Release 0.6.0

[dfornika]

Addresses #6.

[conda-forge-linter]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

[jakirkham]

Would suggest raising this test failure upstream and seeing what they say.

[dfornika]

I will. This sort of issue belongs on the Julia 'discourse' list, not an issue on the julia Github repo, correct? Or is there another place to discuss this?

[jakirkham]

It seems like a bug. So personally would add it to the issue tracker, but will happily defer to your judgement on where to raise it.

[dfornika]

Ok, I've posted it to the julia discourse for now. Will create a github issue if it doesn't get any attention there.

https://discourse.julialang.org/t/test-spawn-failure-in-conda-build-julia-0-6-0/

[tkelman]

Your llvm build is the wrong version or insufficiently patched. Anything that can't be reproduced with the official binaries or recommended build configuration would be closed as a packaging bug.

[jakirkham]

We aren't using the conda-forge packaged LLVM though. We are just using the one that Julia builds itself. 😕

[tkelman]

Does the failure happen if you drop the wrapper script business?

[jakirkham]

I've raised issue ( #10 ) in regards to removing the wrapper script. However I'm not entirely sure it is possible to remove the wrapper script as noted in that issue.

[dfornika]

I've set up a pull request against the v0.5 branch to remove the wrapper script (just got it started, not sure how to configure the .juliarc right now).

[jakirkham]

Toggling for CIs.

[dfornika]

Worker 1 failed running test spawn:
Some tests did not pass: 106 passed, 1 failed, 0 errored, 4 broken.spawn: Test Failed
  Expression: readchomp(pipeline(ignorestatus(@cmd("\$exename --startup-file=no -e '\n            struct Error19864 <: Exception; end\n            Base.showerror(io::IO, e::Error19864) = print(io, \"correct19864\")\n            throw(Error19864())'")), stderr=catcmd)) == "ERROR: correct19864"
   Evaluated: "warning: parsing line table prologue at 0x0001bb18 should have ended at 0x0001beac but it ended at 0x0001bbf7\nwarning: parsing line table prologue at 0x0000be03 should have ended at 0x0000c16a but it ended at 0x0000bed5\nwarning: parsing line table prologue at 0x000074a3 should have ended at 0x0000786e but it ended at 0x000075c6\nwarning: parsing line table prologue at 0x00002673 should have ended at 0x00002a48 but it ended at 0x00002745\nwarning: parsing line table prologue at 0x000095e1 should have ended at 0x00009909 but it ended at 0x000096b3\nwarning: parsing line table prologue at 0x000112ea should have ended at 0x0001166e but it ended at 0x000113bc\nwarning: parsing line table prologue at 0x00000000 should have ended at 0x00000434 but it ended at 0x00000167\nERROR: correct19864" == "ERROR: correct19864"warning: parsing line table prologue at 0x00000000 should have ended at 0x00000434 but it ended at 0x00000167

Stacktrace:
 [1] record(::Base.Test.DefaultTestSet, ::Base.Test.Fail) at ./test.jl:568
 [2] (::##40#46)() at /feedstock_root/build_artefacts/julia_1505088090973/_t_env/share/julia/test/runtests.jl:160
 [3] cd(::##40#46, ::String) at ./file.jl:70
 [4] include_from_node1(::String) at ./loading.jl:569
 [5] include(::String) at ./sysimg.jl:14
 [6] process_options(::Base.JLOptions) at ./client.jl:305
 [7] _start() at ./client.jl:371

We're getting a lot of these messages like:

warning: parsing line table prologue at 0x0001bb18 should have ended at 0x0001beac but it ended at 0x0001bbf7

...in the build log. Am I reading the error message correctly that this is the cause of that test failure? Those messages are getting mixed up in the test input?

[jakirkham]

I think the warning is coming from LLDB, the LLVM debugger. Not sure why it is coming up though.

[jakirkham]

IDK about the warnings causing the issue though. We have those warnings already in the latest 0.5 build. So if they were to cause issues, would have expected that to happen already. I think the failure here is unrelated.

[dfornika]

There's a new section of the spawn test that's not present in v0.5.2.

# issue #19864 (PR #20497)
@test readchomp(pipeline(ignorestatus(
        `$exename --startup-file=no -e '
            struct Error19864 <: Exception; end
            Base.showerror(io::IO, e::Error19864) = print(io, "correct19864")
            throw(Error19864())'`),
    stderr=catcmd)) == "ERROR: correct19864"

JuliaLang/julia#19864

[jakirkham]

So we don't build in an interactive terminal. Not sure if that matters for this test or not.

[tkelman]

If you're getting a bunch of warnings from llvm on startup, something is broken

[jakirkham]

Not seeing these warning when starting up Julia. Only seeing them peppered throughout the tests.

[dfornika]

I do see those messages in an interactive julia session with the current v0.5.2 build. They don't come up after a valid expression but do come up during an error.

[dfornika@sabin ~]$ source activate julia
(julia) [dfornika@sabin ~]$ julia
               _
   _       _ _(_)_     |  A fresh approach to technical computing
  (_)     | (_) (_)    |  Documentation: https://docs.julialang.org
   _ _   _| |_  __ _   |  Type "?help" for help.
  | | | | | | |/ _` |  |
  | | |_| | | | (_| |  |  Version 0.5.2
 _/ |\__'_|_|_|\__'_|  |  conda-forge-julia release
|__/                   |  x86_64-unknown-linux-gnu

julia> 2 + 2
4

julia> akljdsfl
warning: parsing line table prologue at 0x0000e1c8 should have ended at 0x0000e52f but it ended at 0x0000e284
warning: parsing line table prologue at 0x00003c27 should have ended at 0x00003ffc but it ended at 0x00003ce3
warning: parsing line table prologue at 0x0001cd5c should have ended at 0x0001d0f0 but it ended at 0x0001ce25
warning: parsing line table prologue at 0x000086e2 should have ended at 0x00008ae4 but it ended at 0x000087ef
warning: parsing line table prologue at 0x0000a9c0 should have ended at 0x0000ace8 but it ended at 0x0000aa7c
ERROR: UndefVarError: akljdsfl not defined

julia> 

[dfornika]

Not sure how helpful this is, but this is the llvm function where the error originates:

http://llvm.org/doxygen/structllvm_1_1DWARFDebugLine_1_1Prologue.html#a8da40bebd23f4bbc9a1cd4e85df665ab

[dfornika]

Here's the test run by itself in an interactive session:

julia> exename = Base.julia_cmd()
`/home/dfornika/miniconda2/envs/julia/bin/julia -Ccore2 -J/home/dfornika/miniconda2/envs/julia/lib/julia/sys.so --compile=yes --depwarn=yes`

catcmd = `cat`
`cat`

julia> readchomp(pipeline(ignorestatus(`$exename --startup-file=no -e 'struct Error19864 <: Exception; end Base.showerror(io::IO, e::Error19864) = print(io, "correct19864") throw(Error19864())'`), stderr=catcmd))
"warning: parsing line table prologue at 0x0001cd5c should have ended at 0x0001d0f0 but it ended at 0x0001ce25\nwarning: parsing line table prologue at 0x0000e1c8 should have ended at 0x0000e52f but it ended at 0x0000e284\nwarning: parsing line table prologue at 0x000086e2 should have ended at 0x00008ae4 but it ended at 0x000087ef\nwarning: parsing line table prologue at 0x0000a9c0 should have ended at 0x0000ace8 but it ended at 0x0000aa7c\nwarning: parsing line table prologue at 0x0001357a should have ended at 0x0001390c but it ended at 0x00013636\nwarning: parsing line table prologue at 0x00003c27 should have ended at 0x00003ffc but it ended at 0x00003ce3\nwarning: parsing line table prologue at 0x00000000 should have ended at 0x0000041a but it ended at 0x00000151\nERROR: syntax: extra token \"Error19864\" after end of expression\n in eval(::Module, ::Any) at ./boot.jl:234\n in process_options(::Base.JLOptions) at ./client.jl:242\n in _start() at ./client.jl:321"

julia> 

[jakirkham]

An interesting test would be to see if we get the same behavior from Julia during the build. It's possible that prefix replacement by conda-build ends up causing us some issues.

[jakirkham]

Toggling for CIs.

[jakirkham]

Appears this is now passing. 🎉 If there is nothing else, I think we should merge this.

[dfornika]

Great work!