Change Ibef3b834 increased the size of buffers to address Bug: 2537307. This was not an ideal solution, because it consumes more memory and is still vulnerable to failure if package names or class names get larger. Change I127c860b fixes the problem in a different way, so we can now shrink the buffer size back to the original value. Change-Id: I67db90017bb84b2c4a4ff962e5a524a9a32b67e8
Bug: 2537307 The recognizer has a limit on the size of each semantic value in the grammar. Now that we are storing both the package name and class name there, it's pretty easy to overflow the limit. This change doubles the size of the buffer, making this overflow less likely. Change-Id: Ibef3b834cf867e3e4ad7678289460e451e704f2f
Both of these grammars now support the commands "exit" and "cancel", which will result in the voice dialer exiting.
Bug#232200 This grammar accepts the four phone types supported by the VoiceDialer: "home", "mobile", "work", and "other", along with "retry" and "exit". It is used when the voicedialer has determined which person to call, but is not yet sure which phone to call them on. Also get rid of some of the uneeded code from the boolean grammar. It was allowing for changes to be made at runtime, but that is not needed.
…port the new flow for the bluetooth voice dialer."
new flow for the bluetooth voice dialer.
Merge commit 'b888ee1007939e6bcc6e9693edea404219b17e18' * commit 'b888ee1007939e6bcc6e9693edea404219b17e18': eclair snapshot Fix bug 2243155 Set VoiceDialer to use the VOICE_RECOGNITION audio source.
…ION audio source. Merge commit '3133210f7381d7510af72a72a7472ab4fdaef698' into eclair-plus-aosp * commit '3133210f7381d7510af72a72a7472ab4fdaef698': Fix bug 2243155 Set VoiceDialer to use the VOICE_RECOGNITION audio source.
Currently this library is linked via dependency of another shared library. While this works, it is not the right thigh to do.
…coustic model Merge commit '7d56aaf97085593d3aa8080451783fc080731539' into eclair-plus-aosp * commit '7d56aaf97085593d3aa8080451783fc080731539': Fix crash in voice recognizer when it fails to load the acoustic model