Merge plan after v0.3.13: PRs tested by us, not yet in the release

[itaskaev2]

@valentderah, the merge plan for the release after v0.3.13. It replaces #836. Updated 2026-11-10, against dev be7c270; the check of that day is in the comment below.

In dev since v0.3.13

• Performance: chunked replay (Record large submissions in chunks on several threads #628 + Run DLSS on the chunked replay #794 by ipavlin98, merged as 5e56bfc; our carry Chunked replay (#628) and DLSS on it (#794), carried onto dev: the sky and paraglider dips at 4K #941 closed), Let the game run one frame further ahead of the GPU (#676 by ProSep, on dev) #917 sync_ahead (Let the game run one frame further ahead of the GPU (TOTK_SYNC_AHEAD, off by default) #676 by ProSep), Hold the game's dynamic resolution at 100 % and let the scene grow past 4096 wide (#573 by ProSep, on dev) #916 dynamic resolution (Hold the game's dynamic resolution at 100 % and let the scene grow past 4096 wide #573 by ProSep; dynamic_res now defaults to 1), Keep the targets a frame renders in the table that tells per-frame passes from one-shot ones #1040 the per-frame pass table (ipavlin98).
• Fixes: Take the feedback-layout images back to GENERAL before the contact shadows' march #1060 contact shadows' image layouts, Let a validation run with DLSS finish: a valid OpSelect in the motion fragment shader, and set 1 bound first for the motion draws #1061 validation runs with DLSS, Stop the installer space check at an unavailable drive root #1006 the installer at a missing drive (Luxury213), Fix Amiibo file names with non-ASCII characters displaying as garbled text in F1 menu #978 amiibo file names in any script, Linux AppImage: ignore user site-packages so installer.gui imports #990 the AppImage and user site-packages, Make the PC prompts for every game resolution the setup lays the UI out for #979 PC prompts for every resolution.
• Closed: Give the exclusive monitor a store generation, so a store that puts the value back fails STXR (#674 by zzs181, on dev) #868 / Give the exclusive monitor a store generation, so a store that puts the value back fails STXR #674 (no ABA symptom in TotK, the counter's cost), Cheaper chunked replay, uniform snapshot and DLSS motion tracking behind switches (96 to 138 fps on an i7-13700KF) #1053 (to be split).

Thank you for all of them.

Tested by us, waiting

Ready

PR	What	Tested	State
#913 = #638 by STS78721	NVIDIA Reflex (reflex = off / on / boost), with our two fixes: the recorder carries the present ID, and vkLatencySleepNV takes the present lock	RTX 4080 SUPER: on and boost at 60 fps, 0 validation errors (20 threading errors before the fix). Latency not measured	mergeable
#1066	docs/settings.md rows for deswizzle, kspin_us, pcore, rec_async, rec_spin_us	settings_doc_check: 5 problems on dev, 0 with it	mergeable, docs only
#864 = #675 by zzs181	tests/run_all.py, one entry point for the tests	every check without game files passes	mergeable
#865 = #656 by jkoehler11	the runtime compiles on both hosts in CI; the AppImage builds in a pinned container	both workflows pass on the PR (Windows 3 min, Linux 9 min)	mergeable
#995 by deadlynch	replay_async = 1 in the Steam Deck profile	the profile turns it on, 40 fps held; measured by the author on a Deck-like 680M, not on a real Deck	mergeable

Synced with dev on 2026-11-10, mergeable

PR	What	Tested
#915 = #624 by ipavlin98	antialiasing = smaa on the present	Iris Xe: works, costs 14 %; 0 validation errors
#914 = #228 by Lucas-blip-png	AMD FSR 3.1 as the other DLSS / DLAA engine	the device loss on NVIDIA is fixed (eb85d98, rec_drain() before the bridge's dispatch): 0 losses in ~27k FSR frames

Need a sync with dev from their authors

PR	What	Tested
#996 by Wesley-CR	lift again only the chunks whose inputs changed	61 s to 4 s on a second lift; output byte-equal to --no-cache; one docs conflict
#972 by m3rcury02	update to a new release from the settings menu	13 update tests pass; install-and-restart not run; one cfg.c conflict

Our own features, waiting for your review

PR	What	State
#790	The Second Quest on 1.4.3: MSFix as host hooks, IPS data writes, mods without their code module	runs on six saves and in daily play with TSQ + a 4K pack; the plugin part waits for the step 2 design, #845; synced with dev, mergeable
#732	Achievements tab	your eight points are done; mergeable
#497 (draft)	Setup and in-game menu in the game's languages	mergeable
#367 (main)	A working-loop skill for coding agents	unchanged

Tested, optional or waiting for their authors

PR	Result
#1054 usnap_fast	no gain on a Ryzen 7 5800X3D (median 125.9 fps off, 120.1 on); off by default
#975 view formats	never triggers in TotK; a compatibility-class check would help before other games rely on it
#1002 1.1.0 patches	408 Unicorn cases pass; no 1.1.0 dump here
#1024 motion sensor recovery (draft)	the virtual-pad test passes; needs a physical pad
#1050 launcher (base main)	builds and works with dev's installer; downloads are checked by size only, not SHA-256 (review comment there)

Not ready

The rest of #836's "not ready" list stays as it was: #487, #690, #647, #741, #669, #657, #748.

Suggested order

1. NVIDIA Reflex on the recorder thread: carry the present ID and lock the latency sleep (#638 by STS78721, on dev) #913, Describe deswizzle, kspin_us, pcore, rec_async and rec_spin_us in docs/settings.md #1066: mergeable now.
2. Add tests/run_all.py: one entry point for the tests (#675 by zzs181, on dev) #864, Compile the runtime on both hosts in CI, and build the AppImage in a pinned container (#656 by jkoehler11, on dev) #865: tests and CI.
3. SMAA on the present (#624 by ipavlin98, on dev) #915, AMD FSR 3.1 as the other DLSS / DLAA engine, on dev (#228 by Lucas-blip-png) #914, Run The Second Quest on 1.4.3: MSFix as host hooks, IPS data writes, mods without their code module (from ericdallo's fork) #790: synced, mergeable.
4. Achievements tab: statistics from the save and achievements on them #732, Setup and in-game menu in the game's languages, with a Languages table in the README #497: after your review.
5. Steam Deck profile: run the NVN replay on its own thread, and a benchmark script for Deck owners #995; Lift again only the chunks whose inputs changed #996, Update to a new release from the settings menu #972 after their authors' syncs.

[itaskaev2]

@valentderah, a check of dev and of the PRs opened since 2026-11-08. Machine: Windows 11, Ryzen 7 5800X3D, RTX 4080 SUPER (driver 616.92), game 1.4.3, 3200x1800, DLSS Quality, no mods. Each PR was merged onto dev 1503b17 and tested there; the results are in each PR.

dev 1503b17

• --selftest, --probe-kernel, --probe-vkexec, --probe-nvn-tex, --probe-shader, --probe-cfg, --probe-nvn: pass. settings_doc_check: the same 5 undocumented names as before (DESWIZZLE, KSPIN_US, PCORE, REC_ASYNC, REC_SPIN_US).
• Six saves, 40 s each with the camera turning, 60 fps limit: 59.0 to 60.0 fps mean, no faults. The sky dive holds 59.2 to 59.9, 59.9 with replay_workers = 4, replay_chunks = 16.
• Two old problems, found with validation on a config with contact_shadows and DLSS on. Both are also on dev from 2026-11-07, and I opened fixes for both:
  • Take the feedback-layout images back to GENERAL before the contact shadows' march #1060: contact shadows break the image layout rules on every frame (40 validation errors per run, 0 with contact_shadows = 0). This is from our own contact shadows code.
  • Let a validation run with DLSS finish: a valid OpSelect in the motion fragment shader, and set 1 bound first for the motion draws #1061: with DLSS on, a validation run crashes inside the layer, from an invalid OpSelect in the per-object motion fragment shader and the layer's push-set problem that set1_first already works around for the replay's draws. With both PRs a run with DLSS and contact shadows has 0 validation errors.

PRs since 2026-11-08

PR	Result	Recommendation
#917 sync_ahead (ours, = #676)	+15 to +21 % fps without a limit (118 to 135-144); no new validation messages; confirmed on a 4070 Ti and on Fedora	merge
#1006 installer at a missing drive	the hang reproduced on dev (killed after 5 s), the PR returns an error in 0.46 s; 55 installer tests pass	merge
#996 lift cache	61 s to 4 s on a second lift; output byte-equal to --no-cache; 9 tests pass. One docs conflict with the cheat VM paragraph	merge after the conflict
#1040 pipeline target table	no regression; the waits it fixes do not reproduce on NVIDIA here (0 one-shot waits on dev either)	merge if the author's numbers are enough
#995 Deck replay_async	the profile turns it on, 40 fps held; not measured on a real Deck	wait for Deck numbers
#972 in-game update	13 update tests pass, menu rows show; install-and-restart not run. One cfg.c conflict	review
#1054 usnap_fast	no gain on this CPU (median 125.9 off, 120.1 on); off by default, no faults	optional
#975 view formats	never triggers in TotK (also with BC3 ASTC); a compatibility-class check is worth adding before other games rely on it	not needed for TotK
#1002 1.1.0 patches	408 Unicorn cases pass; no 1.1.0 dump here	review
#1024 motion sensor recovery (draft)	the SDL virtual-pad test passes; no physical pad here	wait for hardware tests
#1050 launcher (base main)	builds, works with dev's installer; downloads are checked by size only, not by SHA-256, and an older download can replace a newer install	changes first
#1053 combined perf (draft)	conflicts in 15 files; its parts tested as #1054 and #917	wait for the split
#978, #990	already in dev (7e440be, 35968c1), checked there	can be closed

Still open from the plan above: #913, #916, #915, #865, #864, #790, #732, #497, #367. #914 (FSR 3.1) is fixed (eb85d98) and ready for review.

[itaskaev2]

Closing: everything from this plan that was merged is in v0.4.0. The rest moved to #1072.