Dne 04. 11. 22 v 18:34 Vít Ondruch napsal(a):
Dne 04. 11. 22 v 13:24 Vít Ondruch napsal(a):
>
> Dne 03. 11. 22 v 16:37 Vít Ondruch napsal(a):
>>
>> Dne 03. 11. 22 v 15:37 Mamoru TASAKA napsal(a):
>>> Vít Ondruch wrote on 2022/11/03 22:22:
>>>>
>>>> Dne 03. 11. 22 v 14:07 Mamoru TASAKA napsal(a):
>>>>> Vít Ondruch wrote on 2022/11/03 21:19:
>>>>>> I have provided negative karma for F37 for the moment:
>>>>>>
>>>>>>
https://bodhi.fedoraproject.org/updates/FEDORA-2022-f7975d0e6a
>>>>>>
>>>>>>
>>>>>> However, not sure if it is not too late already, since the
>>>>>> update was submitted for stable. Mamoru, could you please check
>>>>>> the status and our options here?
>>>>>>
>>>>>>
>>>>>> Vít
>>>>>
>>>>> This is:
>>>>>
https://github.com/rspec/rspec-mocks/commit/e931e818b577172b89fb4583fc336...
>>>>>
>>>>> i.e. to ”emphasize" the difference between keyword v.s. hash
>>>>
>>>>
>>>> Thx for pointing out the exact commit.
>>>>
>>>>
>>>>>
>>>>> I "think" the package seeing errors due to the above change
need
>>>>> fixing anyway,
>>>>
>>>>
>>>> No doubt about it, the only question is when :)
>>>>
>>>>
>>>>> however
>>>>> To distinguish "keywords" v.s. "hash" with ruby
3.x seems
>>>>> generally weigh too difficult...
>>>>>
>>>>> As Fedora 37 is not released yet, and is going to be maintained
>>>>> for 13 months,
>>>>> I think fixing F37 packages seeing the above error is desirable
>>>>
>>>>
>>>> In this case, can we postpone landing the F37 update and include
>>>> fixes for the affected packages into the update?
>>>>
>>>> (I wish the notifications were not delayed by one week, but
>>>> hopefully, the FMN is going to be fixed soon).
>>>>
>>>>
>>>>> (and on the other hand,
>>>>> I am not going to upgrade F36 rspec series to 3.12.x)
>>>>
>>>>
>>>> I support that, thx.
>>>>
>>>> Vít
>>>
>>>
>>> Well, the simplest solution for now is to revert the above change
>>> on F-37.
>>
>>
>> You mean the rspec-mocks commit? That is interesting idea.
>>
>> In the man time, I have fixed rubygem-notiffany, but there seems to
>> be several more:
>>
>>
https://koschei.fedoraproject.org/search?q=ruby&order_by=state-f38%2C...
>>
>>
>> rubygem-listen
>>
>> rubygem-guard
>
>
> This one ^^ won't be easy. The test failure comes from here:
>
>
https://github.com/guard/guard/blob/4f5f763d05c113dc607ea39a909ec401441b3...
>
>
> However, it is not clear to me what defines COLUMNS env variable.
This is part of the backtrace from where the `abort` is called:
~~~
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/message_expectation.rb:761:in
`block in call'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/message_expectation.rb:760:in
`map'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/message_expectation.rb:760:in
`call'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/message_expectation.rb:621:in
`invoke_incrementing_actual_calls_by'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/message_expectation.rb:474:in
`invoke'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/proxy.rb:214:in
`message_received'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/proxy.rb:366:in
`message_received'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/method_double.rb:80:in
`proxy_method_invoked'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/verifying_proxy.rb:161:in
`proxy_method_invoked'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/method_double.rb:64:in
`block (2 levels) in define_proxy_method'
/usr/share/ruby/pp.rb:80:in `width_for'
/usr/share/ruby/pp.rb:91:in `pp'
/usr/share/gems/gems/rspec-support-3.12.0/lib/rspec/support/differ.rb:184:in
`object_to_string'
/usr/share/gems/gems/rspec-support-3.12.0/lib/rspec/support/differ.rb:58:in
`diff_as_object'
/usr/share/gems/gems/rspec-support-3.12.0/lib/rspec/support/differ.rb:20:in
`diff'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/error_generator.rb:299:in
`diff_message'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/error_generator.rb:285:in
`error_message'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/error_generator.rb:68:in
`raise_similar_message_args_error'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/message_expectation.rb:548:in
`generate_error'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/message_expectation.rb:500:in
`verify_messages_received'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/method_double.rb:117:in
`block in verify'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/method_double.rb:117:in
`each'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/method_double.rb:117:in
`verify'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/proxy.rb:165:in
`block in verify'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/proxy.rb:165:in
`each_value'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/proxy.rb:165:in
`verify'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/space.rb:74:in
`block in verify_all'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/space.rb:74:in
`each'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks/space.rb:74:in
`verify_all'
/usr/share/gems/gems/rspec-mocks-3.12.0/lib/rspec/mocks.rb:45:in `verify'
/usr/share/gems/gems/rspec-core-3.12.0/lib/rspec/core/mocking_adapters/rspec.rb:23:in
`verify_mocks_for_rspec'
... snip ...
~~~
End this is the check for COLUMNS env variable:
https://github.com/ruby/ruby/blob/13395757fa7fe0abee3a260d5669baec2dc2e0f...
If there are not issues, then this is probably not executed.
> It very likely something related to RSpec but somthing different then
> kwargs.
It is kwargs after all:
https://github.com/rspec/rspec-mocks/pull/1464
But I am probably not able to untangle this.
> When I applied medicine similar to
>
>
https://src.fedoraproject.org/rpms/rubygem-guard/blob/rawhide/f/rubygem-g...
>
>
> I got past this issue, but with 46 failures due to kwargs. However,
> the recent code is heavily modified by this patch:
>
>
https://github.com/guard/guard/commit/cc8d97a27c824e58aa9251c9feca306b29d...
>
And this patch has not landed yet 🤦🏻♂️
This includes basically completely different patch. But luckily, the
same patch works also for F37:
Vít
>
> which seems to touch every file. So it seems it would be reasonable
> to rebase prior even trying to fix the failures.
>
>
> Vít
>
>
>>
>> rubygem-guard-livereload
>>
>> rubygem-memfs
>>
>>
>> And that actually might be it. I'll try to take look at listen.
>>
>>
>> Vít
>>
>>