Age | Commit message (Collapse) | Author |
|
|
|
|
|
Superficial cleanup to replace all instances of "gAgentAvatar" with "gAgentAvatarp".
|
|
Superficial cleanup to replace all instances of "LLVOAvatarSelf *avatarp = gAgent.getAvatarObject" with "gAgentAvatar".
|
|
Standardizes representation of volume as [0.0 - 1.0] logarithmic with 0db at 0.5 within LLVoiceClient
Reviewed by Tofu.
|
|
|
|
Minor superficial cleanup to use proper naming conventions.
|
|
Ignore incoming volume level of zero from SLVoice.
(May also fix EXT-5625, though unconfirmed, removes obsolete hack that was obfuscating that issue).
|
|
disabled).
- Added constant holding maximum number of times(in a row) "stateJoiningSession" case for spatial channel is
reached in LLVoiceClient::stateMachine() which is treated as normal. They are counted using member added to
LLVoiceClient. If this number is exceeded we suspect there is a problem with connection to voice server.
Check of current number against maximum was added to LLVoiceClient::voiceWorking().
--HG--
branch : product-engine
|
|
mVolumeDirty needs to be set when recalling volume levels, for both the participant and the session, to trigger sending a volume change message to Vivox.
Reviewed by: Lynx.
|
|
Checker: REVERSE_INULL
Function: LLVoiceClient::startUserIMSession(const LLUUID &)
File: /indra/newview/llvoiceclient.cpp
|
|
- for now volume level is stored in external (vivox) format [0.0 - 1.0]
- this fix should be included into the Beta 2 release
--HG--
branch : product-engine
|
|
|
|
LLVoiceClient::voiceWorking() in general to clear things up a bit.
--HG--
branch : product-engine
|
|
|
|
from parcel with disabled voice)
- One more voice client state is now treated by voiceWorking() as valid.
--HG--
branch : product-engine
|
|
change in the Vivox API
The master voice mute in preferences wasn't working either, due to the minimum scaled volume now being 30 instead of 0.
Reviewed by Lynx.
|
|
Vivox API
Added <Scope>Audio</Scope> to the SetParticipantMuteForMe message.
Reviewed by Lynx
|
|
--HG--
branch : product-engine
|
|
Checker: UNINIT_CTOR
Function: LLVoiceClient::sessionState::sessionState()
File: /indra/newview/llvoiceclient.cpp
|
|
Checker: UNINIT_CTOR
Function: LLVivoxProtocolParser::LLVivoxProtocolParser()
File: /indra/newview/llvoiceclient.cpp
|
|
LLPanelFriends::refreshNames
--HG--
branch : product-engine
|
|
settings per resident between voice sessions)
- moved default internal voice level value into application settings
- cleanned up include list in llmutelist.cpp
--HG--
branch : product-engine
|
|
settings per resident between voice sessions)
- moved necessary functionality from LLMuteList to LLVoiceClient. It was used only in active speackers floater which is deprecated.
- initialized saving/loading of voice level in voice client.
- also saving voice levels between session is activated.
--HG--
branch : product-engine
|
|
disabled) and EXT-4313
(Should not be able to start a call when voice is disabled). Call buttons state now reacts on
voice changes in time.
- Added voiceWorking() method to LLVoiceClient to determine real availability of voice, because voiceEnabled()
doesn't take into account possible errors. Perhaps there is no need in two methods- some investigation will be
made and depending on its results they may become one non-static method. voiceWorking() uses state of voice
client(mState) to determine voice availability. Also some states which are not currently counted by voiceWorking()
as valid may be added if testing reveals problems.
- To enable/disable call buttons in time, LLVoiceClientStatusObserver is used. Its trigger uses states from its
enum only to skip updating button in some states(to avoid button blinking), but to determine button state
LLVoiceClient's voiceWorking() is used.
--HG--
branch : product-engine
|
|
(following by the 7d600b529ee8 commit)
--HG--
branch : product-engine
|
|
--HG--
branch : product-engine
|
|
leave some LLFloaterChat usages commented for later review...
--HG--
branch : product-engine
|
|
|
|
display when users are in the same voice channel)
-- improvements: replace std::vector with std::set to avoid duplicated speaker UUIDs. Added method to get a set of voice participants UUIDs
--HG--
branch : product-engine
|
|
open, movement controlled, etc)
|
|
session 00000000-0000-0000-0000-000000000000does not exist" from Voice Client)
-- added check of session UUID to null before calling LLIMModel::getSpeakerManager (implemented for the EXT-3544 bug)
--HG--
branch : product-engine
|
|
--HG--
branch : product-engine
|
|
controls panel after muting a resident)
--added updating of speakers in speaker manager when LLVoiceClient::participantUpdatedEvent and
LLViewerChatterBoxSessionAgentListUpdates::post events come.
--HG--
branch : product-engine
|
|
We now scale the mic and speaker volume levels to the Vivox range 30
to 70 (default of 50), as recommended by Vivox for their V3 SDK.
|
|
--HG--
branch : product-engine
|
|
The new Vivox SDK supports Automatic Gain Control (AGC), so we
shouldn't need to change the local gain level from the default of 50%.
This change sets the initial gain level to 50 (instead of 62) and
removes the local gain sliders from the Sound Preferences floater and
the Voice Controls floater. The latter was not hooked up yet anyway.
|
|
for incoming group and ad-hoc calls
now is is able to get a direction of calls through the voice channel's signal for its state changing
also choosing of type of IMSession was fixed and was moved from a separate function to the constructor
--HG--
branch : product-engine
|
|
(EXT-2803) Create notifications and dialogs to negotiate P2P voice chat (shown over the Speak button)
(EXT-2806) Create notification windows and dialogs to negotiate Group Voice Chat (docked to Speak button's chevron)
(EXT-2802) Create notifications and dialogs to negotiate AvaLine calls (shown over Speak button)
(EXT-2805) Create notifications and dialogs to negotiate Ad-Hoc calls (shown over Speak button)
- "NO ANSWER" notification gets closed after specified timeout
- added possibility to check direction of call (but it doesn't work for group and ad-hoc calls yet)
- adjusted layout for the incoming call dialog
--HG--
branch : product-engine
|
|
than hardcode it in llvoiceclient.cpp. Also rename the string from
"Unknown" to "Not connected", to better reflect the fact that we get
the Vivox version once voice is connected.
|
|
Cuts number of includes of llnotifications.h from 300+ to 40.
|
|
Fixed many other includes related to this file.
Cleaned out llfloatertestlistview to be an empty test floater.
|
|
|
|
|
|
Don't show the "Start IM" button in the incoming call request for an
Avaline call. And display the phone number of the incoming call as the
avatar name, instead of "(???) (???)".
|
|
|
|
|
|
|
|
visual state as a special case
- all speak buttons will soon know how to poll (sigh) the ptt state, with more reliability.
|
|
with Vivox V2 and V3 SDKs. We were assuming that the <CaptureDevice>
tag contains only a single tag (<Device>), but the V3 SDK returns
other tags such as <DisplayName> and <Type>.
I also added the current Vivox SDK version to the About window so that
it is easier to tell which version your client is actually using.
|