engine: Use annotations/en.xml from unocode.org but not emoji-list.html

engine: Use annotations/en.xml from unocode.org but not emoji-list.html

Downloading emoji-list would cause a different build by build site.
Now save annotations/en.xml from unicode.org and get emoji.json from
Emoji One.
en.xml is used for Unicode annotations and emoji.json is used for
aliases_ascii, e.g. ":)", and category, e.g. "people".

BUG=https://github.com/ibus/ibus/issues/1865
R=Shawn.P.Huang@gmail.com

Committed: 0ee1896a2b3e75494f8f9fd9d04c27436f0877b8

[Peng, 2016-07-05 12:45:38]

On 2016/07/05 05:28:17, fujiwara wrote:

LGTM. And I imagine using vala or python may make the emoji-parser more simpler.

[fujiwara, 2016-07-06 10:49:24]

On 2016/07/05 12:45:38, Peng wrote:
> On 2016/07/05 05:28:17, fujiwara wrote:
> 
> LGTM. And I imagine using vala or python may make the emoji-parser more
simpler.

Currently I'm thinking if https://github.com/ibus/ibus/issues/1819 can be
handled with this way because GtkCompose outputs one char only and it's a bit
complicated for me to modify the algorithm.

If we convert $HOME/.config/ibus/emoji.json or emoji.xml to
$HOME/.cache/ibus/dicts/emoji.dict, I think emoji-parser will be merged to
libibus.so.

[fujiwara, 2016-07-07 03:47:51]

Committed patchset #2 (id:20001) manually as
0ee1896a2b3e75494f8f9fd9d04c27436f0877b8 (presubmit successful).