When trying to enter a magic text in Chinese, we often end up with the wrong content. Interestingly, the text area underneath has the correct chinese text. So - there is obviously some race pushing this into the core and/or diffing the text. We end up with phonetic characters (?) raw in the document instead of having them combined into the proper characters. Present on master as of now.
Created attachment 157615 [details] log in console of Chromium browser when typing Chinese
Created attachment 157617 [details] screenshot Ctrl-space to launch the input method engine (gcin with phonetic input method) input 測試中文 The keystroke sequence: hk4 3g4 15j/ 1jp61 e.g., to input 測 h appears ㄘ in the input method small window k appears ㄜ in the input method small window 4 appears ˋ in the input method small window, and started the word selection <space> to switch to the next page of word selection 3 to selected the 3rd word in the list (測) But most of the time, the Chinese character would stay only in ㄘ not the whole word 測
Thanks for the report, potential fix here: https://gerrit.libreoffice.org/c/core/+/87928
Merged the fix - would love some feedback Franklin, there should be nightly CODE snapshots with that in it to test soon.
Michael Meeks committed a patch related to this issue. It has been pushed to "master": https://git.libreoffice.org/core/commit/80a1c69a30a55ed648a43976f6ab401ac2b5ceaa tdf#130382 - process key-event handlers synchronously for unipoll. It will be available in 7.0.0. The patch should be included in the daily builds available at https://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: https://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
*** Bug 129692 has been marked as a duplicate of this bug. ***
(In reply to Michael Meeks from comment #4) > Merged the fix - would love some feedback Franklin, there should be nightly > CODE snapshots with that in it to test soon. Details on the snapshots is available at: https://www.collaboraoffice.com/code/tryout-online-snapshot/
Confirmed fixed at LOKit 6.2.10.4-4 (git hash 01f0d42). Thanks.
Great news =)
Would love testing of characters that are not in the BMP if that's easy: https://stackoverflow.com/questions/5567249/what-are-the-most-common-non-bmp-unicode-characters-in-actual-use =)
Michael Meeks committed a patch related to this issue. It has been pushed to "master": https://git.libreoffice.org/online/commit/10caa82b0a2c90d8b7ee1c05ca9a928815c6873c tdf#130382 - unit test for core fix for text input ordering issue.