- separate zh_hant to zh_hant_tw, zh_hant_hk and zh_hant_sg
- add new locales `zh-hant-sg`, `zh-hant-hk`, `zh-hant-tw`
- fix zh_hant and zh_hans typo and layout
- update `bots/browse/translations.js` `bots/console/translations.js`
`bots/mailman/translations.js` `bots/transactor/translations.js`
rename test
fix test `supported-locales-are-actually-supported` and improve failing message
fix some coding standards
rename old `str` namespace alias to `string`
fixing cs
rename translation id to label and explain checkpoints
remove non-existent labels in translations
add some comments
add status 0.9.12 as first checkpoint in translations, plus current supported locales
remove non-existent label from base checkpoint and fix locale->checkpoint and supported locales
format checkpoint-0-9-12-labels in several lines
add fn labels-that-are-not-in-current-checkpoint
previous behavior was using context (which screen we were on) to
determine if account was new or not.
in order to be able to have more straightforward automation this
commit now checks the list of accounts to get that information
which is more accurate and allows to login without navigating screens.
Fix absolute dir and statement order. Regression appears to have been introduced
during network configuration changes 2-3 months ago and noone checked this file
since then.
This fixes bugs #2520 and #2484
The cause of the bugs is the fact that calling `.sorted` method
on the realm resultset produces another resultset which is sorted
when iterated through `Iterator<T>` interface, but does absolutely
nothing to way how resultset object is serialised (printed out
as JSON).
So when we sorted, printed to json, parsed back through transit
and expected the order to be retained, it was not, and it was
failing quite randomly.
To be precise, it failed in a case where we sorted the resultset
according to some property and picked the top element from the result.
As the top element picking was done AFTER the sorted realm resultset
was converted into cljs datastructure, it was sometimes failing.
This fix just ensures that we select the single element from js
resultset (where sorting info is still retained) and only convert
that single element to cljs (it's also much more efficient).