Bug 588428

builds in mock
rpmlint is clean

%ruby_sitelib should be removed, as it is not used.


You could add a %check

%check
cd %{buildroot}%{geminstdir}
rake spec


There isn't really a reason to do this. You can simply mark the normally installed README file a %doc.  
mkdir -p %{buildroot}/usr/share/doc/%{gemdir}-%{version}
mv  %{buildroot}%{geminstdir}/README %{buildroot}/usr/share/doc/%{gemdir}-%{version}

I'd also encourage you to break this package into a separate -doc package. It would allow someone wishing to use addressable as a dep to not install about 3/4 of the total files.

This initially started as a dependency for rubygem-buildr. It is no longer
needed for that. We can cancel this review.

I will take this over, needed for OpenNebula 3.0.

NOW: Part of Dependency packaging for OpenNebula 2.x/3.x

I have updated this to use Rspec for tests

Spec URL:
http://www.sh0n.net/spstarr/fedora/rubygem-addressable/rubygem-addressable.spec
SRPM URL:
http://www.sh0n.net/spstarr/fedora/rubygem-addressable/rubygem-addressable-2.2.6-1.fc17.src.rpm

Description: Improved URI/URL handling

MOCK: PASS

This package is needed for rubygem-webmock as well. Taking the review

* package depends on rubygem-idn, review for that in progress

* package builds fine in mock w/ rubygem-idn manually installed

* rpmlint looks good

* Package matches upstream source's md5sum

$ md5sum rpmbuild/SOURCES/addressable-2.2.6.gem Downloads/addressable-2.2.6.gem 
3eec7c544b664f28023ff5d7fb0116ef  rpmbuild/SOURCES/addressable-2.2.6.gem
3eec7c544b664f28023ff5d7fb0116ef  Downloads/addressable-2.2.6.gem


* Can you remove references to ruby_sitearch as this is a noarch gem package

* As mentioned before could you replace the "%doc %{geminstdir}/[A-Z]*" with the doc files explicitly listed, just to prevent unintentional mistakes in the future (if non-doc files are added to the upstream project that match this)

* since %geminstdir is being marked as a %dir in the files section, does the Rakefile need to be marked w/ %exclude? (why exclude the Rakefile but then include the lib subdir?)

* The README file is currently listed twice, first as a doc in the main package (pulled in as part of the aforementioned wildcard expression) and then again in the docs subpackage, either place is acceptable but it should only appear once

http://fedoraproject.org/wiki/Packaging/Guidelines#DuplicateFiles

Other than that the package looks good. Thanks for the submissions

Also few nits from my side:

* Please replace "Requires: rubygems" with virtual provider "Requires: ruby(rubygems)". The same apply also for BuildRequires.

* It seems that the pushd on line 54 is useless.

* pushd should be paired with popd in %check section

ping? Am getting regular broken dependencies emails regarding rubygem-webmock.

Again appreciate the submission.

Pong, let me get to this on weekend or so. I still need to sync with Ruby SIG so I can make sure I'm not getting swallowed up in the transisition to Ruby 1.9.3.

It would be really good if there was a Ruby SIG IRC channel.

Any updates?  Shawn, while I'm not a ruby expert, I'm happy to help with the review if you like.  Just let me know, and I'll be happy to get my review feedback in.

I'll fix the comments. But as I've said. Ruby is moving to 1.9.x which is going to complicate things that are not RSpec 2.x ready. I should have this and the latter done by the weekend.

This is updated from both feedback. It's still not RSpec 2.x though.

I'd appreciate any help/tips in making each of these rubygems compliant for Ruby 1.9.x and RSpec 2.x.

Updates here:

http://www.sh0n.net/spstarr/fedora/rubygem-addressable/rubygem-addressable.spec
http://www.sh0n.net/spstarr/fedora/rubygem-addressable/rubygem-addressable-2.2.6-2.fc17.src.rpm

I did not checked the complete gem content, but the RSpec 2.x seems to be called correctly ATM. But if you call RSpec directly, then it renders Rake, RDoc and Yarv dependencies probably useless and they should be dropped.

Package looks good. All feedback addressed. rpmlint / mock builds are green (again once rubygem-idn is installed manually).

APPROVED

Can someone change the flag to + I will submit package as soon as this is done.

(In reply to comment #14)
> Can someone change the flag to + I will submit package as soon as this is done.

Vit, removed rake/rdoc/yard dependencies also.

For some reason the '+' isn't available as an option for the fedora-review flag. Does anyone know the reason for this? It has worked in the past for me.

I've set fedora_review+ as requested

ping? Can we get this into Fedora / close this out now. Thanks alot.

Will push to git and build

I will be pushing this today was blocked due to fedora filer issue and being sick ;/

New Package SCM Request
=======================
Package Name: rubygem-addressable
Short Description: Improved URI/URL handling
Owners: spstarr

Can I get an update from the SCM team on this approval?

ah cvs flag... done

Git done (by process-git-requests).

Looks like this fails in rawhide now:

DEBUG: Failures:
DEBUG:   1) Addressable::URI when form encoding a hash should result in correct percent encoded sequence
DEBUG:      Failure/Error: ).should == "%26one=%2F1&%3Dtwo=%3F2&%3Athree=%233"
DEBUG:        expected: "%26one=%2F1&%3Dtwo=%3F2&%3Athree=%233"
DEBUG:             got: "%3Athree=%233&%26one=%2F1&%3Dtwo=%3F2" (using ==)
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-expectations-2.6.0/lib/rspec/expectations/fail_with.rb:29:in `fail_with'
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-expectations-2.6.0/lib/rspec/matchers/operator_matcher.rb:48:in `fail_with_message'
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-expectations-2.6.0/lib/rspec/matchers/operator_matcher.rb:70:in `__delegate_operator'
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-expectations-2.6.0/lib/rspec/matchers/operator_matcher.rb:60:in `eval_match'
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-expectations-2.6.0/lib/rspec/matchers/operator_matcher.rb:29:in `=='
DEBUG:      # /builddir/build/BUILD/rubygem-addressable-2.2.6/usr/lib/ruby/gems/1.8/gems/addressable-2.2.6/spec/addressable/uri_spec.rb:4213
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-core-2.6.4/lib/rspec/core/example.rb:48:in `instance_eval'
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-core-2.6.4/lib/rspec/core/example.rb:48:in `run'
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-core-2.6.4/lib/rspec/core/example.rb:107:in `with_around_hooks'
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-core-2.6.4/lib/rspec/core/example.rb:45:in `run'
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-core-2.6.4/lib/rspec/core/example_group.rb:294:in `run_examples'
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-core-2.6.4/lib/rspec/core/example_group.rb:290:in `map'
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-core-2.6.4/lib/rspec/core/example_group.rb:290:in `run_examples'
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-core-2.6.4/lib/rspec/core/example_group.rb:262:in `run'
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-core-2.6.4/lib/rspec/core/command_line.rb:24:in `run'
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-core-2.6.4/lib/rspec/core/command_line.rb:24:in `map'
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-core-2.6.4/lib/rspec/core/command_line.rb:24:in `run'
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-core-2.6.4/lib/rspec/core/reporter.rb:12:in `report'
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-core-2.6.4/lib/rspec/core/command_line.rb:21:in `run'
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-core-2.6.4/lib/rspec/core/runner.rb:80:in `run_in_process'
DEBUG:      # /usr/lib/ruby/gems/1.8/gems/rspec-core-2.6.4/lib/rspec/core/runner.rb:69:in `run'
DEBUG: Finished in 1.77 seconds
DEBUG: 1019 examples, 1 failure
DEBUG: Failed examples:
DEBUG: rspec ./spec/addressable/uri_spec.rb:4210 # Addressable::URI when form encoding a hash should result in correct percent encoded sequence

This is now blocking two projects, and I've emailed upstream since this never occurred before.

Hi Shawn,
I have run into these problem and here is the explanation:

There was a security problem with hash table collisions in Ruby < 1.9 (see [1] for more info). This was recently solved by introducing Ruby-1.8.7-p357, which deals with this issue by assigning a random seed to each hash (again, look at [1] for the details). Therefore the order of a hash can be different in each interpreter run. The solution is not to test for the exact string, but its contents. BTW, the order of hashes was never guaranteed in Ruby < 1.9, so it's always been a bug, but everything worked well in the simple cases.

In Ruby >= 1.9, the hashes are ordered by default, so the tests won't be failing once we switch to 1.9.3.

[1] https://bugzilla.redhat.com/show_bug.cgi?id=750564

I have an updated package for review:

* Koji scratch build pass 

SPEC: http://fedorapeople.org/~spstarr/packages/rubygem-addressable.spec

SRPM: http://fedorapeople.org/~spstarr/packages/rubygem-addressable-2.2.6-3.fc17.src.rpm

ping, can someone review? Some fixes have been made to rubygem-idn which now make no modifications to rubygem-addressable required for package test to pass.

Shawn, I took just quick look on the spec file and it looks good. Since the package was already approved, you have requested git repo and it was already prepared, nothing can't prevent you from importing the package into Fedora ;)

This is pushed into f17 and rawhide now.

rubygem-addressable-2.2.6-3.fc17 has been submitted as an update for Fedora 17.
https://admin.fedoraproject.org/updates/rubygem-addressable-2.2.6-3.fc17

Could this be pushed into Fedora 16 as well. rubygem-webmock will be broken there until it is.

I've attached the following SRPM / SPEC and diff to the addressable spec adding ruby 1.8 / Fedora 16 support. I've also requested commit rights to the rawhide so that I can request the new branch / do the push if you don't have the cycles.

Thanks.

http://mo.morsi.org/files/rpms/rubygem-addressable-2.2.6-4.fc18.src.rpm
http://mo.morsi.org/files/rpms/rubygem-addressable.spec
http://mo.morsi.org/files/rpms/addressable-2.2.6.spec.diff

Please do request commit, never can have enough maintainers :) even for older versions

Package Change Request
======================
Package Name: rubygem-addressable
New Branches: f16
Owners: mmorsi
InitialCC: 

F16 branch for rubygem-addressable

Git done (by process-git-requests).

rubygem-addressable-2.2.6-3.fc17 has been pushed to the Fedora 17 stable repository.  If problems still persist, please make note of it in this bug report.

Package Change Request
======================
Package Name: rubygem-addressable
New Branches: epel7
Owners: tdawson

Git done (by process-git-requests).