Failiure in misc on master #15341
Comments
|
Is this reproducible? Can this be reproduced with single process test or single test? |
|
Reproducible on |
|
Hmm, for one thing it seems that the error reporting for Can you try to run @show ccall((:OpenClipboard, "user32"), stdcall, Cint, (Ptr{Void},), C_NULL)
e = Libc.GetLastError()
@show e
@show FormatMessage(e)? |
julia> @show ccall((:OpenClipboard, "user32"), stdcall, Cint, (Ptr{Void},), C_NULL)
ccall((:OpenClipboard,"user32"),Cint,(Ptr{Void},),C_NULL,$(Expr(:stdcall))) = 1
1
julia> e = Libc.GetLastError()
0x000001e7
julia> @show e
e = 0x000001e7
0x000001e7
julia> @show FormatMessage(e)
ERROR: UndefVarError: FormatMessage not defined
in eval(::Module, ::Any) at .\boot.jl:267 |
|
Ooops replace the |
Really?.... |
julia> @show ccall((:OpenClipboard, "user32"), stdcall, Cint, (Ptr{Void},), C_NULL)
ccall((:OpenClipboard,"user32"),Cint,(Ptr{Void},),C_NULL,$(Expr(:stdcall))) = 1
1
julia> e = Libc.GetLastError()
0x0000007f
julia> @show e
e = 0x0000007f
0x0000007f
julia> @show Libc.FormatMessage(e)
Libc.FormatMessage(e) = "The specified procedure could not be found. "
"The specified procedure could not be found. Excuse the previous messup! |
|
Well, you still need to call the function. Maybe f() = ccall((:OpenClipboard, "user32"), stdcall, Cint, (Ptr{Void},), C_NULL)
@code_llvm f()
@code_native f()? I don't really see why this issue is threading specific though..... |
julia> f() = ccall((:OpenClipboard, "user32"), stdcall, Cint, (Ptr{Void},), C_NULL)
f (generic function with 1 method)
julia> @code_llvm f()
; Function Attrs: uwtable
define i32 @julia_f_1536() #0 {
top:
%0 = call x86_stdcallcc i32 inttoptr (i64 140734516732800 to i32 (i8*)*)(i8* null)
ret i32 %0
}
julia> @code_native f()
.text
Filename: none
Source line: 0
pushq %rbp
movq %rsp, %rbp
Source line: 1
subq $32, %rsp
movabsq $OpenClipboard, %rax
xorl %ecx, %ecx
callq *%rax
addq $32, %rsp
popq %rbp
retqAbsolutely no idea either. That is the only change I have in my |
|
Is there a difference if you run output the llvm and native code for the non-threading version? |
mikewl
changed the title
|
Hmmm, not limited to threading. Disabled threading and tested just julia> Base.runtests("misc")
* misc ERROR: LoadError: LoadError: SystemError: OpenClipboard: No error
[inlined code] from .\promotion.jl:185
in clipboard(::ASCIIString) at .\interactiveutil.jl:132
[inlined code] from K:\julia\usr\share\julia\test\misc.jl:384
in anonymous at .\no file:4294967295
in include(::UTF8String) at .\boot.jl:264
in include_from_node1(::UTF8String) at .\loading.jl:417
[inlined code] from .\util.jl:179
in runtests(::UTF8String) at K:\julia\usr\share\julia\test\testdefs.jl:7
in run_work_thunk(::Base.##247#248{##16#24,Tuple{UTF8String}}, ::Bool) at .\multi.jl:714
in remotecall_fetch(::Any, ::LocalProcess, ::UTF8String, ::Vararg{UTF8String}) at .\multi.jl:787
in remotecall_fetch(::Any, ::Int64, ::UTF8String, ::Vararg{UTF8String}) at .\multi.jl:803
[inlined code] from K:\julia\usr\share\julia\test\runtests.jl:36
in (::##15#23{Array{Any,1}})() at .\task.jl:431
while loading K:\julia\usr\share\julia\test\misc.jl, in expression starting on line 382
in remotecall_fetch(::Any, ::LocalProcess, ::UTF8String, ::Vararg{UTF8String}) at .\multi.jl:788
in remotecall_fetch(::Any, ::Int64, ::UTF8String, ::Vararg{UTF8String}) at .\multi.jl:803
[inlined code] from K:\julia\usr\share\julia\test\runtests.jl:36
in (::##15#23{Array{Any,1}})() at .\task.jl:431
in sync_end() at .\task.jl:397
[inlined code] from .\task.jl:406
in (::##11#19)() at K:\julia\usr\share\julia\test\runtests.jl:31
in cd(::##11#19, ::ASCIIString) at .\file.jl:58
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 K:\julia\usr\share\julia\test\runtests.jl, in expression starting on line 13
ERROR: A test has failed. Please submit a bug report (https://github.com/JuliaLang/julia/issues)
including error messages above and the output of versioninfo():
Julia Version 0.5.0-dev+2985
Commit d8ec4a7* (2016-03-03 01:09 UTC)
Platform Info:
System: Windows (x86_64-w64-mingw32)
CPU: Intel(R) Core(TM) i7-3770K CPU @ 3.50GHz
WORD_SIZE: 64
BLAS: libopenblas (USE64BITINT DYNAMIC_ARCH NO_AFFINITY Sandybridge)
LAPACK: libopenblas64_
LIBM: libopenlibm
LLVM: libLLVM-3.7.1
in error(::ASCIIString) at .\error.jl:21
in runtests(::ASCIIString, ::Int64) at .\interactiveutil.jl:420
in runtests(::ASCIIString) at .\interactiveutil.jl:409
in eval(::Module, ::Any) at .\boot.jl:267julia> f() = ccall((:OpenClipboard, "user32"), stdcall, Cint, (Ptr{Void},), C_NULL)
f (generic function with 1 method)
julia> @code_llvm f()
; Function Attrs: uwtable
define i32 @julia_f_1404() #0 {
top:
%0 = call x86_stdcallcc i32 inttoptr (i64 140734516732800 to i32 (i8*)*)(i8* null)
ret i32 %0
}
julia> @code_native f()
.text
Filename: none
Source line: 0
pushq %rbp
movq %rsp, %rbp
Source line: 1
subq $32, %rsp
movabsq $OpenClipboard, %rax
xorl %ecx, %ecx
callq *%rax
addq $32, %rsp
popq %rbp
retq
julia>Going to run all tests again. See if the misc tests only pass under no threading and run all. |
|
This time it failed with threading disabled and all tests ran. So it is intermittently passing. |
|
Could #15033 have possibly affected this? OT: cygwin keeps on failing with: |
|
Maybe, can you bisect? What windows version are you on? I haven't seen this but I've been using windows 10 lately. |
|
I am on windows 10. Not sure what is causing it. Will start a bisect in the morning. |
|
The cygwin issue might be if you're running in a clone that wasn't downloaded by cygwin's git. Some of the new ACL permissions stuff in cygwin works differently for cygwin's git vs mingw git. |
|
It was that it wasn't Cygwin's git thanks! I did the bisect and I am going to do it again just to be sure but bisect gives: I will redo bisect just in case. |
|
If it's an intermittent failure, you may need to run the misc test 5 |
|
I did it properly and got an amusing result: Will continue tomorrow with including the misc tests manually. |
|
d'oh, should've seen that coming! whoops |
|
Maybe those clipboard tests are just too problematic and should be disabled entirely? |
|
Maybe. Worth figuring out why they're so problematic though. |
|
My guess is that it's an interactive feature that's likely to be unsupported on pure server systems. |
|
Does apple even have a "pure server" variant of os x any more? Their man page for pbcopy didn't indicate any obvious restrictions. And this issue is on Windows, still not clear why me and appveyor see different behavior than @Mike43110 |
|
don't be fooled by the pretty gui, underneath it is still unix. it's not hard to make it fail, all you need is an account without a running quartz server: |
|
Ah. Sure would've been useful for the man page to mention that. |
|
So I guess we should check for a running quartz server before running these tests? |
|
marking as a windows error handling bug since it is printing the wrong error code. |
|
Possibly related to #15117 where the wrong error was being reported? |
|
stale |
Make.usercontainsJULIA_THREADS=1.Tests pass fine with threading disabled.
The text was updated successfully, but these errors were encountered: