search

yodalee / mozc

Mozc - a Japanese Input Method Editor designed for multi-platform
BSD 3-Clause "New" or "Revised" License
0 stars 0 forks source link

For some? machine-dependent character ("機種依存文字"), it doesn't report that when U+xxxxx is typed for the first time. #76

Open yodalee opened 6 years ago

yodalee commented 6 years ago 
I noticed that there is slight behavioral difference between first-time and 
afterwards, that when we type U+xxxxx to convert a character based on codepoint 
it may or may not report the warning that the character is machine-dependent 
(機種依存文字).

What steps will reproduce the problem?
1. Enter mozc hiragana mode.
2. Type U+2F81E.

What is the expected output?

- See this screenshot: https://twitter.com/atsushieno/status/552138494982643715

What do you see instead?

- See this screenshot: https://twitter.com/atsushieno/status/552136801335267328

What version of the product are you using? On what operating system?

- Ubuntu 14.04, mozc-server 1.13.1651.102-2, fcitx-mozc 1.13.1651.102-2

Please provide any additional information below.

- Should be cosmetic issue, but this kind of behavioral differences could be a 
glitch when reporting bugs :-)

Original issue reported on code.google.com by atsushi...@gmail.com on 5 Jan 2015 at 4:37

yodalee commented 6 years ago 
> 2. Type U+2F81E.

I mean, literally type those ASCII characters.

Original comment by atsushi...@gmail.com on 5 Jan 2015 at 4:39

yodalee commented 6 years ago 
Thank you for the report.  I can confirm the glitch you reported with the 
following steps on Ubuntu 14.04 with ibus-mozc 2.16.2013.102.

1. Enter mozc hiragana mode.
2. Type "U+2F81E".
3. Hit space key to convert the preedit text.
4. Hit backspace key to revert the converted text to the preedit state.
5. Hit backspace to delete the last character of "U+2F81E" with keeping preedit 
state.
6. Hit E key to add the "E" at the last of "U+2F81" with keeping preedit state.

Basically Mozc tries to add annotations like <機種依存文字> only for 
conversion results, not for suggestion results.  So the expected behavior is 
that <機種依存文字> is shown only after step 3, where you intentionally 
entered conversion state by hitting space key.  The state after step 2, 4, 5, 
and 6 is the so-called "suggestions" hence <機種依存文字> shouldn't be 
shown there.

However, as you reported, we see <機種依存文字> even after step 4 for 
some reasons. So the expected behavior and actual behavior would be:

Expected behavior:
The candidate is annotated with <機種依存文字> after step 3 only.

Actual behavior:
The candidate is annotated with <機種依存文字> after step 3 and step 4.

Original comment by yukawa@google.com on 5 Jan 2015 at 11:32

yodalee commented 6 years ago 
Oops, I didn't intend to change the issue title.  Revert to the original one.

Original comment by yukawa@google.com on 5 Jan 2015 at 11:34