Fix: edit_user_link rendering

[thedanielhanke]

if config.show_gravatar is e.g. false, "false" becomes part of the render output as false is not nil and therefore not removed on #compact

if there is no edit-user action to be found, the rendered output was 2 nested spans, which messed up layout.

[coveralls]

Changes unknown
when pulling cfd6d67 on thedanielhanke:patch-1
into ** on railsadminteam:master**.

[mshibuya]

Would you mind adding some tests for this fix? They should fit into somewhere around here:
https://github.com/railsadminteam/rails_admin/blob/master/spec/helpers/rails_admin/application_helper_spec.rb#L463-L506

[thedanielhanke]

@mshibuya

concerning the tests: sure, the only issue i had was getting rspec up and running when using sprockets.

it failed at require sprockets/railstie - is this a known thing? will try webpacker later.

as rspec did not start and the current specs just dont really care for the generated html, my first attempt was just to quick-fix the rendering issue.

would love to add some actual improvements (i owe you :))

[mshibuya]

Ah you need to run rspec via appraisal, like this:

$  appraisal rails-7.0 rspec spec/helpers/rails_admin/application_helper_spec.rb

[thedanielhanke]

thanks, working now.

question concerning "#{request.ssl? ? 'https://secure' : 'http://www'}.":

is there any real world requirement for the http://www part? why not always do https here?

[thedanielhanke]

@mshibuya
i did some changes, please have a look.

paused now, as i changed quite a bit - if you are not entirely against these kind of changes, i`ll continue.
kind of changes: imo. readability, maintainability, specs care a bit more about the actual outcome.

the 'include Pundit::Authorization' commit is not PR-related, i can move it to another branch later too.

[mshibuya]

@thedanielhanke

kind of changes: imo. readability, maintainability, specs care a bit more about the actual outcome.

Please don't do it now. You can propose these changes in a separate PR, but do not mix it with the issue fix.

(And I personally think your style of writing specs has more downsides, it's less descriptive and needs more execution time. But this is not relevant here anyway...)

[mshibuya]

Merged in, thanks!