fix: turning JoyAI audio memory off clears the H3 v2a chip

The gate added in f40ae82 was right, but fpActionbar never listed
joyai_audio_memory — so switching it off changed nothing the fingerprint
could see and the actionbar zone skipped its repaint, leaving the chip
stranded next to Generate until some other field forced a redraw.

Adds joyai_audio_memory, plus models.disable_core (every chip is gated on
the Chain Sampler running at all), and reads the models config through the
same state.models || project.models fallback PipelineCaps uses.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This commit is contained in:
digital-garbage
2026-08-05 11:36:29 +03:00
co-authored by Claude Opus 5
parent 7af38e2e55
commit 05a97ada2d
+8 -2
View File
@@ -130,10 +130,16 @@
sampler: st.project?.sampler_slot,
// The MiniMax H3 "this setting can't run" chips live in the same zone, so every
// input they read belongs here too — switching the family or turning one of these
// off must clear its chip, not leave it sitting next to Generate.
family: st.models?.model_family,
// off must clear its chip, not leave it sitting next to Generate. That includes
// the models config every chip is gated on, read the same way PipelineCaps reads
// it (state.models in the Editor, project.models otherwise).
family: (st.models || st.project?.models)?.model_family,
core: (st.models || st.project?.models)?.disable_core,
h3Bounded: st.project?.sampler_inputs?.bounded_attention_enabled,
h3Detail: st.project?.sampler_inputs?.segmented_detailing,
// v2a_grad_scale only warns while joyai_audio_memory is on, so the toggle is part
// of the chip's message — without it, switching JoyAI off leaves the chip stranded.
h3Joyai: st.project?.sampler_inputs?.joyai_audio_memory,
h3V2a: st.project?.sampler_inputs?.v2a_grad_scale,
h3Pass2: st.project?.sampler_inputs?.second_pass,
h3Pass2Op: st.project?.sampler_inputs?.second_pass_op,