This sometimes crashes mspec worker like
https://ci.appveyor.com/project/ruby/ruby/build/9806/job/37tx2atojy96227m
and it doesn't generate helpful output and it seems hard to fix it soon.
As AppVeyor is too unstable by too many factors, let me skip this
at least for a short term. until AppVeyor gets stable.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64933 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Remove .name since SEGV on MinGW is fixed.
[ruby-core:89200]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64880 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* Otherwise it seems to SEGV on Windows, see
f00bf24272 (commitcomment-30650955)
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64855 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
On Appveyor, WIN32OLE sometimes fails due to a system shutdown, and
`@ie` is not assigned.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64815 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Otherwise, I get NameError when running these IPv6 tests individually
or in parallel.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64663 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* Most of these seem OS bugs.
* See 20180905T103302Z.fail.html.gz
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64644 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* This reverts commit 12f624b673:
"Try 4 times for WIN32OLE specs"
* It was a machine problem, it needed to be rebooted.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64498 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Under pipe page memory pressure on Linux, a pipe may only be
created with a single buffer[1]. And as of Linux v4.18, the
fs/pipe.c::pipe_poll callback does not account for merging
done in fs/pipe::pipe_write; only the number of usable buffers
in the pipe. Thus it is possible for a pipe to be writable
(if only by a small amount) despite IO.select saying it is not.
With the default 16384 /proc/sys/fs/pipe-user-pages-soft value
and the pipe having 16 pages of buffers, this issue is trivially
reproducible by creating 1024 pipes in a background process
before running the spec:
$ ulimit -n 2053 # or something higher
$ ./miniruby -e '1024.times.map { IO.pipe }; sleep' &
$ make test-spec MSPECOPT=spec/ruby/core/io/select_spec.rb
So, we create a new pipe we never write to for testing
writability of IO.select. This may cause the test to fail with
ENFILE on an overloaded system, but at least that's an obvious
failure (unlike having a test get stuck). This should reduce
failures on our overloaded CI machines:
http://ci.rvm.jp/results/trunk-nopara@silicon-docker/1239426
[1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/fs/pipe.c?h=v4.18#n642
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64478 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* The specs above already do that, and Solaris needs it too:
20180819T111806Z.fail.html.gz
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64470 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* Where localhost is an alias but not the primary name of 127.0.0.1.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64463 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* Only Solaris 2.10 i386 and Windows seem to return 0 it and other
Solaris seem to fill the value.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64462 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* Happens when VALIDATE_SOCKLEN() actually checks the length such as on
FreeBSD.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64455 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* I believe this should be default behavior, see [Feature #2250].
* Now make test-spec MSPECOPT='-R100 spec/ruby/library/socket' works fine.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64454 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
