The test done in this PR checks #13969 has been fixed, but there is no test to check that #13966.

@LittleStar-VIP, would you mind test the build of this PR and check that all the descriptions are reported for a character if you enable delayed character description? Thanks.

@CyrilleB79
Thank you for tagging me, but sorry, where can I get the pr build of nvda here? I cannot find the link to download.

@LittleStar-VIP Here it is: https://ci.appveyor.com/api/buildjobs/gmt9bpa7aaud0yvv/artifacts/output%2Fnvda_snapshot_pr13972-26067%2C8c74c19b.exe

@seanbudd
I tested this pr and there is a very strange result.
After creating a portable copy with the link provided above, I enabled the delay character description in nvda setting.
I test the function with the same word "香" which I had mentioned in the issue.
This time, both delay character description and pressing numpad 2 generate the same character description
however, both tw and hk locale still generate the same character description for the character 香 and the character description is
香 香港 香蕉 香煙 蚊香 香味 香料 香水 芳香 檀香 中國香港 在香港以外

After comparing the character description of the respective folder, the above line of character description is found in the file "X:\NVDA PR\locale\zh\characterDescriptions.dic"

I remember somewhere else mentioned that the character description here is a no longer maintained file of character description.

@LittleStar-VIP - if the description is consistent across gestures, I think #13966 is fixed.
The changes since 2022.3beta1 should have no affect on the numpad2 character description.
zh and zh_tw both have the same description for 香 in locale\zh\characterDescriptions.dic and locale\zh_tw\characterDescriptions.dic: 香 香港 香蕉 香煙 蚊香 香味 香料 香水 芳香 檀香 中國香港 在香港以外

@seanbudd
Please clarify with the character description maintainer, to my understanding, the cn, hk and tw character description are maintained by different group of people.
So it should not generate the same set of character description under hk, cn or tw locale.

the character description for the word 香should be "香 香水 香港 香味 芳香 百里香 焚香頂禮" in tw locale
the above line of text is copied from the file X:\NVDA 2023.1 Beta 1\locale\zh_TW\characterDescriptions.dic

This PR is blocked by a test sample and STR for #13966, so that the behaviour for the sample can be compared with 2022.2 and this PR.

I have created a new issue regarding the character description problem which I mentioned above
#13978

as the problem described in #13978 is not specific to delayed character description, I'm unblocking this PR and suggesting it closes #13966 as is