 34.26 | 1.68 | 4.9 | 661.74 | 565.20
channels (7) | 27.74 | 2.65 | 9.5 | 530.75 | 585.53
logging (2) | 25.48 | 0.61 | 2.4 | 467.10 | 657.41
reinterpretarray (4) | 29.89 | 1.58 | 5.3 | 760.70 | 863.30
asyncmap (3) | 29.08 | 1.65 | 5.7 | 696.63 | 533.74
missing (9) | 34.04 | 1.45 | 4.3 | 567.98 | 651.17
subarray (5) | 490.57 | 32.71 | 6.7 | 24178.32 | 730.59
Pkg/resolve (4) | 111.53 | 12.19 | 10.9 | 6461.94 | 916.42
cmdlineargs (6) | 339.49 | 0.46 | 0.1 | 282.26 | 564.53
LinearAlgebra/schur (6) | 89.05 | 4.56 | 5.1 | 2071.92 | 584.50
LinearAlgebra/qr (9) | 349.23 | 27.03 | 7.7 | 10402.17 | 878.09
LinearAlgebra/matmul (4) | 251.77 | 15.46 | 6.1 | 5256.24 | 968.18
LinearAlgebra/special (6) | 84.62 | 3.75 | 4.4 | 1518.12 | 622.92
SparseArrays/higherorderfns (8) | 492.97 | 30.73 | 6.2 | 12247.26 | 752.01
LinearAlgebra/eigen (9) | 119.19 | 10.42 | 8.7 | 2868.32 | 953.38
LinearAlgebra/bunchkaufman (4) | 146.49 | 9.02 | 6.2 | 3612.94 | 1011.61
LinearAlgebra/svd (6) | 175.11 | 11.80 | 6.7 | 3901.14 | 710.16
SparseArrays/sparse (7) | 650.06 | 105.60 | 16.2 | 10478.00 | 730.51
LinearAlgebra/tridiag (9) | 193.28 | 9.98 | 5.2 | 2925.90 | 1002.51
LinearAlgebra/lapack (8) | 201.86 | 13.52 | 6.7 | 3894.11 | 865.50
SparseArrays/sparsevector (2) | 730.09 | 34.43 | 4.7 | 18190.98 | 905.14
LinearAlgebra/bidiag (4) | 246.43 | 14.00 | 5.7 | 4614.16 | 1126.11
LinearAlgebra/dense (5) | 684.14 | 59.40 | 8.7 | 15176.97 | 1060.49
LinearAlgebra/generic (2) | 88.59 | 5.39 | 6.1 | 1830.64 | 928.02
LinearAlgebra/uniformscaling (4) | 74.30 | 4.72 | 6.3 | 1282.67 | 1160.13
LinearAlgebra/hessenberg (2) | 18.78 | 1.07 | 5.7 | 519.05 | 931.70
LinearAlgebra/adjtrans (2) | 94.01 | 6.48 | 6.9 | 2253.27 | 939.31
LinearAlgebra/lu (9) | 267.55 | 21.77 | 8.1 | 4766.23 | 1124.69
LinearAlgebra/cholesky (7) | 306.55 | 17.56 | 5.7 | 6416.01 | 853.47
LinearAlgebra/blas (4) | 126.59 | 7.11 | 5.6 | 1936.44 | 1228.89
LinearAlgebra/givens (9) | 29.65 | 1.48 | 5.0 | 336.92 | 1126.14
LinearAlgebra/pinv (2) | 48.39 | 4.35 | 9.0 | 1260.57 | 1031.24
LinearAlgebra/lq (5) | 206.12 | 20.89 | 10.1 | 6268.26 | 1230.05
Dates/query (5) | 4.97 | 0.11 | 2.1 | 23.10 | 1235.89
Dates/adjusters (2) | 20.35 | 0.68 | 3.3 | 271.59 | 1035.46
LinearAlgebra/symmetric (8) | 366.30 | 24.82 | 6.8 | 5999.72 | 1005.37
LinearAlgebra/structuredbroadcast (7) | 96.84 | 7.25 | 7.5 | 2231.46 | 893.03
Dates/types (7) | 10.60 | 0.38 | 3.6 | 68.31 | 909.90
Dates/rounding (8) | 19.56 | 0.90 | 4.6 | 125.10 | 1013.34
Dates/ranges (2) | 54.79 | 3.70 | 6.8 | 1093.80 | 1047.39
Dates/periods (5) | 76.51 | 5.72 | 7.5 | 965.24 | 1256.53
Dates/conversions (2) | 7.52 | 0.21 | 2.9 | 58.59 | 1048.29
CRC32c (2) | 7.02 | 1.25 | 17.9 | 114.69 | 1048.76
Base64 (5) | 12.71 | 0.96 | 7.6 | 190.64 | 1315.90
LinearAlgebra/diagonal (6) | 498.22 | 38.09 | 7.6 | 9931.00 | 911.45
Dates/arithmetic (8) | 40.82 | 2.37 | 5.8 | 649.40 | 1035.79
Dates/io (7) | 50.39 | 2.67 | 5.3 | 867.50 | 922.62
DelimitedFiles (2) | 24.34 | 1.67 | 6.9 | 501.67 | 1048.76
Libdl (2) | 0.78 | 0.02 | 3.1 | 28.97 | 1048.76
Future (8) | 4.40 | 0.23 | 5.2 | 83.15 | 1036.29
Logging (2) | 14.05 | 1.03 | 7.4 | 350.58 | 1048.76
Markdown (8) | 21.74 | 1.73 | 7.9 | 479.97 | 1037.42
InteractiveUtils (7) | 31.53 | 2.02 | 6.4 | 587.83 | 977.88
Dates/accessors (9) | 173.97 | 9.29 | 5.3 | 3765.97 | 1128.21
FileWatching (6) | 49.28 | 8.22 | 16.7 | 195.11 | 918.34
Profile (7) | 12.74 | 0.71 | 5.6 | 305.13 | 981.47
From worker 4: The following 'Resetting the helper list...' warning is expected:
Mmap (2) | 36.95 | 25.06 | 67.8 | 296.74 | 1048.76
┌ Warning: Resetting the helper list is currently unsupported:
│ ignoring all git credential helpers
└ @ LibGit2 ~/julia/src/julia/usr/share/julia/stdlib/v1.0/LibGit2/src/gitcredential.jl:229
Serialization (2) | 20.33 | 1.86 | 9.2 | 672.75 | 1048.76
Printf (8) | 99.25 | 10.08 | 10.2 | 2987.53 | 1064.66
┌ Warning: TerminalMenus: Unable to enter raw mode: ArgumentError("stream is closed or unusable")
└ @ REPL.TerminalMenus ~/julia/src/julia/usr/share/julia/stdlib/v1.0/REPL/src/TerminalMenus/util.jl:21
From worker 9:
┌ Warning: TerminalMenus: Unable to enter raw mode: ArgumentError("stream is closed or unusable")
└ @ REPL.TerminalMenus ~/julia/src/julia/usr/share/julia/stdlib/v1.0/REPL/src/TerminalMenus/util.jl:21
From worker 9:
REPL (9) | 135.78 | 10.01 | 7.4 | 2211.22 | 1203.76
Random (6) | 146.50 | 9.37 | 6.4 | 2722.16 | 1006.78
UUIDs (6) | 0.42 | 0.00 | 0.0 | 6.92 | 1006.78
Unicode (6) | 16.64 | 0.59 | 3.5 | 247.89 | 1018.78
Statistics (2) | 133.77 | 9.77 | 7.3 | 2396.21 | 1055.93
┌ Warning: Omitting `@test_deprecated` tests which can't yet
│ be tested in --depwarn=error mode
└ @ Main.Test77Main_Test ~/julia/src/julia/usr/share/julia/stdlib/v1.0/Test/test/runtests.jl:736
Test (9) | 88.86 | 2.32 | 2.6 | 388.82 | 1204.68
SuiteSparse (8) | 201.04 | 12.39 | 6.2 | 3520.37 | 1142.39
LibGit2/libgit2 (4) | 483.44 | 6.14 | 1.3 | 1532.96 | 1603.24
Distributed (5) | 369.16 | 0.23 | 0.1 | 26.94 | 1315.90
SHA (7) | 323.17 | 2.16 | 0.7 | 647.92 | 1030.83
LinearAlgebra/triangular (3) | 1505.82 | 89.13 | 5.9 | 35050.41 | 1407.03
Executing tests that run on node 1 only:
precompile (1) | 43.08 | 0.22 | 0.5 | 466.87 | 239.95
SharedArrays (1) | 40.55 | 1.43 | 3.5 | 1756.68 | 268.86
stress: Test Failed at /build/julia/src/julia/test/stress.jl:98
Expression: begin
ccall(:kill, Cvoid, (Cint, Cint), getpid(), 2)
for i = 1:10
Libc.systemsleep(0.1)
ccall(:jl_gc_safepoint, Cvoid, ())
end
end
Expected: InterruptException
No exception thrown
Stacktrace:
[1] top-level scope at /build/julia/src/julia/test/stress.jl:98
[2] include at ./boot.jl:317 [inlined]
[3] include_relative(::Module, ::String) at ./loading.jl:1038
[4] include at ./sysimg.jl:29 [inlined]
[5] include(::String) at /build/julia/src/julia/test/testdefs.jl:13
[6] macro expansion at /build/julia/src/julia/test/testdefs.jl:22 [inlined]
[7] macro expansion at /build/julia/src/julia/usr/share/julia/stdlib/v1.0/Test/src/Test.jl:1083 [inlined]
[8] macro expansion at /build/julia/src/julia/test/testdefs.jl:21 [inlined]
[9] macro expansion at ./util.jl:289 [inlined]
[10] top-level scope at /build/julia/src/julia/test/testdefs.jl:19 [inlined]
[11] top-level scope at ./none:0
stress: Error During Test at none:1
Test threw exception
Expression: stress
Some tests did not pass: 118 passed, 1 failed, 0 errored, 0 broken.
Test Summary: | Pass Error Broken Total
Overall | 37408366 1 327816 37736183
compiler/validation | 26 26
strings/search | 522 522
strings/io | 12749 12749
unicode/utf8 | 19 19
strings/util | 449 449
worlds | 64 64
strings/types | 2302688 2302688
compiler/compiler | 545 1 546
keywordargs | 127 127
triplequote | 28 28
ambiguous | 62 2 64
intrinsics | 34 34
char | 1471 1471
strings/basic | 87281 87281
iobuffer | 200 200
staged | 59 59
hashing | 18762 18762
subtype | 337379 15 337394
tuple | 483 483
dict | 144314 144314
reduce | 292 292
intfuncs | 4386 4386
simdloop | 248 248
vecelement | 533 533
core | 8423072 8423072
offsetarray | 324 324
copy | 512 512
rational | 97519 97519
reducedim | 673 673
fastmath | 811 811
functional | 95 95
numbers | 1380107 1380107
operators | 12922 12922
path | 233 12 245
parse | 10291 10291
abstractarray | 3972 3972
math | 1505744 1505744
bigfloat | 32 32
loading | 156014 156014
ccall | 5244 5244
iterators | 1497 1497
bigint | 2061 2061
backtrace | 25 1 26
version | 2464 2464
file | 781 781
namedtuple | 154 1 155
arrayops | 1766 1766
spawn | 130 4 134
floatapprox | 49 49
mpfr | 760 760
regex | 28 28
reflection | 330 330
float16 | 124 124
sysinfo | 2 2
combinatorics | 98 98
env | 56 56
complex | 8250 2 8252
mod2pi | 80 80
sorting | 4864 4864
euler | 12 12
rounding | 880 880
read | 3085 3085
goto | 19 19
llvmcall | 13 13
llvmcall2 | 6 6
grisu | 683 683
errorshow | 172 172
some | 62 62
meta | 55 55
stacktraces | 43 43
sets | 736 1 737
docs | 197 197
enums | 84 84
broadcast | 422 422
threads | 190579 190579
misc | 1279371 1279371
int | 10727 10727
bitarray | 898011 898011
inline | 13 13
bitset | 186 186
show | 329 329
cartesian | 7 7
checked | 1219 1219
osutils | 26 26
error | 30 30
iostream | 27 27
specificity | 126 126
ranges | 12108035 327680 12435715
boundscheck | No tests
secretbuffer | 12 12
syntax | 579 1 580
codegen | 94 94
floatfuncs | 134 134
channels | 182 182
logging | 115 115
reinterpretarray | 114 114
asyncmap | 292 292
missing | 403 1 404
subarray | 280 280
Pkg/resolve | 182 182
cmdlineargs | 206 206
LinearAlgebra/schur | 390 390
LinearAlgebra/qr | 3118 3118
LinearAlgebra/matmul | 597 597
LinearAlgebra/special | 947 947
SparseArrays/higherorderfns | 7073 72 7145
LinearAlgebra/eigen | 406 406
LinearAlgebra/bunchkaufman | 5145 5145
LinearAlgebra/svd | 412 412
SparseArrays/sparse | 2144 2144
LinearAlgebra/tridiag | 1159 1159
LinearAlgebra/lapack | 780 780
SparseArrays/sparsevector | 9919 9919
LinearAlgebra/bidiag | 1750 1750
LinearAlgebra/dense | 7467 2 7469
LinearAlgebra/generic | 408 408
LinearAlgebra/uniformscaling | 313 313
LinearAlgebra/hessenberg | 40 40
LinearAlgebra/adjtrans | 252 252
LinearAlgebra/lu | 1191 1191
LinearAlgebra/cholesky | 2186 2186
LinearAlgebra/blas | 628 628
LinearAlgebra/givens | 1840 1840
LinearAlgebra/pinv | 288 288
LinearAlgebra/lq | 1253 1253
Dates/query | 988 988
Dates/adjusters | 3147 3147
LinearAlgebra/symmetric | 1966 1966
LinearAlgebra/structuredbroadcast | 408 408
Dates/types | 167 167
Dates/rounding | 296 296
Dates/ranges | 349123 349123
Dates/periods | 681 681
Dates/conversions | 160 160
CRC32c | 664 664
Base64 | 1015 1015
LinearAlgebra/diagonal | 1531 1531
Dates/arithmetic | 312 312
Dates/io | 258 258
DelimitedFiles | 81 81
Libdl | 63 1 64
Future | 33 33
Logging | 34 34
Markdown | 231 231
InteractiveUtils | 104 104
Dates/accessors | 7723858 7723858
FileWatching | 510 510
Profile | 10 10
Mmap | 131 131
Serialization | 106 106
Printf | 738 738
REPL | 984 5 989
Random | 203068 203068
UUIDs | 10 10
Unicode | 752 752
Statistics | 574 574
Test | 285 14 299
SuiteSparse | 759 759
LibGit2/libgit2 | 666 1 667
Distributed | No tests
SHA | 1 1
LinearAlgebra/triangular | 33194 33194
precompile | 89 89
SharedArrays | 110 110
stress | 1 1
FAILURE
The global RNG seed was 0xa3d406745cd6ad08ccaffb2bb9e1ccea.
Error in testset stress:
Error During Test at none:1
Test threw exception
Expression: stress
Some tests did not pass: 118 passed, 1 failed, 0 errored, 0 broken.
ERROR: LoadError: Test run finished with errors
in expression starting at /build/julia/src/julia/test/runtests.jl:61
mpfr:
Test (Worker) | Time (s) | GC (s) | GC % | Alloc (MB) | RSS (MB)
compiler/validation (4) | 9.52 | 0.36 | 3.8 | 154.03 | 215.82
strings/search (7) | 14.52 | 0.44 | 3.1 | 246.68 | 215.82
unicode/utf8 (7) | 1.09 | 0.00 | 0.0 | 2.25 | 215.82
strings/io (9) | 17.36 | 0.54 | 3.1 | 257.39 | 221.89
From worker 2: Skipping Base.Filesystem.JL_O_RANDOM
From worker 2: Skipping Base.Filesystem.JL_O_SEQUENTIAL
From worker 2: Skipping Base.Filesystem.JL_O_SHORT_LIVED
From worker 2: Skipping Base.Filesystem.JL_O_TEMPORARY
From worker 2: Skipping Base.active_repl
From worker 2: Skipping Base.active_repl_backend
worlds (9) | 9.27 | 0.19 | 2.0 | 132.57 | 223.74
strings/util (8) | 27.57 | 0.59 | 2.1 | 309.89 | 222.52
strings/types (4) | 19.64 | 0.61 | 3.1 | 1183.92 | 216.50
From worker 2: Skipping Base.active_repl
From worker 2: Skipping Base.active_repl_backend
From worker 2: Skipping Base.Filesystem.JL_O_RANDOM
From worker 2: Skipping Base.Filesystem.JL_O_SEQUENTIAL
From worker 2: Skipping Base.Filesystem.JL_O_SHORT_LIVED
From worker 2: Skipping Base.Filesystem.JL_O_TEMPORARY
From worker 2: Skipping Base.active_repl
From worker 2: Skipping Base.active_repl_backend
compiler/compiler (3) | 40.73 | 1.16 | 2.9 | 690.09 | 241.13
ambiguous (2) | 41.32 | 0.95 | 2.3 | 1153.28 | 241.18
triplequote (2) | 0.15 | 0.00 | 0.0 | 1.66 | 241.43
intrinsics (2) | 1.41 | 0.01 | 0.4 | 23.02 | 241.95
keywordargs (9) | 17.57 | 0.11 | 0.6 | 111.47 | 226.97
char (3) | 12.67 | 0.63 | 5.0 | 227.22 | 254.43
strings/basic (6) | 58.43 | 0.93 | 1.6 | 667.02 | 243.43
iobuffer (3) | 14.67 | 0.19 | 1.3 | 117.25 | 269.30
staged (6) | 15.98 | 0.29 | 1.8 | 266.15 | 244.63
hashing (9) | 84.08 | 1.87 | 2.2 | 1445.07 | 261.82
tuple (9) | 24.98 | 0.53 | 2.1 | 264.26 | 277.89
subtype (4) | 132.49 | 3.09 | 2.3 | 1797.38 | 694.62
dict (2) | 133.49 | 3.99 | 3.0 | 1970.74 | 302.38
reduce (9) | 62.72 | 2.35 | 3.7 | 1215.56 | 432.69
intfuncs (9) | 15.26 | 0.18 | 1.2 | 132.14 | 432.69
simdloop (9) | 12.03 | 0.45 | 3.8 | 261.24 | 432.69
core (7) | 253.20 | 23.66 | 9.3 | 10726.55 | 618.90
vecelement (9) | 47.51 | 1.65 | 3.5 | 1325.62 | 432.69
offsetarray (3) | 222.78 | 5.97 | 2.7 | 3356.87 | 349.89
copy (3) | 9.81 | 0.29 | 3.0 | 158.77 | 354.37
rational (7) | 39.08 | 1.45 | 3.7 | 883.32 | 618.90
reducedim (4) | 165.27 | 5.61 | 3.4 | 3315.79 | 694.62
fastmath (7) | 32.50 | 0.94 | 2.9 | 516.64 | 618.90
functional (4) | 34.12 | 1.01 | 3.0 | 637.04 | 694.62
operators (4) | 25.96 | 1.19 | 4.6 | 527.16 | 694.62
numbers (8) | 364.72 | 6.24 | 1.7 | 5053.62 | 388.96
path (4) | 11.51 | 0.06 | 0.5 | 35.58 | 694.62
parse (4) | 18.90 | 0.49 | 2.6 | 435.20 | 694.62
abstractarray (2) | 281.42 | 11.13 | 4.0 | 5278.09 | 405.94
ccall (8) | 120.58 | 82.56 | 68.5 | 781.08 | 404.97
math (3) | 203.76 | 3.71 | 1.8 | 3008.74 | 414.69
bigfloat (8) | 0.97 | 0.05 | 5.3 | 14.28 | 404.97
iterators (7) | 174.61 | 8.43 | 4.8 | 3839.43 | 618.90
bigint (2) | 24.38 | 0.65 | 2.7 | 310.37 | 419.30
From worker 8: [stdio passthrough ok]
backtrace (7) | 4.29 | 0.06 | 1.4 | 33.20 | 618.90
From worker 4: WARNING: replacing module Foo.
loading (4) | 108.50 | 5.04 | 4.6 | 3567.07 | 694.62
version (4) | 4.24 | 0.14 | 3.4 | 98.13 | 694.62
file (2) | 20.55 | 0.66 | 3.2 | 358.37 | 423.94
namedtuple (4) | 13.84 | 0.38 | 2.8 | 210.90 | 694.62
arrayops (6) | 480.10 | 17.12 | 3.6 | 9209.84 | 536.76
spawn (8) | 45.58 | 0.76 | 1.7 | 600.03 | 418.00
floatapprox (8) | 6.70 | 0.18 | 2.7 | 153.87 | 418.99
reflection (8) | 29.19 | 1.01 | 3.5 | 469.78 | 434.38
regex (8) | 2.99 | 0.11 | 3.7 | 36.22 | 434.63
float16 (8) | 8.85 | 0.22 | 2.5 | 109.87 | 441.00
combinatorics (8) | 5.21 | 0.09 | 1.7 | 53.56 | 443.35
sysinfo (8) | 1.14 | 0.06 | 5.3 | 14.99 | 443.63
mpfr (2) | 81.85 | 1.78 | 2.2 | 1285.43 | 472.34
env (8) | 3.86 | 0.04 | 1.0 | 19.65 | 443.79
read (7) | 126.78 | 7.11 | 5.6 | 3178.49 | 759.98
mod2pi (7) | 4.00 | 0.24 | 5.9 | 32.74 | 760.79
complex (6) | 82.11 | 2.01 | 2.4 | 853.82 | 536.76
rounding (2) | 34.20 | 1.21 | 3.5 | 526.32 | 473.41
euler (7) | 6.41 | 0.58 | 9.0 | 133.48 | 768.95
sorting (3) | 157.33 | 4.14 | 2.6 | 2385.30 | 448.07
goto (3) | 0.34 | 0.00 | 0.0 | 1.11 | 448.07
llvmcall (3) | 1.31 | 0.05 | 3.9 | 17.83 | 448.33
llvmcall2 (3) | 0.28 | 0.00 | 0.0 | 2.00 | 448.38
grisu (3) | 14.97 | 0.37 | 2.4 | 185.69 | 450.13
some (3) | 5.46 | 0.14 | 2.5 | 92.86 | 456.32
meta (3) | 6.13 | 0.24 | 3.8 | 92.27 | 456.50
errorshow (2) | 51.45 | 1.86 | 3.6 | 655.74 | 483.05
stacktraces (3) | 14.85 | 0.52 | 3.5 | 240.63 | 462.17
sets (7) | 76.36 | 3.46 | 4.5 | 762.69 | 780.58
From worker 7: Warning: threaded loop executed in order
docs (2) | 41.90 | 1.89 | 4.5 | 652.58 | 496.82
enums (2) | 20.32 | 0.92 | 4.5 | 327.73 | 504.07
broadcast (4) | 235.53 | 11.55 | 4.9 | 4717.70 | 755.91
threads (7) | 64.42 | 4.56 | 7.1 | 238.76 | 791.42
From worker 3: Test Summary: | Pass Total
From worker 3: @deprecate | 6 6
From worker 3: Test Summary: | Pass Total
From worker 3: firstcaller | 1 1
From worker 3: Test Summary: | Broken Total
From worker 3: parser syntax deprecations | 1 1
misc (3) | 98.72 | 6.97 | 7.1 | 1804.86 | 559.47
int (4) | 38.84 | 0.62 | 1.6 | 236.55 | 778.24
bitset (3) | 18.34 | 0.28 | 1.5 | 73.20 | 559.47
inline (3) | 2.45 | 0.10 | 4.0 | 30.99 | 559.47
ranges (8) | 217.98 | 14.79 | 6.8 | 8552.28 | 524.17
checked (7) | 43.11 | 0.26 | 0.6 | 78.18 | 828.18
cartesian (7) | 0.73 | 0.00 | 0.0 | 8.99 | 828.25
show (6) | 195.05 | 8.10 | 4.2 | 2808.78 | 571.95
osutils (7) | 1.54 | 0.00 | 0.0 | 0.61 | 828.28
iostream (7) | 3.52 | 0.32 | 9.1 | 67.13 | 828.42
error (8) | 6.50 | 0.12 | 1.9 | 72.62 | 525.46
specificity (8) | 0.40 | 0.00 | 0.0 | 4.71 | 525.46
floatfuncs (4) | 34.54 | 1.54 | 4.5 | 662.73 | 795.75
secretbuffer (7) | 3.57 | 1.19 | 33.2 | 32.93 | 828.46
boundscheck (3) | 21.20 | 0.00 | 0.0 | 0.90 | 559.47
codegen (8) | 8.78 | 1.08 | 12.3 | 26.83 | 525.46
syntax (7) | 8.74 | 0.24 | 2.7 | 100.64 | 828.46
channels (6) | 29.10 | 3.65 | 12.6 | 555.13 | 655.88
bitarray (9) | 592.55 | 27.95 | 4.7 | 12450.13 | 583.75
logging (3) | 31.76 | 0.93 | 2.9 | 457.77 | 559.47
reinterpretarray (4) | 35.22 | 2.21 | 6.3 | 690.14 | 813.97
asyncmap (7) | 37.39 | 1.70 | 4.6 | 710.01 | 828.46
missing (8) | 46.93 | 1.91 | 4.1 | 679.52 | 528.38
subarray (5) | 1014.56 | 62.94 | 6.2 | 24177.84 | 734.57
Pkg/resolve (4) | 130.51 | 14.70 | 11.3 | 6311.65 | 874.10
cmdlineargs (2) | 413.00 | 1.29 | 0.3 | 274.76 | 531.19
LinearAlgebra/schur (2) | 135.49 | 6.26 | 4.6 | 2397.13 | 548.52
LinearAlgebra/matmul (4) | 318.65 | 18.89 | 5.9 | 5171.46 | 941.84
LinearAlgebra/qr (8) | 462.25 | 27.21 | 5.9 | 10351.28 | 757.92
LinearAlgebra/special (2) | 111.84 | 4.44 | 4.0 | 1700.68 | 597.65
SparseArrays/higherorderfns (6) | 622.89 | 36.50 | 5.9 | 12283.17 | 756.53
LinearAlgebra/bunchkaufman (8) | 146.89 | 8.66 | 5.9 | 2726.68 | 805.26
LinearAlgebra/eigen (4) | 190.79 | 11.62 | 6.1 | 3576.19 | 981.35
SparseArrays/sparse (9) | 759.78 | 151.64 | 20.0 | 10134.73 | 828.83
LinearAlgebra/svd (2) | 203.91 | 12.97 | 6.4 | 4052.89 | 672.47
LinearAlgebra/lapack (6) | 197.11 | 12.01 | 6.1 | 3678.70 | 839.48
LinearAlgebra/tridiag (8) | 202.19 | 8.33 | 4.1 | 2922.24 | 866.04
SparseArrays/sparsevector (3) | 858.32 | 39.28 | 4.6 | 17877.06 | 932.37
LinearAlgebra/bidiag (4) | 263.44 | 15.53 | 5.9 | 4262.74 | 1089.36
LinearAlgebra/dense (5) | 789.77 | 68.77 | 8.7 | 15170.40 | 1061.60
LinearAlgebra/generic (3) | 78.97 | 4.72 | 6.0 | 1668.48 | 948.98
LinearAlgebra/hessenberg (3) | 37.56 | 2.41 | 6.4 | 508.19 | 952.35
LinearAlgebra/uniformscaling (4) | 91.78 | 5.37 | 5.8 | 1274.78 | 1125.23
LinearAlgebra/adjtrans (4) | 117.13 | 8.32 | 7.1 | 1640.48 | 1143.37
LinearAlgebra/cholesky (2) | 412.22 | 24.67 | 6.0 | 6204.20 | 795.05
LinearAlgebra/pinv (4) | 54.11 | 4.26 | 7.9 | 1186.28 | 1247.96
LinearAlgebra/blas (3) | 209.65 | 11.40 | 5.4 | 2640.84 | 986.56
LinearAlgebra/givens (2) | 21.15 | 1.01 | 4.8 | 355.31 | 806.00
LinearAlgebra/lq (5) | 297.26 | 27.93 | 9.4 | 6268.08 | 1238.22
LinearAlgebra/lu (6) | 412.54 | 28.86 | 7.0 | 6821.57 | 1042.70
LinearAlgebra/symmetric (8) | 391.84 | 26.09 | 6.7 | 5038.90 | 956.71
Dates/query (6) | 6.28 | 0.08 | 1.3 | 23.19 | 1042.98
Dates/adjusters (5) | 25.08 | 1.08 | 4.3 | 251.25 | 1262.91
Dates/rounding (5) | 11.44 | 0.50 | 4.4 | 111.32 | 1270.69
Dates/types (5) | 9.95 | 0.36 | 3.6 | 81.91 | 1277.62
LinearAlgebra/structuredbroadcast (4) | 107.30 | 7.98 | 7.4 | 2232.33 | 1247.96
Dates/ranges (6) | 60.51 | 3.77 | 6.2 | 1129.78 | 1055.29
Dates/periods (8) | 66.44 | 3.25 | 4.9 | 926.75 | 984.34
Dates/conversions (6) | 6.62 | 0.14 | 2.2 | 56.88 | 1060.83
CRC32c (6) | 3.83 | 0.28 | 7.3 | 75.97 | 1062.58
Base64 (8) | 8.52 | 1.37 | 16.1 | 178.66 | 984.42
Dates/io (5) | 64.81 | 3.69 | 5.7 | 863.01 | 1348.06
Dates/arithmetic (4) | 48.44 | 2.66 | 5.5 | 672.17 | 1247.96
DelimitedFiles (6) | 33.36 | 2.04 | 6.1 | 463.72 | 1071.77
Future (4) | 8.43 | 1.20 | 14.3 | 86.82 | 1247.96
Libdl (4) | 3.01 | 0.10 | 3.4 | 25.67 | 1247.96
Logging (4) | 17.84 | 1.41 | 7.9 | 359.76 | 1247.96
FileWatching (5) | 59.57 | 14.68 | 24.6 | 179.44 | 1348.06
InteractiveUtils (6) | 44.98 | 2.69 | 6.0 | 625.29 | 1127.43
Markdown (4) | 28.98 | 2.14 | 7.4 | 492.88 | 1247.96
LinearAlgebra/diagonal (9) | 648.36 | 51.75 | 8.0 | 10777.88 | 1031.84
Profile (4) | 21.08 | 1.64 | 7.8 | 312.66 | 1250.49
Dates/accessors (2) | 218.61 | 9.18 | 4.2 | 3768.73 | 815.97
From worker 3: The following 'Resetting the helper list...' warning is expected:
┌ Warning: Resetting the helper list is currently unsupported:
│ ignoring all git credential helpers
└ @ LibGit2 ~/julia/src/julia/usr/share/julia/stdlib/v1.0/LibGit2/src/gitcredential.jl:229
Mmap (5) | 51.83 | 38.68 | 74.6 | 275.21 | 1348.06
Serialization (5) | 30.09 | 3.48 | 11.6 | 753.64 | 1348.06
Printf (6) | 105.35 | 9.32 | 8.8 | 2958.97 | 1140.58
┌ Warning: TerminalMenus: Unable to enter raw mode: ArgumentError("stream is closed or unusable")
└ @ REPL.TerminalMenus ~/julia/src/julia/usr/share/julia/stdlib/v1.0/REPL/src/TerminalMenus/util.jl:21
From worker 9:
┌ Warning: TerminalMenus: Unable to enter raw mode: ArgumentError("stream is closed or unusable")
└ @ REPL.TerminalMenus ~/julia/src/julia/usr/share/julia/stdlib/v1.0/REPL/src/TerminalMenus/util.jl:21
From worker 9:
REPL (9) | 117.92 | 10.08 | 8.5 | 2210.09 | 1101.95
Random (4) | 147.27 | 10.95 | 7.4 | 2706.87 | 1368.16
UUIDs (4) | 0.33 | 0.00 | 0.0 | 6.64 | 1368.18
Unicode (4) | 21.35 | 1.34 | 6.3 | 308.39 | 1381.14
Statistics (5) | 124.03 | 10.94 | 8.8 | 2470.17 | 1348.06
┌ Warning: Omitting `@test_deprecated` tests which can't yet
│ be tested in --depwarn=error mode
└ @ Main.Test97Main_Test ~/julia/src/julia/usr/share/julia/stdlib/v1.0/Test/test/runtests.jl:736
Test (9) | 119.12 | 1.63 | 1.4 | 417.78 | 1106.88
SuiteSparse (6) | 211.72 | 13.68 | 6.5 | 3369.41 | 1225.41
LibGit2/libgit2 (3) | 520.40 | 8.16 | 1.6 | 1488.83 | 1369.19
Distributed (8) | 403.80 | 0.05 | 0.0 | 28.14 | 984.42
SHA (2) | 319.92 | 1.74 | 0.5 | 687.32 | 867.99
LinearAlgebra/triangular (7) | 1773.89 | 111.24 | 6.3 | 35336.51 | 1502.91
Executing tests that run on node 1 only:
precompile (1) | 38.24 | 0.22 | 0.6 | 454.97 | 239.77
SharedArrays (1) | 33.86 | 1.30 | 3.8 | 1756.70 | 270.38
stress: Test Failed at /build/julia/src/julia/test/stress.jl:98
Expression: begin
ccall(:kill, Cvoid, (Cint, Cint), getpid(), 2)
for i = 1:10
Libc.systemsleep(0.1)
ccall(:jl_gc_safepoint, Cvoid, ())
end
end
Expected: InterruptException
No exception thrown
Stacktrace:
[1] top-level scope at /build/julia/src/julia/test/stress.jl:98
[2] include at ./boot.jl:317 [inlined]
[3] include_relative(::Module, ::String) at ./loading.jl:1038
[4] include at ./sysimg.jl:29 [inlined]
[5] include(::String) at /build/julia/src/julia/test/testdefs.jl:13
[6] macro expansion at /build/julia/src/julia/test/testdefs.jl:22 [inlined]
[7] macro expansion at /build/julia/src/julia/usr/share/julia/stdlib/v1.0/Test/src/Test.jl:1083 [inlined]
[8] macro expansion at /build/julia/src/julia/test/testdefs.jl:21 [inlined]
[9] macro expansion at ./util.jl:289 [inlined]
[10] top-level scope at /build/julia/src/julia/test/testdefs.jl:19 [inlined]
[11] top-level scope at ./none:0
stress: Error During Test at none:1
Test threw exception
Expression: stress
Some tests did not pass: 118 passed, 1 failed, 0 errored, 0 broken.
Test Summary: | Pass Error Broken Total
Overall | 37391103 1 327816 37718920
compiler/validation | 26 26
strings/search | 522 522
unicode/utf8 | 19 19
strings/io | 12749 12749
worlds | 64 64
strings/util | 449 449
strings/types | 2302688 2302688
compiler/compiler | 545 1 546
ambiguous | 62 2 64
triplequote | 28 28
intrinsics | 34 34
keywordargs | 127 127
char | 1471 1471
strings/basic | 87281 87281
iobuffer | 200 200
staged | 59 59
hashing | 18762 18762
tuple | 483 483
subtype | 337379 15 337394
dict | 144314 144314
reduce | 292 292
intfuncs | 4386 4386
simdloop | 248 248
core | 8423072 8423072
vecelement | 533 533
offsetarray | 324 324
copy | 512 512
rational | 97519 97519
reducedim | 673 673
fastmath | 811 811
functional | 95 95
operators | 12922 12922
numbers | 1380107 1380107
path | 233 12 245
parse | 10291 10291
abstractarray | 2752 2752
ccall | 5244 5244
math | 1505744 1505744
bigfloat | 32 32
iterators | 1497 1497
bigint | 2061 2061
backtrace | 25 1 26
loading | 141854 141854
version | 2464 2464
file | 781 781
namedtuple | 154 1 155
arrayops | 1766 1766
spawn | 130 4 134
floatapprox | 49 49
reflection | 330 330
regex | 28 28
float16 | 124 124
combinatorics | 98 98
sysinfo | 2 2
mpfr | 760 760
env | 56 56
read | 3085 3085
mod2pi | 80 80
complex | 8250 2 8252
rounding | 880 880
euler | 12 12
sorting | 4864 4864
goto | 19 19
llvmcall | 13 13
llvmcall2 | 6 6
grisu | 683 683
some | 62 62
meta | 55 55
errorshow | 172 172
stacktraces | 43 43
sets | 736 1 737
docs | 197 197
enums | 84 84
broadcast | 422 422
threads | 190579 190579
misc | 1279371 1279371
int | 10727 10727
bitset | 186 186
inline | 13 13
ranges | 12108108 327680 12435788
checked | 1219 1219
cartesian | 7 7
show | 329 329
osutils | 26 26
iostream | 27 27
error | 30 30
specificity | 126 126
floatfuncs | 134 134
secretbuffer | 12 12
boundscheck | No tests
codegen | 94 94
syntax | 579 1 580
channels | 182 182
bitarray | 896055 896055
logging | 115 115
reinterpretarray | 114 114
asyncmap | 292 292
missing | 403 1 404
subarray | 280 280
Pkg/resolve | 182 182
cmdlineargs | 206 206
LinearAlgebra/schur | 390 390
LinearAlgebra/matmul | 597 597
LinearAlgebra/qr | 3118 3118
LinearAlgebra/special | 947 947
SparseArrays/higherorderfns | 7073 72 7145
LinearAlgebra/bunchkaufman | 5145 5145
LinearAlgebra/eigen | 406 406
SparseArrays/sparse | 2144 2144
LinearAlgebra/svd | 412 412
LinearAlgebra/lapack | 780 780
LinearAlgebra/tridiag | 1159 1159
SparseArrays/sparsevector | 9919 9919
LinearAlgebra/bidiag | 1750 1750
LinearAlgebra/dense | 7467 2 7469
LinearAlgebra/generic | 408 408
LinearAlgebra/hessenberg | 40 40
LinearAlgebra/uniformscaling | 313 313
LinearAlgebra/adjtrans | 252 252
LinearAlgebra/cholesky | 2186 2186
LinearAlgebra/pinv | 288 288
LinearAlgebra/blas | 628 628
LinearAlgebra/givens | 1840 1840
LinearAlgebra/lq | 1253 1253
LinearAlgebra/lu | 1191 1191
LinearAlgebra/symmetric | 1966 1966
Dates/query | 988 988
Dates/adjusters | 3147 3147
Dates/rounding | 296 296
Dates/types | 167 167
LinearAlgebra/structuredbroadcast | 408 408
Dates/ranges | 349123 349123
Dates/periods | 681 681
Dates/conversions | 160 160
CRC32c | 664 664
Base64 | 1015 1015
Dates/io | 258 258
Dates/arithmetic | 312 312
DelimitedFiles | 81 81
Future | 33 33
Libdl | 63 1 64
Logging | 34 34
FileWatching | 510 510
InteractiveUtils | 104 104
Markdown | 231 231
LinearAlgebra/diagonal | 1531 1531
Profile | 10 10
Dates/accessors | 7723858 7723858
Mmap | 131 131
Serialization | 106 106
Printf | 738 738
REPL | 984 5 989
Random | 203068 203068
UUIDs | 10 10
Unicode | 752 752
Statistics | 574 574
Test | 285 14 299
SuiteSparse | 759 759
LibGit2/libgit2 | 666 1 667
Distributed | No tests
SHA | 1 1
LinearAlgebra/triangular | 33194 33194
precompile | 89 89
SharedArrays | 110 110
stress | 1 1
FAILURE
The global RNG seed was 0x8130efd3ea41bc2925fc76477fec2dd0.
Error in testset stress:
Error During Test at none:1
Test threw exception
Expression: stress
Some tests did not pass: 118 passed, 1 failed, 0 errored, 0 broken.
ERROR: LoadError: Test run finished with errors
in expression starting at /build/julia/src/julia/test/runtests.jl:61
libm fails with a highly uninformative internal make error:
CC src/llvm-api.o
LINK usr/lib/libjulia.so.1.0
make[1]: Leaving directory '/build/julia/src/julia/src'
make[1]: Entering directory '/build/julia/src/julia/ui'
Warning: git information unavailable; versioning information limited
CC ui/repl.o
LINK usr/bin/julia
make[1]: Leaving directory '/build/julia/src/julia/ui'
Creating usr/etc/julia/startup.jl
Copying in usr/share/man/man1/julia.1
/build/julia/src/julia/contrib/install.sh 755 /build/julia/src/julia/contrib/julia-config.jl /build/julia/src/julia/usr/share/julia/
make[1]: Entering directory '/build/julia/src/julia/base'
Warning: git information unavailable; versioning information limited
PERL base/pcre_h.jl
PERL base/errno_h.jl
PERL base/build_h.jl.phony
PERL base/file_constants.jl
PERL base/uv_constants.jl
make[1]: *** [Makefile:168: /build/julia/src/julia/usr/lib/julia/libm.so] Error 1
make[1]: Leaving directory '/build/julia/src/julia/base'
make: *** [Makefile:63: julia-base] Error 2
make: Leaving directory '/build/julia/src/julia'
libunwind fails due to trying to do integer math on something the system version defines as a floating-point number (but if I locally patch out the whole if (!defined...) mess and unconditionally set JL_UNW_HAS_FORMAT_IP, it works fine and even passes the testsuite!):
make[1]: Entering directory '/build/julia/src/julia/src'
Warning: git information unavailable; versioning information limited
CC src/jltypes.o
CC src/gf.o
CC src/ast.o
CC src/typemap.o
CC src/builtins.o
CC src/module.o
CC src/interpreter.o
CC src/symbol.o
CC src/dlload.o
In file included from /usr/include/libunwind-x86_64.h:129,
from /usr/include/libunwind.h:25,
from /build/julia/src/julia/src/julia_internal.h:623,
from /build/julia/src/julia/src/module.c:7:
/build/julia/src/julia/src/julia_internal.h:627:35: error: floating constant in preprocessor expression
(UNW_VERSION_MAJOR == 1 && UNW_VERSION_MINOR > 1))
^~~~~~~~~~~~~~~~~
In file included from /usr/include/libunwind-x86_64.h:129,
from /usr/include/libunwind.h:25,
from /build/julia/src/julia/src/julia_internal.h:623,
from /build/julia/src/julia/src/gf.c:14:
/build/julia/src/julia/src/julia_internal.h:627:35: error: floating constant in preprocessor expression
(UNW_VERSION_MAJOR == 1 && UNW_VERSION_MINOR > 1))
^~~~~~~~~~~~~~~~~
In file included from /usr/include/libunwind-x86_64.h:129,
from /usr/include/libunwind.h:25,
from /build/julia/src/julia/src/julia_internal.h:623,
from /build/julia/src/julia/src/typemap.c:6:
/build/julia/src/julia/src/julia_internal.h:627:35: error: floating constant in preprocessor expression
(UNW_VERSION_MAJOR == 1 && UNW_VERSION_MINOR > 1))
^~~~~~~~~~~~~~~~~
In file included from /usr/include/libunwind-x86_64.h:129,
from /usr/include/libunwind.h:25,
from /build/julia/src/julia/src/julia_internal.h:623,
from /build/julia/src/julia/src/interpreter.c:9:
/build/julia/src/julia/src/julia_internal.h:627:35: error: floating constant in preprocessor expression
(UNW_VERSION_MAJOR == 1 && UNW_VERSION_MINOR > 1))
^~~~~~~~~~~~~~~~~
In file included from /usr/include/libunwind-x86_64.h:129,
from /usr/include/libunwind.h:25,
from /build/julia/src/julia/src/julia_internal.h:623,
from /build/julia/src/julia/src/ast.c:14:
/build/julia/src/julia/src/julia_internal.h:627:35: error: floating constant in preprocessor expression
(UNW_VERSION_MAJOR == 1 && UNW_VERSION_MINOR > 1))
^~~~~~~~~~~~~~~~~
In file included from /usr/include/libunwind-x86_64.h:129,
from /usr/include/libunwind.h:25,
from /build/julia/src/julia/src/julia_internal.h:623,
from /build/julia/src/julia/src/builtins.c:24:
/build/julia/src/julia/src/julia_internal.h:627:35: error: floating constant in preprocessor expression
(UNW_VERSION_MAJOR == 1 && UNW_VERSION_MINOR > 1))
^~~~~~~~~~~~~~~~~
make[1]: *** [Makefile:135: module.o] Error 1
make[1]: *** Waiting for unfinished jobs....
In file included from /usr/include/libunwind-x86_64.h:129,
from /usr/include/libunwind.h:25,
from /build/julia/src/julia/src/julia_internal.h:623,
from /build/julia/src/julia/src/symbol.c:11:
/build/julia/src/julia/src/julia_internal.h:627:35: error: floating constant in preprocessor expression
(UNW_VERSION_MAJOR == 1 && UNW_VERSION_MINOR > 1))
^~~~~~~~~~~~~~~~~
make[1]: *** [Makefile:135: typemap.o] Error 1
make[1]: *** [Makefile:135: interpreter.o] Error 1
In file included from /usr/include/libunwind-x86_64.h:129,
from /usr/include/libunwind.h:25,
from /build/julia/src/julia/src/julia_internal.h:623,
from /build/julia/src/julia/src/jltypes.c:14:
/build/julia/src/julia/src/julia_internal.h:627:35: error: floating constant in preprocessor expression
(UNW_VERSION_MAJOR == 1 && UNW_VERSION_MINOR > 1))
^~~~~~~~~~~~~~~~~
make[1]: *** [Makefile:135: symbol.o] Error 1
make[1]: *** [Makefile:135: gf.o] Error 1
make[1]: *** [Makefile:135: builtins.o] Error 1
make[1]: *** [Makefile:135: jltypes.o] Error 1
In file included from /usr/include/libunwind-x86_64.h:129,
from /usr/include/libunwind.h:25,
from /build/julia/src/julia/src/julia_internal.h:623,
from /build/julia/src/julia/src/dlload.c:10:
/build/julia/src/julia/src/julia_internal.h:627:35: error: floating constant in preprocessor expression
(UNW_VERSION_MAJOR == 1 && UNW_VERSION_MINOR > 1))
^~~~~~~~~~~~~~~~~
make[1]: *** [Makefile:135: dlload.o] Error 1
make[1]: *** [Makefile:135: ast.o] Error 1
make[1]: Leaving directory '/build/julia/src/julia/src'
make: *** [Makefile:69: julia-src-release] Error 2
make: Leaving directory '/build/julia/src/julia'
$ grep UNW_VERSION_MINOR /usr/include/libunwind-common.h
#define UNW_VERSION_MINOR 2.1
openblas and lapack link fine, but fail the testsuite:
Same results with just system openblas:
openblas-0.3.3-no-lapack-fail.log
LLVM of course fails dismally with:
JULIA usr/lib/julia/sys-o.a
Generating precompile statements... 762 generated in 32.810796 seconds
Illegal inttoptr
%scevgep9 = ptrtoint i32 addrspace(13)* %scevgep to i64
Illegal inttoptr
%scevgep1011 = ptrtoint i32 addrspace(13)* %scevgep10 to i64
signal (6): Aborted
in expression starting at no file:0
gsignal at /usr/lib/libc.so.6 (unknown line)
abort at /usr/lib/libc.so.6 (unknown line)
runOnFunction at /build/julia/src/julia/src/llvm-gc-invariant-verifier.cpp:178
_ZN4llvm13FPPassManager13runOnFunctionERNS_8FunctionE at /usr/lib/libLLVM-6.0.so (unknown line)
_ZN4llvm13FPPassManager11runOnModuleERNS_6ModuleE at /usr/lib/libLLVM-6.0.so (unknown line)
_ZN4llvm6legacy15PassManagerImpl3runERNS_6ModuleE at /usr/lib/libLLVM-6.0.so (unknown line)
operator() at /build/julia/src/julia/src/jitlayers.cpp:1182 [inlined]
jl_dump_native at /build/julia/src/julia/src/jitlayers.cpp:1191
jl_write_compiler_output at /build/julia/src/julia/src/precompile.c:84
jl_atexit_hook at /build/julia/src/julia/src/init.c:233
main at /build/julia/src/julia/ui/repl.c:234
__libc_start_main at /usr/lib/libc.so.6 (unknown line)
_start at /build/julia/src/julia/usr/bin/julia (unknown line)
Allocations: 57622075 (Pool: 57611147; Big: 10928); GC: 126
/bin/sh: line 1: 30109 Aborted (core dumped) /build/julia/src/julia/usr/bin/julia -O3 -C "x86-64" --output-o /build/julia/src/julia/usr/lib/julia/sys-o.a.tmp --startup-file=no --warn-overwrite=yes --sysimage /build/julia/src/julia/usr/lib/julia/sys.ji /build/julia/src/julia/contrib/generate_precompile.jl /build/julia/src/julia/usr/lib/julia/sys.ji
But I'm sure you know and expect that, you have locally patched llvm in order to fix it. What's more alarming to me is that we were confidently assured that julia 1.0 would finally actually work with a system llvm, which has so far failed to be true.
I tried to see if I could apply the patches in question to llvm 6.0.1 but first they didn't apply cleanly, and then attempting to backport related patches for our system llvm package, resulted in llvm failing to build -- at which point I rapidly gave up.
Apparently, julia has no interest any time soon in being remotely easy to compile with a system llvm? Whatever happened to "We just have to nominate them" for llvm 6.0.1? I'm no expert at llvm, and our llvm packager is distinctly unimpressed with this whole trainwreck and refuses to "babysit llvm patches for julia".
Is this going to be the same story with llvm 7 once that promised rainbow appears?
eli-schwartz
All 21 comments
The fact that it's almost impossible to build julia with any system dependencies, on top of which shenanigans like https://github.com/JuliaLinearAlgebra/Arpack.jl/issues/5 occur, makes it seem increasingly likely that julia's plan for Linux distribution is to ship a small linux operating system inside a VM, in order to run successfully.
What's next, are you going to bootstrap your own gcc build so that thirdparty ecosystem packages can reliably depend on the private libgfortran.so version?
eli-schwartz
on 17 Sep 2018
Oh, also when setting up the testsuite to actually run, your make test target doesn't let me skip tests, it just runs them all. Which is only problematic because apparently you dont run them all either -- your CI slips the sockets test. I've reimplemented the testsuite that way for now, but it would be nice to know if that's actually expected to happen, and if so, what is the Makefile target for?
eli-schwartz
on 17 Sep 2018
As noted in the README, we require patched versions of our dependencies, because our dependencies have bugs. We do our best to upstream patches, but that takes time, esp if you're only using major releases.
we were confidently assured that julia 1.0 would finally actually work with a system llvm, which has so far failed to be true.
I'm sorry if you felt misled, but we were never in a position to promise this, and to the best of my knowledge we didn't.
Separately, this issue comes across as very ranty. I appreciate that your intention is to package julia for Arch Linux, but comments like
Apparently, julia has no interest any time soon in being remotely easy to compile with a system llvm? [...] Is this going to be the same story with llvm 7 once that promised rainbow appears?
are extremely off putting and make me less likely to want to help you fix these issues.
With respect to https://github.com/JuliaLinearAlgebra/Arpack.jl/issues/5, as mentioned there we're in the process of addressing these issues, but our first priority is getting it right for the generic binaries, second users building from source locally, and third integration with system package managers. We have rough plans for the third, including allowing distro packagers to re-use binaries that they already ship (assuming they are compatible), but we haven't gotten to that yet. On that point, I would ask for a bit of patience.
Keno
on 17 Sep 2018
stress test failure seems strange.
libm build issue could be a real bug.
libunwind and llvm are both upstream bugs so ask them why they haven't got a new release yet.
blas issue appears to be not actually using openblas (I used to get that error before making sure I'm actually loading openblas).
yuyichao
on 17 Sep 2018
are extremely off putting and make me less likely to want to help you fix these issues.
On the other hand, I explicitly receive the strong impression that you all actively find the existence of distro packages to be distasteful and unsupported in the strongest terms. That makes me feel like we're fools (and unwelcome ones at that) for daring to try packaging it.
eli-schwartz
on 17 Sep 2018
stress test failure doesn't really seems strange.
Yes -- I sometimes get nondeterministic failures there, so it takes 3 or 4 times to rebuild the package before it works. But in the case of pcre2 and mpfr it is consistently failing.
libunwind and llvm are both upstream bugs so ask them why they haven't got a new release yet.
Actually this libunwind thing seems to be remarked upon by https://github.com/JuliaLang/julia/pull/29082 (but it's still float comparison, whether you use > or !=)
As for llvm, they did release a new version, which is why I asked:
Whatever happened to "We just have to nominate them" for llvm 6.0.1?
eli-schwartz
on 17 Sep 2018
but it's still float comparison, whether you use > or !=
It's a libunwind bug, whether any attempt is made to workaround it or not. https://github.com/libunwind/libunwind/issues/30
they did release a new version
The point release isn't the issue. We are hitting bugs from them faster than they are making releases.
distasteful and unsupported in the strongest terms
Correct and that is a statement of the fact rather than a statement of what could be done.
Given the bugs and misalignment in release schedule with LLVM, we don't expect being able to include all necessary patches in the latest point release and neither do we expect distributions to be willing to carry all the LLVM patches which necessarily mean that their julia package will likely be broken (and again as a fact, many of them are). Therefore, we have to strongly discourage users from using distro packages at this time after getting many reports about broken features.
Those messages are also only for the users and distro maintainers are still more than welcome to include dependencies that have our bug fix.
yuyichao
on 17 Sep 2018
@eli-schwartz I've been able to package Julia in Fedora recently. It's annoying that I haven't been able to use neither the system's libunwind nor the system's mpfr (not sure why for the latter, need to investigate), but it works. EDIT: of course I don't use the system LLVM either, since the maintainer of that package didn't think applying Julia patches to it was a good idea. I haven't had problems with PCRE2. See https://src.fedoraproject.org/rpms/julia/blob/master/f/julia.spec if that can help.
nalimilan
on 17 Sep 2018
On the other hand, I explicitly receive the strong impression that you all actively find the existence of distro packages to be distasteful and unsupported in the strongest terms. That makes me feel like we're fools (and unwelcome ones at that) for daring to try packaging it.
I wouldn't say distasteful, but your impression that we consider it unsupported iz correct. This is primarily for three reasons:
- Many distros ship broken versions of julia (often because of missing upstream patches, or occasionally additional patches that cause new bugs), which frustrates users and increases support volume.
- Many distros' release schedule takes so long that by the time the packages make it in, the packaged julia version is no longer supported, making those packages a bit useless for a fast moving project like julia (I appreciate that Arch moves faster and this is probably less applicable with 1.0, but still).
- Distro packaging (esp for packages) is incompatible with our reproducibility goals (we want users to be able to go back years later and get the exact same code that they were using when they originally wrote the code).
As a result of this, supporting distro package users is third priority, behind the generic binaries (first priority) and people building from source (second priority). We still have a bunch of work to do to fix those two, but we will get to improved distro packager support eventually.
Keno
on 17 Sep 2018
I'd suggest filing separate issues about the libm and the stress failures. I don't see any PCRE-related failure (the make output mixes unrelated things AFAICT, -j1 should help).
Also, USE_SYSTEM_LIBM=0 USE_SYSTEM_OPENLIBM=1 should work if Arch includes openlibm, and that's the recommended approach to ensure computations give exactly the same results as the official binaries.
nalimilan
on 17 Sep 2018
I'm hopeful that the LLVM patching process will eventually converge and we'll be compatible with some released version of LLVM. If that's all that's required, we might get there. However the problem in the past has been that distros will only let us use whichever LLVM version they picked, making compatibility far less likely. From our perspective, we're being good citizens by fixing bugs and upstreaming the patches. If we could increase compatibility by throwing in a few #ifdefs we would, and have done so in the past. But for most of these issues that is not practical.
JeffBezanson
on 17 Sep 2018
I think at this point distros accept to package multiple LLVM versions, since Julia is not the only package which needs a specific version (e.g. Haskell or Rust also do). But it would be really nice to get all patches upstreamed and included in an LLVM point release, so that we can share a single versioned LLVM package. If the LLVM release process doesn't allow that, we should complain.
nalimilan
on 17 Sep 2018
But it would be really nice to get all patches upstreamed and included in an LLVM point release, so that we can share a single versioned LLVM package. If the LLVM release process doesn't allow that, we should complain.
I'm pretty sure the llvm process does allow that, since I mentioned it in my initial comment here, and linked to https://github.com/JuliaLang/julia/pull/26626#issuecomment-376587822
eli-schwartz
on 30 Sep 2018
Attempting to build julia using the system llvm completely fails for entirely unexpected reasons now that the system llvm6 is not the one pointed to by the /usr/lib/libLLVM.so symlink -- because Arch Linux actually has no problem providing versioned packages when needed as dependencies for things.
eli-schwartz
on 30 Sep 2018
So, it turns out the failures when using the system openblas, are due to the fact that our openblas package is built with NO_CBLAS=1.
This makes sense I think, because we already provide libcblas.so which is obviously linked to libblas.so (and on Arch Linux, openblas is a drop-in replacement for the reference blas, providing a symlink to libblas.so as well as package provides/conflicts markers for blas). I think it's reasonable for packages which need cblas symbols, to link/dlopen cblas in order to get them instead of just sort of assuming every openblas lib out there has a convenience copy of cblas statically linked into it.
Julia already expects and copes with lapack being external to openblas, as build_h.jl contains both libblas_name and liblapack_name. I think it should do the same with libcblas...
In the meantime I got the package to build and pass the tests by lying and saying:
override LIBBLAS = -lcblas
override LIBBLASNAME = libcblas
Feel free to file a PR to support this with a less hacky solution.
nalimilan
on 5 Oct 2018
Since we build our own BLAS by default, could we link it explicitly as cblas instead of as blas since technically cblas is what we need to link against? That seems like it would solve this problem without breaking our usage of our own OpenBLAS build.
StefanKarpinski
on 5 Oct 2018
Well, I'm currently testing this as a solution, and then building with USE_SYSTEM_CBLAS: https://github.com/eli-schwartz/julia/commit/0c442318196389d653ee21eba65d8c4f7beb72a0
eli-schwartz
on 5 Oct 2018
Related on building from the source; I'm encountering test failures in Sockets and Distributed when building from the source on Nixos. When the failing tests are reordered such that the failing ones are at the top, everything passes just fine, so it appears some of the tests affect the state of the other ones.
f--t
on 21 Oct 2018
I have that issue too, which I fixed by emulating one of the the CI settings (all of which explicitly don't use the apparently unsupported make test target) that disables the sockets test in order to claim success. Presumably those tests don't matter.
eli-schwartz
on 21 Oct 2018
Better file a separate issue. A big issue with lots of problems doesn't help tracking progress.
nalimilan
on 21 Oct 2018
Related issues
wilburtownsend
·
3Comments
iamed2
·
3Comments
m-j-w
·
3Comments
ararslan
·
3Comments
Keno
·
3Comments
Most helpful comment
As noted in the README, we require patched versions of our dependencies, because our dependencies have bugs. We do our best to upstream patches, but that takes time, esp if you're only using major releases.
I'm sorry if you felt misled, but we were never in a position to promise this, and to the best of my knowledge we didn't.
Separately, this issue comes across as very ranty. I appreciate that your intention is to package julia for Arch Linux, but comments like
are extremely off putting and make me less likely to want to help you fix these issues.
With respect to https://github.com/JuliaLinearAlgebra/Arpack.jl/issues/5, as mentioned there we're in the process of addressing these issues, but our first priority is getting it right for the generic binaries, second users building from source locally, and third integration with system package managers. We have rough plans for the third, including allowing distro packagers to re-use binaries that they already ship (assuming they are compatible), but we haven't gotten to that yet. On that point, I would ask for a bit of patience.