Change-Id: I957d883d1ccf188cb98eb11031aa0a32fce4ed76

Also fix a typo, and increase the time allowed to load the main
dictionary to 2 secs instead of 1.

Bug: 6114326
Change-Id: I5f70b34fc4277c55977b18466253152aa98a2507

Bug: 6023947
Change-Id: I779da32708fca7333aff74929cc98173246ee685

Bug: 4343280

Change-Id: I17f8f6295b01900948b98680d0267753f33a46cf

Bug: 6023947
Change-Id: I53a5c0eb99100a0dfe3e025808f5dc55747d1287

Change-Id: I98e77e5432a228e4455998bf09a2cc671b113c4e

Bug: 6106407
Change-Id: I43d9c89610a9ed7a9f6825bd90334d7b343df52c

Bug: 6121571
Change-Id: Ib16fbef1087da3b7da5539901ec57156f242f089

\o/

Bug: 5975068
Change-Id: Ia18b55af6d555e7ee7af1cdd7c6227fc1ee6786b

Change-Id: Ie86675ff2facd37430051bf129ddd930fb933dc9

Change-Id: I9d2f053b6f7bad20fc6afe5f5086a965e5014a5c

This duplicates the functionality in ComposingStateManager at
a fraction of the cost. It's not used yet.

Change-Id: Ica2607fb2f7c41e11189fcb566e253db8f62596f

Bug: 5759092
Change-Id: Ic7d7d4812976654c0e85fecabd1128eadfc52c5c

Change-Id: Ifcc79357adcc53f93b190b6e98d5569f42f74655

This exports some voiceproxy complexity to voiceproxy rather
than keeping it in latinime

Change-Id: I0b3523e77af4104cc3ed28417956059c734c8bfe

Change-Id: I2c8b093b59ad36ffe860c3c4d360d87251d101c4

Bug: 4343280

Change-Id: I0f6a7e6912ed4abea07c10d266da4c7ccb0dae76

This fixes two defects:
- One where some separators should promote a phantom space to
a real space. This bug had a unit test but wasn't tracked.
- One where French punctuations would behave incorrectly. Tracked
Bug: 6113693

Change-Id: Ia7f86fc960e00141757632ab2c9bce9168dd6966

Change-Id: I1b0f7b857e89307c987187c1969a2846aa97fdcc

Change-Id: I72a2f71d7f6565a30c06e563b0a64f496542cde9

Change-Id: I894c0850e803bdfc7b590a62ddd4076a3a4e4658

This reverts commit c208f4df

Change-Id: I2f92c01fa1ff6b318128d387f26942b5d47de3e0

Change-Id: I4b4dc6f86d659491d8b9a74ed120092eab964345

Bug: 6114350
Change-Id: I419bc1736c27046dfea254ac510f5da478b6412c

Change-Id: I26bd82aee5ead84e40abfc3db5a48ed6d1e42361

It makes more sense to catch words with the safety net in
updateSuggestions() than in showSuggestions(). Also, it's useless
to do it in updateBigramPredictions, because the result will
always be provably false.
This will also help with writing unit tests.

Change-Id: Ifab81e6c5a74b1aa75cf85ef7742fee69a98f5ce

This is a followup change to Ie10b3e0b.

Change-Id: I1f91d96e32357b83bd0546537d561d638b6ff0c6

Change-Id: I11f87448f689cf828bff960ad7a3e4739a3457dc

Change-Id: I2002662ee07af2fcc626a52ea4e9ad8977b29ca2

This is provably always false

Change-Id: I88d783fa59fec0b7a729558ba974a6a58c158645

With the current logic, this test provably always resolves
to false.

Change-Id: Iba84ef5f3a14cd8ac48c6a813519ab283bf605f8

This is follow up of  Ie10b3e0.

Change-Id: Idefbe1575915bb87f1c86cb3ddb5efe578c5b760

Change-Id: Ic837018c702c8c575de3e5e81fe8e31fee16fe22

This issue has been introduced by I2bd4daf2

Change-Id: Ie10b3e0bb360986597a0f5a8e5b8253400f924c4

Bug: 6080806
Change-Id: Iacd78c6aab6abf948717d34da7489a79c41316d2

Change-Id: I445277cb3af469a0145a8da0ffb801b7e19048cd

Change-Id: I038e6a82b6463a2a8ec494171a596d76fbc75fe1

I wish "are we autocorrecting?" was not computed in a dozen
places all depending on a hundred code paths
More than likely, this fixes very subtle discrepancies
between auto-correction indicator with the underline and with
the LED on the spacebar - which is not displayed any more in
the current version anyway. Especially, the LED probably
would have been off when the word was caught by the safety net.

Change-Id: Idda3021771081d6155b06915e728ecd64d9e042e

Avoid doing some useless processing.

Change-Id: I4ce0a188e10db30322bd25751dc5c09492cea301