[rspec-users] changes in rspec 2.4 break ci_reporter
pathsny at gmail.com
Fri Jan 28 05:24:12 EST 2011
ah. Because people currently depend on the override behaviour? This
seems to be a historical artifact though. I'd expect something that
wants to override other configuration to have to specify that it does,
not the other way around?
so maybe SPEC_OPTS should have an --overide option to override
behaviour specified through other configuration means?
On Thu, Jan 27, 2011 at 3:35 PM, David Chelimsky <dchelimsky at gmail.com> wrote:
> On Jan 26, 2011, at 8:43 AM, vishnu wrote:
>> I use the ci_reporter gem for builds on my CI machine (hudson).
>> Prior to rspec 2.4, environment options would override commandline
>> options. So the gem used this to change the formatter and require its
>> In 2.4, that order has been reversed.
> This has since been restored to its previous behavior, to be released in 2.5. See https://github.com/rspec/rspec-core/issues/276.
>> the thing is 2.4 supports multiple formatters, so there is another suggestion.
>> def parse_options
>> @options = begin
>> options_to_merge = 
>> if custom_options_file
>> options_to_merge << custom_options
>> options_to_merge << global_options
>> options_to_merge << local_options
>> options_to_merge << env_options
>> options_to_merge << command_line_options
>> options_to_merge.inject do |merged, options|
>> the last section, instead of merging hashes, could combine the values
>> of all hashes with the keys "--require" or "--format". So you get the
>> default formatter and the extra formatter that ci_reporter introduces.
> I think I'd want a separate API for this - something like:
> # ~/.rspec
> --format-that-is-not-overridden html
> --out-that-is-not-overridden ~/rspec/this-project/report.html
> But maybe a bit less verbose :) Either that or maybe a marker option like this:
> # ~/.rspec
> --require ~/path/to/file/to/always/require
> --format html
> --out ~/rspec/this-project/report.html
> Then require, format, and out options following the --no-override option would be additive.
> WDYT? Other ideas?
> rspec-users mailing list
> rspec-users at rubyforge.org
More information about the rspec-users