chore(l10n): translate Tor connection guidance - #2586
Conversation
|
Claude finished @ethicnology's task in 6m 19s —— View job Review: Tor translation keys across 26 locales
SummaryThe mechanical claims in the PR description hold up. No correctness, security, or regression blockers — a few worthwhile observations below, the first of which questions whether one of the translated keys should exist at all. Verified (all pass):
Findings1. 2. The new English text of that key is ambiguous about RecoverBull, and the ambiguity propagates into all 26 translations. 3. Register inconsistency in French and German (minor, translation quality).
The PR already flags translations as pending native-speaker review, but these two keys are the concrete instances to catch in that pass. 4. Cosmetic, French only: I couldn't run |
49bb155 to
e559a56
Compare
e2049ca to
9e0ebb2
Compare
9e0ebb2 to
ce5819b
Compare
ce5819b to
76b2f14
Compare
6020cd5 to
76bf06a
Compare
76bf06a to
c747d8a
Compare
c747d8a to
073e74b
Compare
073e74b to
75417c1
Compare
75417c1 to
a5bafe0
Compare
a5bafe0 to
56388c7
Compare
56388c7 to
285197f
Compare
285197f to
e7d023e
Compare
Stacked on the
feat/tor-recoverbull-status-uiPR. Review that one first.The Tor work in this stack added its keys to the
app_en.arbtemplate only, so 26 otherwise-complete locales fell back to English on the Tor settings and RecoverBull connecting screens — the two screens a censored user is most likely to be reading, and the least useful ones to serve in a language they may not speak.All the new keys are now present in every locale. Applied with
tools/arb.dartrather than by hand, so untouched keys stay byte-for-byte identical and the diff is limited to the added entries.Two things a reviewer should know:
audit-identicalwill flag as expected rather than as a miss.Placeholders survive in all locales —
audit-placeholdersreports no mismatch.