Align main realtime v2 wire and runtime flow

Port the realtime v2 session, parser, app-server, and conversation runtime behavior onto the split websocket-method base without pulling in the startup-context or TUI playback slices.

Co-authored-by: Codex <noreply@openai.com>
This commit is contained in:
Ahmed Ibrahim
2026-03-16 10:44:10 -07:00
parent 06e299ea56
commit f8c56e66ee
23 changed files with 744 additions and 48 deletions

View File

@@ -176,6 +176,7 @@ async fn conversation_start_audio_text_close_round_trip() -> Result<()> {
sample_rate: 24000,
num_channels: 1,
samples_per_channel: Some(480),
item_id: None,
},
}))
.await?;
@@ -409,6 +410,7 @@ async fn conversation_audio_before_start_emits_error() -> Result<()> {
sample_rate: 24000,
num_channels: 1,
samples_per_channel: Some(480),
item_id: None,
},
}))
.await?;
@@ -518,6 +520,7 @@ async fn conversation_second_start_replaces_runtime() -> Result<()> {
sample_rate: 24000,
num_channels: 1,
samples_per_channel: Some(480),
item_id: None,
},
}))
.await?;
@@ -1469,6 +1472,7 @@ async fn inbound_handoff_request_clears_active_transcript_after_each_handoff() -
sample_rate: 24000,
num_channels: 1,
samples_per_channel: Some(480),
item_id: None,
},
}))
.await?;
@@ -1954,6 +1958,7 @@ async fn inbound_handoff_request_steers_active_turn() -> Result<()> {
sample_rate: 24000,
num_channels: 1,
samples_per_channel: Some(480),
item_id: None,
},
}))
.await?;