fix: load this plugin's locales by absolute path, not via ep_ai_core#20
Merged
JohnMcLear merged 1 commit intomainfrom May 2, 2026
Merged
fix: load this plugin's locales by absolute path, not via ep_ai_core#20JohnMcLear merged 1 commit intomainfrom
JohnMcLear merged 1 commit intomainfrom
Conversation
ep_ai_core/i18n.js's loadLocale() does
require.resolve(`${pluginName}/locales/${lang}.json`)
which only works if `pluginName` is reachable from ep_ai_core's own
node_modules. In every real install ep_ai_core and ep_ai_chat are
sibling plugins, so this lookup silently fails for any ep_ai_chat
key — loadLocale catches the resolve error, returns {}, and t()
falls back to the literal key.
User-visible symptom: chat replies for AI errors say things like
"ep_ai_chat.error_auth" instead of the translated string. Spotted
when the running server was launched without ANTHROPIC_API_KEY in
its env, which makes Anthropic return 401 on every call; the
auth-error branch then sent the bare key to chat.
Fix: ship a tiny i18n.js inside ep_ai_chat that reads
./locales/<lang>.json by absolute path rooted at __dirname. No more
cross-plugin require.resolve and no more silent fallback. index.js
imports from './i18n' instead of 'ep_ai_core/i18n'.
Verified: `node -e "console.log(require('./i18n').t('ep_ai_chat.error_auth'))"`
prints the English string instead of the key.
A follow-up in ep_ai_core could harden its loadLocale to accept an
explicit basePath, but that's a separate repo and a bigger surface;
this PR fixes ep_ai_chat's chat output regardless of what
ep_ai_core does.
ⓘ You've reached your Qodo monthly free-tier limit. Reviews pause until next month — upgrade your plan to continue now, or link your paid account if you already have one. |
JohnMcLear
added a commit
that referenced
this pull request
May 3, 2026
…lay name
Override the default "@<name> " chat-input prefill with the configured
AI trigger ("@ai " by default) when a user clicks the AI's row in the
user list. Without this, clicking the AI chip would prefill
"@AI_Assistant ", which doesn't match anything the server-side mention
extractor recognises.
What this PR adds:
- clientVars server hook exposing { trigger, authorName, authorId } at
clientVars.ep_ai_chat so the client can recognise the AI's row.
- chatPrefillFromUser client hook that returns the trigger string when
the clicked authorId matches the AI's; otherwise returns nothing so
core's default ("@<name> ") wins.
- Mocha unit test (static/tests/backend/specs/chat_prefill.ts) covering
AI/non-AI/missing-clientVars/custom-trigger/missing-trigger paths.
Depends on ether/etherpad#7660 which adds the chatPrefillFromUser hook
to core. On older cores the hook is never fired and this PR is a
benign no-op — graceful degradation, no install requirement bump.
Originally this PR shipped the entire user-list click handler inside
the plugin. Per review feedback, the generic "click a user → prefill
@-mention" UX belongs in core (it's a discoverability win for any
multi-user pad, AI or no AI), so the bulk moved to ether/etherpad#7660
and this PR is now ~40 lines of glue.
(Replaces the previous 3-commit history on this branch with a clean
rebase against the new main, which absorbed PRs #18 and #20 since the
branch was opened.)
JohnMcLear
added a commit
that referenced
this pull request
May 3, 2026
…lay name (#19) Override the default "@<name> " chat-input prefill with the configured AI trigger ("@ai " by default) when a user clicks the AI's row in the user list. Without this, clicking the AI chip would prefill "@AI_Assistant ", which doesn't match anything the server-side mention extractor recognises. What this PR adds: - clientVars server hook exposing { trigger, authorName, authorId } at clientVars.ep_ai_chat so the client can recognise the AI's row. - chatPrefillFromUser client hook that returns the trigger string when the clicked authorId matches the AI's; otherwise returns nothing so core's default ("@<name> ") wins. - Mocha unit test (static/tests/backend/specs/chat_prefill.ts) covering AI/non-AI/missing-clientVars/custom-trigger/missing-trigger paths. Depends on ether/etherpad#7660 which adds the chatPrefillFromUser hook to core. On older cores the hook is never fired and this PR is a benign no-op — graceful degradation, no install requirement bump. Originally this PR shipped the entire user-list click handler inside the plugin. Per review feedback, the generic "click a user → prefill @-mention" UX belongs in core (it's a discoverability win for any multi-user pad, AI or no AI), so the bulk moved to ether/etherpad#7660 and this PR is now ~40 lines of glue. (Replaces the previous 3-commit history on this branch with a clean rebase against the new main, which absorbed PRs #18 and #20 since the branch was opened.)
|
🎉 This PR is included in version 1.2.0 🎉 The release is available on: Your semantic-release bot 📦🚀 |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Bug
Users see literal i18n keys in chat (e.g.
AI Assistant: ep_ai_chat.error_auth) instead of the translated message.Root cause
ep_ai_core/i18n.js'sloadLocale()does:That lookup runs from
ep_ai_core's context, but in every real installep_ai_coreandep_ai_chatare siblings underplugins/, soep_ai_chatisn't reachable fromep_ai_core/node_modules/. The resolve throws, the catch returns{}, andt()falls back to the literal key.Reproduced cleanly:
Spotted when the running server was launched without
ANTHROPIC_API_KEYin its env, which makes Anthropic return 401 on every call. The auth-error branch then sent the bare key to chat.Fix
Ship a tiny
i18n.jsinsideep_ai_chatthat reads./locales/<lang>.jsonby absolute path rooted at__dirname. No cross-pluginrequire.resolve, no silent fallback.index.jsnow importstfrom./i18ninstead ofep_ai_core/i18n.A follow-up in
ep_ai_corecould harden itsloadLocaleto accept an explicitbasePath, but that's a separate repo and a bigger surface; this PR fixesep_ai_chat's chat output regardless of whatep_ai_coredoes.🤖 Generated with Claude Code