sozialhelden / wheelmap-classic

:wheelchair: Legacy "classic" wheelmap.org (deprecated)
http://www.wheelmap.org
GNU Affero General Public License v3.0
47 stars 16 forks source link

Switching the default locale to :en #688

Open spacekookie opened 6 years ago

spacekookie commented 6 years ago

Addresses an issue discovered in #681 (related #686)

spacekookie commented 6 years ago

@1000miles Any idea about this failure?

1000miles commented 6 years ago

@spacekookie cc @Svenyo

Please keep :de as default locale.

  1. A lot of test cases are based on german locale, if you change it to :en you would need to adjust/fix all the test cases as well. I assume that's why your PR https://github.com/sozialhelden/wheelmap/pull/688 fails (some of the test failures indicate it).

  2. In the past we also wanted to unify the default locale for the wheelmap project AND transifex but as far as I remember @holgerd once said they want to keep the german locale because of historical reasons, e.g. search engines already have :de indexed as default locale.

  3. If we introduce new keys & translations in english without providing the german translations as well, the system would throw an error message like translation key is missing anywhere where no translations in other locales exist yet.

CORRECTED:

My suggestion: Keep it as it is and wait until new translations are coming up for "Datenschutzbestimmungen".

Note: Even though it works at the moment I think for consistency the key privacy should be put to the keys structure for footer instead of header. Here is an example where the structure maps to the translations resource files:

https://github.com/sozialhelden/wheelmap/blob/8df541107ed785f82a2436ebe53d11129e7e2fba/config/locales/en/wheelmap.yml#L351-L353

=> https://github.com/sozialhelden/wheelmap/blob/master/app/views/relaunch/_footer.html.haml#L30

spacekookie commented 6 years ago

@1000miles That would be fine for me. Depends on what @Svenyo thinks.

Also, no I didn't run the tests locally, I was moving from place to place a lot. But…checking tests asynchronously is kinda what failing PR CI is for, isn't it :wink: