Segmentation fault on Julia master

[giordano]

I'm not sure whether this is a bug in StaticArrays.jl or also this issue is actually a Julia bug, but it involves only StaticArrays, so let me first try reporting it here :-)

This code on Julia master causes a segmentation fault:

julia> using StaticArrays

julia> sin.((SMatrix{3,2}(1.0, 2.0, 3.0, 4.0, 5.0, 6.0) * SVector(1.0, 2.0)))
3-element SVector{3,Float64}:
  0.4121184852417566
 -0.5365729180004349
  0.6502878401571168

julia> SVector(1, -1, -4) - [1, -1, -4] # ← segfault happens here

signal (11): Segmentation fault
in expression starting at no file:0
_ZNK4llvm4Type9isEmptyTyEv at /home/mose/repo/julia/usr/bin/../lib/libLLVM-3.9.so (unknown line)
type_is_ghost at /home/mose/repo/julia/src/codegen.cpp:271 [inlined]
emit_bits_compare at /home/mose/repo/julia/src/codegen.cpp:2156
emit_f_is at /home/mose/repo/julia/src/codegen.cpp:2223 [inlined]
emit_builtin_call at /home/mose/repo/julia/src/codegen.cpp:2288
emit_call at /home/mose/repo/julia/src/codegen.cpp:3053
emit_expr at /home/mose/repo/julia/src/codegen.cpp:3767
emit_condition at /home/mose/repo/julia/src/codegen.cpp:3637 [inlined]
emit_function at /home/mose/repo/julia/src/codegen.cpp:5740
jl_compile_linfo at /home/mose/repo/julia/src/codegen.cpp:1155
jl_compile_for_dispatch at /home/mose/repo/julia/src/gf.c:1704
jl_compile_method_internal at /home/mose/repo/julia/src/julia_internal.h:334 [inlined]
jl_call_method_internal at /home/mose/repo/julia/src/julia_internal.h:381 [inlined]
jl_apply_generic at /home/mose/repo/julia/src/gf.c:2003
do_call at /home/mose/repo/julia/src/interpreter.c:70
eval at /home/mose/repo/julia/src/interpreter.c:262
jl_interpret_toplevel_expr_in at /home/mose/repo/julia/src/interpreter.c:50
jl_toplevel_eval_flex at /home/mose/repo/julia/src/toplevel.c:640
jl_toplevel_eval_in at /home/mose/repo/julia/src/builtins.c:626
eval at ./repl/REPL.jl:3
jl_call_fptr_internal at /home/mose/repo/julia/src/julia_internal.h:366 [inlined]
jl_call_method_internal at /home/mose/repo/julia/src/julia_internal.h:385 [inlined]
jl_apply_generic at /home/mose/repo/julia/src/gf.c:2003
eval_user_input at ./repl/REPL.jl:69
jl_call_fptr_internal at /home/mose/repo/julia/src/julia_internal.h:366 [inlined]
jl_call_method_internal at /home/mose/repo/julia/src/julia_internal.h:385 [inlined]
jl_apply_generic at /home/mose/repo/julia/src/gf.c:2003
macro expansion at ./repl/REPL.jl:100 [inlined]
#1 at ./event.jl:96
jl_call_fptr_internal at /home/mose/repo/julia/src/julia_internal.h:366 [inlined]
jl_call_method_internal at /home/mose/repo/julia/src/julia_internal.h:385 [inlined]
jl_apply_generic at /home/mose/repo/julia/src/gf.c:2003
jl_apply at /home/mose/repo/julia/src/julia.h:1451 [inlined]
start_task at /home/mose/repo/julia/src/task.c:268
unknown function (ip: 0xffffffffffffffff)
Allocations: 6406588 (Pool: 6406196; Big: 392); GC: 13

No problem on Julia 0.6, no problem with regular arrays.

[mschauer]

This is worth reporting (Julia shouldn't segfault), but of course can ultimately be caused by bad code in this package.

[c42f]

The segfault in type_is_ghost looks suspiciously familiar, I'm trying to remember where I've seen this before.

I think there's a reasonable chance this is a bug in julia master - it's been changing fairly rapidly and has been a bit unstable for me lately.

[giordano]

Ok, ticket opened at JuliaLang/julia#24193

[julia master has] been changing fairly rapidly and has been a bit unstable for me lately.

For me too ;-)

[c42f]

Looks like this is fixed upstream now.