- Extract audioStreamGetPanFactors() into the shared layer, replacing
identical pan-to-LR math duplicated in the PSP and Dolphin backends
(and dropping a dead clamp - directionality's int8_t range already
guarantees pan stays in [-1, 1]).
- Implement audioStreamSetPosition() for real (was previously a no-op
that computed a value and threw it away) and add
audioStreamSetLoopPoints() to actually drive loopStart/loopTo, which
were previously dead fields with no setter at all. Both are threaded
through all three platform backends:
- PSP: the feeder thread now bounds each pass by the loop segment
and wraps to loopTo instead of always frame 0, while still filling
hardware chunks gaplessly.
- Dolphin: loopTo/loopStart map directly onto ansnd's existing
loop_start_offset/loop_end_offset, and startFrame onto start_offset.
- Linux: Buffer() now queues only the current segment, clearing the
SDL queue on an explicit seek but preserving the existing
overlap-based gapless loop restart otherwise.
- Linux: skip the mix scratch-buffer entirely at full volume, queuing
stream->data directly instead of allocating/zeroing/copying into one
every buffer call for no reason.
Verified no regression in the default (no seek, no loop points) case on
Linux/PPSSPP/Dolphin (-a LLE), and verified seek + loop-region behavior
manually via a temporary engine.c smoke-test tweak (reverted) showing
the expected faster loop cadence and seek-then-loop sequencing.
ansnd_configure_pcm_voice requires frame_data_ptr to be a physical
address - it rejects cached virtual pointers (0x8xxxxxxx) as
ANSND_ERROR_INVALID_MEMORY. Convert via MEM_VIRTUAL_TO_PHYSICAL and flush
the CPU cache first so the DSP's DMA read sees what was actually written.
Root-caused by disassembling libansnd's ansnd_configure_pcm_voice (no
source available, only the static lib) and confirming via a temporary
debug print that frame_data_ptr's top bit being set was what tripped the
check. Confirmed fixed by running the Wii DOL build in Dolphin under -a
LLE - voice configure/start now succeed with no errors.
Splits the asset archive into a compressed (DEFLATE) zip and an
uncompressed (STORED) zip back to back behind a small header, instead of
one plain zip. DEFLATE-compressed zip entries aren't reliably seekable in
libzip, which caused locale string lookups (repeated rewind/reopen of the
same entry) to silently skip content on Dolphin specifically. Uncompressed
entries don't have that problem, so locale files now go in the stored
archive without needing to buffer the whole thing in memory.
Adds tools/asset/pack as the new packer (replacing the plain
`tar --format=zip`), a shared assetdsk.h/.c opener used by all platforms,
and a second ASSET.zipStored handle with fallback lookup in assetfile.c.
Confirmed working: Linux, PSP (PPSSPP), and Dolphin-FAT (Wii DOL under
-a LLE) - the original Dolphin locale lookup failure no longer reproduces.
Root cause of the crackle (confirmed on real PSP hardware): the
persistent feeder thread called stream->onLoop directly in the middle
of its tight chunk-feeding loop. onLoop can do arbitrary work (the
test callback does consolePrint, which locks a mutex, moves the
console history buffer, and fflushes stdout) - if that takes anywhere
close to one chunk's playback time (~23ms), the next chunk isn't
ready and the hardware channel starves. Audio callbacks must never be
invoked directly from a real-time audio thread.
Fixed generically: audiostream_t gains loopCount (incremented by
platform code from whatever context it runs in) and lastLoopCount
(main-thread-only bookkeeping). audioStreamUpdate() detects the
change and fires onLoop safely from the main thread, regardless of
which thread/interrupt actually noticed the loop. PSP's feeder thread
now increments the counter instead of calling onLoop inline.
Also eliminated the ~21.7ms of real silence padding baked into every
PSP loop pass (found while chasing the timing gap that preceded the
crackle fix): the final chunk's padding is now filled with the start
of the next loop instead of zero, avoiding sceAudioSetChannelDataLen
(the likely real cause of an earlier, separate click) while keeping
every output call the same constant size. Loop period measured via
PPSSPP is now ~1.000-1.002s for a 1.000s tone, down from a consistent
~1.02-1.03s before.
The PSP feeder thread is also now persistent for the stream's whole
lifetime (created once in Init, idles between plays) rather than
respawned via threadStartRequest on every single loop restart - real,
avoidable OS thread creation overhead that was contributing to the
gap before the padding was identified as the dominant cause.
Brought Dolphin in line architecturally rather than mirroring PSP/
Linux's restart-and-detect approach: ansnd_pcm_voice_config_t has
native loop_start_offset/loop_end_offset fields, so a looping Dolphin
voice loops entirely in DSP hardware with zero host involvement at
the loop boundary - no restart latency to create a gap in the first
place. Trade-off, clearly documented in code: onLoop never fires for
Dolphin this way (no ANSND_VOICE_STATE for "wrapped") and it requires
cleanly-authored loop content (no per-wrap fade like PSP's, matching
the same assumption). Compiles cleanly for both gamecube and wii;
not yet verified on real hardware.
Confirmed on real PSP hardware: no more gap, crackle fix pending
final hardware confirmation.
Allocates an ansnd voice per stream and configures/starts it in
single-buffer mode (no continuous re-feed needed yet, matching the
PSP/Linux single-shot approach). Real linear panning from
directionality via left/right voice volume. ansnd has no polling API
for voice state, so finish detection uses a voice_callback (fired
from ansnd's own audio DMA interrupt, not the main thread) that sets
a flag audioStreamDolphinIsFinished() reads.
Links libansnd (new for this project) and wires duskdolphin/audio
into the CMake build for the first time.
Verified via the dusk-dolphin toolchain: both GameCube and Wii
targets compile and link cleanly. Runtime verification in headless
Dolphin was inconclusive - confirmed via an A/B test against a
pre-audio baseline build that a pre-existing MMIO flood (unrelated
to this change, reproduces identically with zero audio code present)
keeps the harness from reaching the game's own steady state before
timing out.
Cross-platform audiostream/audio API with per-platform hooks for
PSP/Dolphin/Linux. Linux is fully wired end-to-end through SDL2
(device open, volume-mixed queueing, finish detection via
SDL_GetQueuedAudioSize) and verified playing a generated test tone
in engine.c. PSP and Dolphin hooks are stubbed for now.
Memory sticks/cards don't suit N+1 separate save files the way Linux's
filesystem does, so add an opt-in SAVE_DEVICE_DATA_RAW mode: settings and
all save slots get serialized to JSON, concatenated, zlib-compressed, and
framed with a magic/version/checksum/generation header, then read/written
as a single blob through one combined platform hook instead of four.
- PSP and Dolphin's SD/NAND backends write via a temp file + atomic rename,
so a crash mid-write can never leave a half-written save behind.
- GameCube memory cards have no rename or resize primitive, so they
ping-pong between two fixed files instead, picking whichever is valid and
has the higher generation counter on load.
- Linux is untouched (still four separate JSON files); saveslot.h/
savesettings.h drop their now-unnecessary pack(1) now that nothing
persists them as raw bytes.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
CARD_Mount() alone reliably reports CARD_ERROR_NOCARD for a card physically
present in slot B - unlike slot A, which the IPL polls automatically at
boot, slot B needs an explicit CARD_Probe() first (the standard pattern in
every official devkitPro CARD sample). Also fixes
saveDeviceDolphinCardHasFreeSpace() treating CARD_GetDirectory()'s
CARD_ERROR_NOFILE (a totally empty card, not an actual error) as a hard
failure, which made any card with zero existing save files - slot B's in
this case, since slot A already had leftover test data on it - get
reported as full.
Confirmed via Dolphin with slot A disabled/slot B set to a memory card:
previously reported "no save device found", now correctly detects it.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
PSP checks memory-stick reachability via sceIoGetstat. GameCube gets two
devices (CARD_SLOTA/CARD_SLOTB), each mounted via CARD_Init/CARD_Mount and
checked for free space via CARD_GetDirectory/CARD_GetBlockCount. Wii picks
between three storage methods at compile time (DUSK_SAVE_WII_METHOD =
NAND/CARD/SD in wii.cmake, default NAND via ISFS) since real hardware
behavior for the default is unverified.
Also fixes two pre-existing bugs found while build/runtime-testing this
across PPSSPP and Dolphin: wrong libogc language macros in
systemGetLocaleDolphin, and a missing SYS_STDIO_Report(true) call that
silently swallowed all guest console output in Dolphin.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Removes the per-platform save/savestream/autosave implementations, the settings UI frames, and the autosave overlay, replacing them with a savemanager.c/h + savefile.h/savedevice.h scaffold to build the new save system on top of. Also separates the concrete UI_ELEMENTS registration into uielementlist.h/.c so uielement.c only holds the generic per-element lifecycle logic.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Splits the platform-agnostic network core out of dusk into a top-level
dusknetwork module, mirroring the duskgl/dusksdl2 pattern. Adds a
DUSK_NETWORK cmake option (default ON) so builds can exclude all
networking code, including the per-platform implementations.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Renames savefile_t to saveslot_t and folds last session's standalone
settings.h/.c module back in as savemeta_t, so there's one save system
(SAVE.slots[] + SAVE.meta) instead of two parallel ones - while letting
each platform pick its own physical format for the two concepts:
- Linux now writes human-editable JSON (slot0.json, settings.json, ...)
via yyjson's mutable writer API, so players can hand-fix a bad setting.
- PSP folds meta into the same sceUtilitySavedata binary payload as its
one save slot (SAVE_SLOT_COUNT_MAX=1 there - a future save picker will
let players manage multiple named saves via the OS's own browser).
- GameCube consolidates the 3 per-slot memory card files and the separate
settings file into one combined card file.
Also fixes two bugs surfaced while building this: the CRC finalize step
seeked to a hardcoded offset (only safe for one section per file, breaks
once meta+slots share a buffer), and save.c's async/sync dispatch left an
unconditional fallback call that doesn't exist on PSP-only platforms.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Deadzone (and future prefs like locale) now live in their own
settingsfile_t/settings.c, loaded eagerly at boot and saved immediately
on Apply, instead of inside savefile_t - a setting shouldn't reset or
diverge just because the player is on a different save slot, and this
also fixes settings changes not actually reaching disk until the next
full game Save.
PSP settings use a new plain sceIo path rather than sceUtilitySavedata,
since that dialog would flash its native icon on every settings tweak.
GameCube reuses the save system's existing memory card mount rather than
mounting it twice (settingsInit() now runs after saveInit()).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Adds three new pieces of save-file state, all following the same shape:
the save file is the single source of truth, not a separate live runtime
copy that gets synced in/out.
- globalitemstore.h/.c: per-global-entity-ID "collected" flags
(savefile_t.globalItemCollected), so a global item entity's init
callback can check whether it was already picked up in a prior session
without needing to keep the entity itself alive to remember that.
- Gamepad deadzone: removed input_t.deadzone entirely. The setting UI and
every platform's actual deadzone-applying code (inputGetDeadzoneDolphin/
SDL2, previously hardcoded per-platform literals that the settings menu
didn't actually affect) now read savefile_t.deadzone directly via
saveGet(SAVE_ACTIVE_SLOT). Default lives in savefile.h
(SAVE_DEADZONE_DEFAULT), stamped onto every slot in saveInit().
- Story flags: STORY_FLAG_VALUES (a live, codegen-initialized array) is
replaced by savefile_t.storyFlags, read/written via the existing
storyFlagGet()/storyFlagSet() call sites (now macros/functions over the
active save file instead of a separate array). tools/story.py now
generates STORY_FLAG_DEFAULTS (const) instead; storyFlagInitDefaults()
stamps those onto a save the first time it's used (file->exists false),
called from rpgInit().
Added SAVE_ACTIVE_SLOT (0) to savefile.h as the one shared "which slot is
actually being played" constant, replacing three different local/implicit
0s (uigamemenu.c, rpg.c, and now the settings/input call sites).
Verified round-trip on Linux (all three together in one save/load cycle);
both Linux and PSP build clean.
- Fixed the actual reason saving never worked on any platform: saveWrite()
never stamped file->header/file->version before serializing, so every
written save file had a zeroed magic header and failed its own
validation on the next load. Confirmed via a manual write/load round
trip that this alone fully explains "saving doesn't work."
- Re-enabled saveInit()/saveDispose() in engine.c (previously commented out
under "Temporarily disable save code").
- Added SAVE.available + saveIsAvailable(), refreshed by every real
save/load/delete attempt. saveInit() no longer treats an unreachable
save medium as fatal to booting - it logs and continues, since a missing
memory card/stick shouldn't prevent playing.
- Hardened PSP's saveInitPSP() to actually detect a missing memory stick
(sceIoGetstat on ms0:/) instead of assuming success, and fixed
single-level sceIoMkdir to build the full PSP/SAVEDATA directory chain.
- Added busy-retry (CARD_ERROR_BUSY) and a not-mounted guard to Dolphin's
live savestreamdolphin.c path, extending the same handling already
backported into savedolphin.c.
- Added a "Save" entry to the game menu wired to saveWrite(0), showing a
clear message on success, on failure, and when saveIsAvailable() is false.
Backported from branch ac2 (commit 85b61097) - CARD_Mount was being called
without CARD_Init first, leaving per-channel control blocks and the DSP
unlock sequence unset. On real Dolphin/hardware this surfaced as a hard
MMIO crash instead of a clean CARD_ERROR_* failure.
Co-Authored-By: Dominic Masters <dominic@domsplace.com>
YourWishes