"#test_cartesian not defined" error in subarrays test

[nalimilan]

This failure happens when building the RPM nightlies (LLVM 3.7.1 with Julia patches), both on 32- and 64-bit. It appeared for the first time on on February 5th. It doesn't seem to be present in nightlies from the 4th.

    From worker 2:       * subarray             A = [1 14 27 40 53 66 79 92 105 118 131 144 157]
    From worker 2:  
    From worker 2:  [170 183 196 209 222 235 248 261 274 287 300 313 326]
    From worker 2:  
    From worker 2:  [339 352 365 378 391 404 417 430 443 456 469 482 495]
    From worker 2:  
    From worker 2:  [508 521 534 547 560 573 586 599 612 625 638 651 664]
    From worker 2:  
    From worker 2:  [677 690 703 716 729 742 755 768 781 794 807 820 833]
    From worker 2:  
    From worker 2:  [846 859 872 885 898 911 924 937 950 963 976 989 1002]
    From worker 2:  
    From worker 2:  [1015 1028 1041 1054 1067 1080 1093 1106 1119 1132 1145 1158 1171]
    From worker 2:  
    From worker 2:  [1184 1197 1210 1223 1236 1249 1262 1275 1288 1301 1314 1327 1340]
    From worker 2:  
    From worker 2:  [1353 1366 1379 1392 1405 1418 1431 1444 1457 1470 1483 1496 1509]
    From worker 2:  
    From worker 2:  [1522 1535 1548 1561 1574 1587 1600 1613 1626 1639 1652 1665 1678]
    From worker 2:  
    From worker 2:  [1691 1704 1717 1730 1743 1756 1769 1782 1795 1808 1821 1834 1847]
    From worker 2:  
    From worker 2:  [1860 1873 1886 1899 1912 1925 1938 1951 1964 1977 1990 2003 2016]
    From worker 2:  
    From worker 2:  [2029 2042 2055 2068 2081 2094 2107 2120 2133 2146 2159 2172 2185]
    From worker 2:  B = [1 14 27 40 53 66 79 92 105 118 131 144 157]
    From worker 2:  
    From worker 2:  [170 183 196 209 222 235 248 261 274 287 300 313 326]
    From worker 2:  
    From worker 2:  [339 352 365 378 391 404 417 430 443 456 469 482 495]
    From worker 2:  
    From worker 2:  [508 521 534 547 560 573 586 599 612 625 638 651 664]
    From worker 2:  
    From worker 2:  [677 690 703 716 729 742 755 768 781 794 807 820 833]
    From worker 2:  
    From worker 2:  [846 859 872 885 898 911 924 937 950 963 976 989 1002]
    From worker 2:  
    From worker 2:  [1015 1028 1041 1054 1067 1080 1093 1106 1119 1132 1145 1158 1171]
    From worker 2:  
    From worker 2:  [1184 1197 1210 1223 1236 1249 1262 1275 1288 1301 1314 1327 1340]
    From worker 2:  
    From worker 2:  [1353 1366 1379 1392 1405 1418 1431 1444 1457 1470 1483 1496 1509]
    From worker 2:  
    From worker 2:  [1522 1535 1548 1561 1574 1587 1600 1613 1626 1639 1652 1665 1678]
    From worker 2:  
    From worker 2:  [1691 1704 1717 1730 1743 1756 1769 1782 1795 1808 1821 1834 1847]
    From worker 2:  
    From worker 2:  [1860 1873 1886 1899 1912 1925 1938 1951 1964 1977 1990 2003 2016]
    From worker 2:  
    From worker 2:  [2029 2042 2055 2068 2081 2094 2107 2120 2133 2146 2159 2172 2185]
Worker 2 terminated.
ERROR (unhandled task failure): UndefVarError: #test_cartesian not defined
Exception running test subarray :
ProcessExitedException()
ERROR: LoadError: Some tests exited with errors.
 in (::##11#19)() at /builddir/build/BUILD/julia/test/runtests.jl:64
 in cd(::##11#19, ::ASCIIString) at ./file.jl:47
 in include(::ASCIIString) at ./boot.jl:264
 in include_from_node1(::UTF8String) at ./loading.jl:417
 in process_options(::Base.JLOptions) at ./client.jl:244
 in _start() at ./client.jl:300
while loading /builddir/build/BUILD/julia/test/runtests.jl, in expression starting on line 13

https://copr-be.cloud.fedoraproject.org/results/nalimilan/julia-nightlies/fedora-23-i386/00158644-julia/build.log.gz

[vtjnash]

the serializer is hiding the real error. in particular, it is trying to serialize the full method signature for the improved stacktrace, but the target function doesn't actually exist on the master processor. i think a custom serializer for StackTrace should resolve this.

[tkelman]

Thanks to #15097 this gives the right backtrace now (https://ci.appveyor.com/project/StefanKarpinski/julia/build/1.0.13454/job/p4qa39w5oy3vrbpe):

Exception running test subarray : 
On worker 3:
LoadError: Mismatch 
 in test_cartesian at C:\projects\julia\test\subarray.jl:152 
 in runtests at C:\projects\julia\test\subarray.jl:296 
 in runviews at C:\projects\julia\test\subarray.jl:319 
 [inlined code] from C:\projects\julia\test\subarray.jl:396 
 in anonymous at .\no file:-1 
 [inlined code] from .\essentials.jl:78 
 in include_string at .\loading.jl:371 
 in include_from_node1 at .\loading.jl:420 
 [inlined code] from .\util.jl:179 
 in runtests at C:\projects\julia\test\testdefs.jl:7 
 in #16 at C:\projects\julia\test\runtests.jl:36 
 in run_work_thunk at .\multi.jl:714 
 [inlined code] from .\multi.jl:1010 
 in #255 at .\task.jl:59 
while loading C:\projects\julia\test\subarray.jl, in expression starting on line 378 
ERROR: LoadError: Some tests exited with errors.
 in (::##11#19)() at C:\projects\julia\test\runtests.jl:64
 in cd(::##11#19, ::ASCIIString) at .\file.jl:57
 in include(::ASCIIString) at .\boot.jl:264
 in include_from_node1(::UTF8String) at .\loading.jl:417
 in process_options(::Base.JLOptions) at .\client.jl:262
 in _start() at .\client.jl:318
while loading C:\projects\julia\test\runtests.jl, in expression starting on line 13

[nalimilan]

#15151 has a similar backtrace.

[tkelman]

@nalimilan were you ever able to reproduce this locally, or only on remote build servers?

[nalimilan]

I have never trying reproducing locally. But I've not seen it in a while on the build servers, so it may well be the same stack size issue as #15079 and #14543 (comment) (I've increased the stack limit).

So maybe close this. EDIT: hadn't seen #15271

[tkelman]

Let's close this given #15271 is a bit more specific and up to date.