https://gitlab.synchro.net/main/sbbs/-/commit/4d97d64aebc91297d0d578fe
Modified Files:
src/doors/syncretro/DESIGN.md M4_AUDIO.md README.md syncretro_audio.c syncretro_chunk.h
Log Message:
syncretro: document M4's audio and correct the spec's drop rule

The spec asked for a bounded pending queue that could never fill and
forbade the only congestion signal a stream actually has. Say what the
code does: a chunk is dropped when the staged-output backlog stays
over the threshold across two chunk boundaries, and it is dropped
before it is encoded rather than after it is queued.

That is not the check SyncDOOM reverted. Theirs was instantaneous and
fired constantly, because the frame path keeps the socket busy
essentially always; sustained backlog is a different signal.

Mark M4 done, point DESIGN.md's audio section at the design doc, and
tell the sysop that audio needs a SyncTERM with libsndfile, costs
about 20 KB/s, and is tuned in syncretro.ini. A game sitting on a
silent title screen sends nothing at all, which is correct rather than
broken.

---
þ Synchronet þ Vertrauen þ Home of Synchronet þ [vert/cvs/bbs].synchro.net

https://gitlab.synchro.net/main/sbbs/-/commit/b107c79dd4357f8f6dab312c
Modified Files:
src/doors/syncretro/DESIGN.md M2_INPUT.md retro_env.c xtrn/syncarcade/README.md syncretro.example.ini
Log Message:
syncarcade: fix vector-game claims, pin Battlezone's core options

probe_core measured Battlezone's native vector resolution at 400x300
(120k px/frame, under Pac-Man's 200k budget) and 40fps, not the
1024x768/60fps the docs described -- that figure is the core's own
supersampled default, and antialiasing alone (not translucency) is
what blows out the sixel palette. The template also lacked the two
vector core-option pins the live install already carries, so a fresh
install got the unplayable defaults. Ported those pins into syncretro.example.ini, corrected the pixel-count/control claims in it
and in README.md, and pointed sysops at games.ini's button/stick2
labelling and the arcade profile's I/K second-stick keys.

M2_INPUT.md and DESIGN.md predate the arcade profile's I/K axis keys
and the games.ini answer to MAME 2003-Plus sending no
SET_INPUT_DESCRIPTORS at all; document both, and correct DESIGN.md's
stale claim that descriptor data already drives the help overlay.
Also narrow retro_env.c's M5 TODO so it doesn't read as unaware that
games.ini now covers the no-descriptor case.

---
þ Synchronet þ Vertrauen þ Home of Synchronet þ [vert/cvs/bbs].synchro.net