Possibly more i686 issues

[staticfloat]

Launchpad builds are now throwing some InexactErrors() on the 32-bit builds (which I just recently switched from targeting pentium4 to instead targeting i686):

exception on 5: ERROR: test failed: ishermitian(A' * A)
 in expression: ishermitian(A' * A)
 in error at error.jl:21
 in default_handler at test.jl:25
 in do_test at test.jl:50
 in runtests at /build/buildd/julia-0.4.0/test/testdefs.jl:5
 in anonymous at multi.jl:828
 in run_work_thunk at multi.jl:593
 in anonymous at task.jl:828
while loading arrayops.jl, in expression starting on line 897
exception on 4: ERROR: test error in expression: Dates.DateTime("[14:51:00.118]","[HH:MM:SS.sss]") == t
InexactError()
 in slotparse at dates/io.jl:99
 in getslot at dates/io.jl:108
 in parse at dates/io.jl:120
 in anonymous at test.jl:83
 in do_test at test.jl:47
 in include_string at loading.jl:97
 in include_from_node1 at ./loading.jl:131
 in runtests at /build/buildd/julia-0.4.0/test/testdefs.jl:5
 in anonymous at multi.jl:828
 in run_work_thunk at multi.jl:593
 in anonymous at task.jl:828
while loading /build/buildd/julia-0.4.0/test/dates/io.jl, in expression starting on line 272
while loading dates.jl, in expression starting on line 14
ERROR: test error in expression: Dates.DateTime("[14:51:00.118]","[HH:MM:SS.sss]") == t
InexactError()
 in include_from_node1 at ./loading.jl:131
 in start_reading at stream.jl:617
 in worker_id_from_socket at multi.jl:343
 in Worker at multi.jl:95
 in anonymous at task.jl:898
while loading /build/buildd/julia-0.4.0/test/dates/io.jl, in expression starting on line 272
while loading dates.jl, in expression starting on line 14
while loading /build/buildd/julia-0.4.0/test/runtests.jl, in expression starting on line 40

[staticfloat]

@quinnj pinging you on the dates stuff

[tkelman]

Which test is the hermitian failure in? I can reproduce the dates failure with the linux-i386 nightly (why does it say master is a fork?)

  | | |_| | | | (_| |  |  Version 0.4.0-dev+1132 (2014-10-18 20:43 UTC)
 _/ |\__'_|_|_|\__'_|  |  master/636f481 (fork: 40 commits, 4 days)
|__/                   |  i386-redhat-linux

but it passes with a win32 build (without MARCH set).

[staticfloat]

It's from arrayops.jl, right?
On Oct 19, 2014 12:04 AM, "Tony Kelman" notifications@github.com wrote:

Which test is the hermitian failure in? I can reproduce the dates failure
with the linux-i386 nightly (why does it say master is a fork?)

| | || | | | (| | | Version 0.4.0-dev+1132 (2014-10-18 20:43 UTC)
/ |__'|||'_| | master/636f481 (fork: 40 commits, 4 days)
|/ | i386-redhat-linux

but it passes with a win32 build (without MARCH set).

—
Reply to this email directly or view it on GitHub
#8731 (comment).

[ivarne]

@tkelman master (or any branch) is considered a fork if it contains commits that is not present in the origin/master branch.

I tried not to hardcode origin as remote, but rather searched for the remote with JuliaLang/julia.git in its url. When that fails, my intention was to just assume that master was a valid branch, but it seems like @staticfloat compiles without creating a local master branc on his buildbots, so he removed what I thought was a reasonable fallback.

What is the output of git config -l 2>/dev/null | grep 'remote\.\w*\.url'?

[staticfloat]

I didn't think I removed anything; all I did was ensure that $origin would be set to something at all, right? Because if it's not, then we get git erroring out on every operation that involves $origin.

[staticfloat]

Closing this in favor of #8812