1.3.0 - #80
Merged
Merged
Conversation
…build tooling Global-plugin call gestures (NVDA+Windows+A/D/S/F) were running blocking OCR and time.sleep on NVDA's main thread instead of the app module's worker-thread wrappers, defeating the beta7 fix; they now delegate properly. Also fixes a cursor-restore gap in _sendMouseClick, an OCR-buffer leak on LINE exit, and a busy-poll file-dialog watcher. Replaces private-API usage (_gestureMap mutation, direct braille buffer poking) with public NVDA APIs, unifies script categories, adds translatable descriptions to previously undecorated gestures, and adds speakOnDemand to pure-reporting scripts. Raises minimumNVDAVersion to 2024.1. Adds a one-time consent dialog before AI image description uploads a screenshot using the bundled shared API key; users with their own key are never asked. Syncs build tooling with the current NVDA AddonTemplate: speechDictionaries support, PEP 735 dependency groups, prek (replacing pre-commit), lint+test CI gates, tag-based prerelease detection, and removal of redundant one-off build scripts and stale committed artifacts. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
…eImage crash _hasAiConsent/_recordAiConsent: a failed disk write after the user chose Yes left every subsequent call still refusing with "not consented"; add an in-memory session fallback so a grant stays honored for the rest of the NVDA session regardless of disk state. _suppressAddonForFileDialog: replace the two-phase appear-then-disappear wait with a single poll that requires observing the dialog open before declaring it closed, fixing a race where a file dialog slower than 2s to appear caused the addon to resume key interception while the dialog was still opening. script_describeImage: defer the consent dialog (and the rest of the flow) via wx.CallAfter instead of calling it synchronously from the script, which crashed on the first NVDA+Windows+I use. Scripts run on the keyboard hook's call stack; every dialog-opening script in NVDA's own globalCommands.py defers the same way. Also includes the excludedFiles fix from the prior build (excludes __pycache__ from the packaged addon) and its regenerated locale manifests. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
…watchers Consent for one cloud AI provider no longer implies consent for another: _hasAiConsent/_recordAiConsent are now scoped per provider (stored as one provider ID per line instead of a single "1" marker), and the worker chokepoint checks consent against the provider it is actually about to call, so switching provider mid-conversation can't inherit a prior provider's consent. Consent is now asked on first use of each provider regardless of whether the request uses the add-on's bundled key or the user's own: owning a key doesn't answer which company receives the screenshot. Removed the now-unused _isImageDescriptionUsingUserKey and updated the dialog wording, readmes, and changelog in all four languages to match. _suppressAddonForFileDialog's watcher now bails out as soon as its watch token is invalidated, and terminate() invalidates it; previously an orphaned watcher (LINE closed/restarted before the dialog was first seen) could poll a recycled PID for up to 30 minutes into the next LINE session. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
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
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.
…build tooling
Global-plugin call gestures (NVDA+Windows+A/D/S/F) were running blocking OCR and time.sleep on NVDA's main thread instead of the app module's worker-thread wrappers, defeating the beta7 fix; they now delegate properly. Also fixes a cursor-restore gap in _sendMouseClick, an OCR-buffer leak on LINE exit, and a busy-poll file-dialog watcher.
Replaces private-API usage (_gestureMap mutation, direct braille buffer poking) with public NVDA APIs, unifies script categories, adds translatable descriptions to previously undecorated gestures, and adds speakOnDemand to pure-reporting scripts. Raises minimumNVDAVersion to 2024.1.
Adds a one-time consent dialog before AI image description uploads a screenshot using the bundled shared API key; users with their own key are never asked.
Syncs build tooling with the current NVDA AddonTemplate: speechDictionaries support, PEP 735 dependency groups, prek (replacing pre-commit), lint+test CI gates, tag-based prerelease detection, and removal of redundant one-off build scripts and stale committed artifacts.
Greptile Summary
此次更新完成先前意見的修正:
Confidence Score: 5/5
此 PR 看起來可以安全合併。
先前提出的問題均已獲得完整修正。未發現仍存在的阻擋性失敗。
What T-Rex did
Important Files Changed
Reviews (3): Last reviewed commit: "Make AI consent per-provider, always ask..." | Re-trigger Greptile