Allow to use_given_order for languages too

[whalebot-helmsman]

No description provided.

[codecov]

Codecov Report

Merging #997 (585a8db) into master (41f9478) will not change coverage.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master     #997   +/-   ##
=======================================
  Coverage   98.29%   98.29%           
=======================================
  Files         234      234           
  Lines        2694     2694           
=======================================
  Hits         2648     2648           
  Misses         46       46           

Impacted Files	Coverage Δ
dateparser/date.py	99.24% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 41f9478...585a8db. Read the comment docs.

[Gallaecio]

Is that all that needs changing? (i.e. was the feature there but this exception was getting in the way?)

Could you add a test?

[whalebot-helmsman]

Is that all that needs changing? (i.e. was the feature there but this exception was getting in the way?)

Yes, e.g. for French format is dd/mm and for English it is mm/dd . English is used as a fallback, so we always add it as last one. You don't have a list of locales, just these two languages. You try to use class and get mm/dd .

To solve it I wrote this class

class OrderedDateDataParser(DateDataParser):
    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)
        self.use_given_order = True

And everything starts working

Could you add a test?

There is a test of loading languages with use_given_order

dateparser/tests/test_loading.py

Lines 21 to 26 in 507dc6d

	param(given_languages=['es', 'fr', 'en'],
	given_locales=None,
	given_region=None,
	loaded_languages=['en', 'es', 'fr'],
	loaded_locales=['en', 'es', 'fr'],
	expected_locales=['es', 'fr', 'en']),

[Gallaecio]

There is a test of loading languages with use_given_order

But it was not failing before. We need a test that would have caught the issue before your change, i.e. a test that fails if we revert your change, and passes after we apply it. And I don’t mean https://github.com/scrapinghub/dateparser/pull/997/files#diff-fd61b87100377a706498644e5ff42b8baecf18918169f6d4082af10d5a053477R658.