This fixes the input logic tests that were broken and suppresses
their dependency upon the spell checker proximity. Instead, it
gets the Keyboard instance from Latin IME and uses the actual
coordinates, which results in a test run closer to what actually
happens during typing.
Change-Id: I3a81d249ee7fb3ac6ae6940aa2e8b2421e829e5c
The scripts/languages such as Georgian, Hndi have no notion of case.
In such scripts/language, shift locked keyboard should be the same of
manually shifted keyboard.
Change-Id: Iadea3a332674a52188de7f1566727e4b56dced8c
This wasn't very obvious, but the actual values are critical, so
make it more readable
Bug: 6181080
Change-Id: I07461d2df4c796dad0cb6f5283e41a28036d023c
Follow up to I4d2ef504. Address a compiler warning and a small optimization as well.
bug: 6188977
bug: 6209651
Change-Id: Ibc9da51d48ebf0b8815ad0bb2f697242970ba8f7
The new format has time, time as int, and the new logGroup code
(m=motionEvent, k=key, s=statechange, c=correction)
followed by additional text whose format depends on the logGroup code.
The additional text will be cleaned up in later CL's.
20120321-170548.067+0900 1332317148067 m
[Down]143116235,0,691,378,0.40000004,0.435
20120321-170548.133+0900 1332317148133 m
[Up]143116302,0,691,378,0.40000004,0.435
20120321-170548.146+0900 1332317148146 k <enter> 691 378
20120321-170602.485+0900 1332317162485 s
[onStartInputView]com.socialnmobile.dictapps.notepad.color.note,2131624043,147457,1140850694
20120321-170606.508+0900 1332317166508 m
[Down]143134679,0,40,154,0.4666667,0.64
20120321-170606.567+0900 1332317166567 m
[Move]143134723,0,42,154,0.4666667,0.765
20120321-170606.577+0900 1332317166577 m
[Move]143134746,0,44,154,0.4666667,0.775
20120321-170606.652+0900 1332317166652 m
[Up]143134824,0,44,154,0.4666667,0.775
20120321-170606.662+0900 1332317166662 k A 40 154
20120321-170607.200+0900 1332317167200 m
[Down]143135372,0,411,390,0.53333336,0.585
20120321-170607.311+0900 1332317167311 m
[Up]143135483,0,411,390,0.53333336,0.585
20120321-170607.313+0900 1332317167313 c [----]A
Bug: 6188932
Change-Id: I8694eb9016d8cf0389ef582f6c7d2820aa4d2c92
Having it in a translatable file would sometimes cause the translation
importer to try to replace the locale-specific configuration file.
Change-Id: Ib450454b63c1a4de8b85905561a59224b9410147
UserBigramDictionary -> UserHistoryDictionary.
Also update all methods, strings and comments, except those needed
for backward compatibility (which only include the name of the
database file).
Change-Id: I0bccea29880dc566b90100575b83baaa947b03ae
UserBigramDictionary now assumes both functionalities. It will
be renamed to UserHistoryDictionary in a future change.
There are several reasons to do this. First, there is a lot of
duplicate code in User{Unigram,Bigram}Dictionaries that are
factored by the few lines of code in this change. Also, other
dictionaries like BinaryDictionary or ContactsDictionary all
assume both responsibilities, as should be the case
theoretically.
It is also possible to do this because previous versions don't
write any unigram data that we'd want to reuse. For even older
versions that do write data, we can't really make any sense out
of it. Bigram data however can be useful, and this allows us to
reuse it easily.
Change-Id: I755525f92744e1536eaef097527e8151b7859a30
