Ruby 2.2

Vít Ondruch vondruch at redhat.com
Tue Oct 14 08:35:27 UTC 2014


Hi František,

Thanks for testing!


Dne 5.10.2014 v 21:17 František Dvořák napsal(a):
> Hi,
>
> On Mon, 2014-09-15 at 17:38 +0200, Vít Ondruch wrote:
>> Hi everybody,
>>
>> Since Ruby 2.2 is going to be released during Christmas and -preview1
>> release is imminent (this Wednesday?), it is probably time to start
>> looking into its packaging. So here is the updated .spec file [1] and
>> scratch build [2], which can be finally build on all platforms. Sorry,
>> no Copr for you, since Ruby's build fails there due to old RHEL kernel :/.
>>
> It look like koji is OK.

Yes, koji is OK, it uses F20 kernel to my knowledge.

> I've created repositories for experimenting
> (the preview1 version):
>
> http://scientific.zcu.cz/repos/ruby22-epel7.repo
> http://scientific.zcu.cz/repos/ruby22-fedora20.repo
> http://scientific.zcu.cz/repos/ruby22-fedora21.repo

Nice! Glad to see that the .spec file works fine with the preview1.

>> * Some prevailing test failures were resolved, some others introduced,
>> but hopefully they'll get resolved prior stable release.
>>
> There is one unstable test
> (https://koji.fedoraproject.org/koji/getfile?taskID=7770442&name=build.log&offset=-4000):
>
>   5) Failure:
> TestException#test_machine_stackoverflow_by_define_method
> [/builddir/build/BUILD/ruby-2.2.0-preview1/test/ruby/test_exception.rb:555]:
> pid 27865 killed by SIGSEGV (signal 11)
> 15848 tests, 2507477 assertions, 1 failures, 0 errors, 33 skips
>
> It passed OK on second build.

Yes, I met the issue as well:

https://bugs.ruby-lang.org/issues/9739#note-6

But it passed without any attention. May be I should open separate issue.

>
> And the build machine needs to have IPv4 enabled, but that's probably
> nothing which can be easily changed (launching local network tests are
> usefull too).

What do you mean by this? The build does not pass on IPv6 only machine
or how should I understand it?

>
>> Please test the packaging if you can and let me know about any issues
>> you encountered.
>>
> rpmlint returns some messages. Bad thing may be the "evil" permissions
> 0666 on /usr/share/gems/specifications/*.gemspec.

That is good idea to check rpmlint output. There seems to be more issues
than this one.

Thanks for the feedback.


Vít



More information about the ruby-sig mailing list