这个版本更新了什么#
Hover reads Chrome's text. Hovering a word in Google Chrome used to fall back to reading the screen's pixels on every hover. XiaolaiDict now asks Chrome for its page text: the first hover after Chrome starts still reads the pixels while Chrome prepares its text, and every hover after that reads the words themselves — whole sentence included.
A hover is one request, start to finish. A word read after you had already moved on, or after you pressed the lookup shortcut, no longer replaces the lookup you made since. A reading is recorded only for a card that was actually drawn on screen — including when the lookup window is reused for the next word.
Hover stops waiting on apps that do not answer. Reading text from another app now has a time budget and stops at the first app that does not reply, instead of asking it again and again. A screen capture that gets stuck is reported in Settings rather than silently blocking every hover after it.
Site exclusions are enforced before anything is read. An excluded site is refused before any of its text or pixels are read, and a page whose address cannot be read is refused rather than assumed to be somewhere else. Typing example.com:443 excludes example.com.
Fixes
- 悬停操作不再读取 XiaolaiDict 自己窗口下方的应用内容。
- 整次查询只使用同一本学习词典，因此你看到的卡片和被标记的释义来自同一本词典。
- 如果一个短语的词典无法全部读取，卡片会明确说明这一点。
- 无法检查某项权限时，设置窗口不再显示所有设置都已完成。
- 在读取记录之前保存的释义，会保留保存时使用的语言。
- 同一个单词的卡片未能显示后再次悬停，现在会重新查询该单词。
以上译自作者写在 GitHub 上的英文更新说明。查看原页面