https://gitlab.synchro.net/main/sbbs/-/commit/5936da4d1c60ec677eff815a
Modified Files:
src/doors/syncconquer/door/door_io.c
Log Message:
syncalert: play FMV cutscenes by default (-NOMOVIES now opt-in)

The door forced -NOMOVIES into the engine argv, which suppressed Red
Alert's VQA intro/campaign cutscenes. That suppression was intentional
early on, but the movies actually stream fine over the terminal: the VQA
frames route through the game page-flip into door_present(), and the
movie loop keeps pumping the door's input/pacing (so they're skippable
and don't stall the connection).

Stop forcing the flag. -NOMOVIES is now injected only when the sysop
sets the SYNCALERT_NOMOVIES environment variable, so cutscenes play by
default. The door-arg neutering (sanitizing door-owned tokens out of the engine's view) is unchanged.

Cutscenes are silent for now -- the no-video build links the null VQA
audio backend; wiring VQA audio into termgfx is a separate follow-up.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
--- SBBSecho 3.37-Linux
* Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)

https://gitlab.synchro.net/main/sbbs/-/commit/0edf7a3435f121d8097ee0d6
Modified Files:
src/doors/syncconquer/door/door_io.c
Log Message:
syncalert: keep the PPM tier out of the F4 cycle when JXL is available

The uncompressed PPM image tier is only a useful fallback for a SyncTERM
build that lacks JPEG XL -- and a raw P6 carries no colorspace tag, so
SyncTERM decodes it as BT.709 rather than our sRGB and the palette looks
wrong. Real SyncTERM has JXL, so cycling into PPM with F4 was only ever a surprise with off colors.

Add it to the F4 tier list only when JXL is not available. sa_auto_tier()
still selects PPM as the automatic fallback for a no-JXL SyncTERM, so that niche keeps working.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
--- SBBSecho 3.37-Linux
* Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)

https://gitlab.synchro.net/main/sbbs/-/commit/bdd485515c8db099897fc76a
Modified Files:
src/doors/syncconquer/door/door_io.c
Log Message:
syncalert: merge REDALERT.INI instead of clobbering it

door_setup_engine_paths() rewrote the per-user REDALERT.INI from scratch on every launch with only a [Paths] section. That wiped the engine's own
persisted settings (music/sound volume, etc.) each session, and left
[Intro] PlayIntro absent -- so it defaulted true, Special.IsFromInstall
stayed true, and the door auto-started a Soviet campaign game on Normal
every session instead of showing the main menu.

Now merge with xpdev's INI reader/writer: read the existing file, (re)write only the runtime [Paths] UserPath/DataPath, and seed [Intro] PlayIntro=false ONLY when it is absent. So a fresh install boots to the main menu (with the Allies/Soviet and difficulty choosers), a sysop's or the engine's own later values survive, and iniWriteFile() rewinds+truncates so everything else in
the file is preserved. Existing broken per-user INIs self-heal on the next launch.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
--- SBBSecho 3.37-Linux
* Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)

https://gitlab.synchro.net/main/sbbs/-/commit/28800fb9114d146eaf591669
Modified Files:
src/doors/syncconquer/door/door_io.c
Log Message:
syncalert: fit-toggle fixes and a tighter stats overlay

Three fixes to the Ctrl-F Aspect<->Fill display toggle and the Ctrl-S
stats overlay:

- Clearing: door_fit_toggle() forced a repaint but not a screen clear,
so switching Fill (full-width) -> Aspect (narrower, centered) left the
old wider frame's pixels bleeding through the newly-exposed side
margins. Request a clear, as door_tier_cycle() already does when the
image re-centers.

- Fill centering: Fill set the emitted width to the whole canvas width
and centered on that, but door_emit_sixel then clamps sixel width to
DOOR_SCALE_MAX. On a canvas wider than the cap the image landed
flush-left with a lopsided right margin. Apply the same cap in
door_calc_rect (sixel only) before centering so it stays symmetric.

- Overlay: add a "fill" token (shown only in Fill mode, like "blob"),
drop the space before the rate unit, round bytes/frame to KB, and
remove the redundant trailing space, so the widest line (JXL with
both blob and fill) fits one 80-column row instead of truncating.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
--- SBBSecho 3.37-Linux
* Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)

https://gitlab.synchro.net/main/sbbs/-/commit/54de3085923bf9535feec2ad
Modified Files:
src/doors/syncconquer/door/door_io.c
Log Message:
syncconquer: write the INI the ENGINE looks for, per title (syncdawn)

SyncDawn replayed Tiberian Dawn's from-install path on every launch -- the
long intro, then Choose_Side, then straight into a mission -- instead of reaching the main menu.

door_setup_engine_paths() hardcoded "REDALERT.INI" as the file it writes into the per-user -home dir, for both titles. But each engine asks PathsClass::Init() for its OWN config filename -- redalert/startup.cpp passes "REDALERT.INI", tiberiandawn/startup.cpp passes "CONQUER.INI" -- and
PathsClass looks for exactly that name. Under syncdawn the door was therefore writing a file the engine never opens, so [Intro] PlayIntro=false never
reached it, Get_Bool("Intro", "PlayIntro", true) defaulted true, and Special.IsFromInstall fired every session (tiberiandawn/startup.cpp).

Add DOOR_ENGINE_INI to the per-title identity block (CONQUER.INI for TD, REDALERT.INI for RA) and write that.

It hid behind a door that otherwise worked. TD takes its mixes from door_engine_data_dir() rather than the INI's [Paths] DataPath, so the game
data loaded fine and nothing looked wrong -- while PlayIntro, and the savegame location, silently never reached the engine.

It also LOOKED platform-specific, and wasn't. With no INI beside argv[0] the engine's ini.Save() falls back to PathsClass::User_Path(): ~/.config/vanilla-conquer/vanillatd on POSIX, %APPDATA%\Vanilla-Conquer\vanillatd on Windows. That write persists for the *nix BBS user -- so from the second run on, Linux found PlayIntro=false and behaved, quietly papering over the bug -- but not for the Windows service account, where every run was thus a first run. That fallback is wrong for a door either way: it is ONE file shared by every caller, the very thing -home exists to prevent, and it is where TD's savegames had been going instead of
the player's own directory.

Verified on both hosts: data/user/<n>/tibdawn/CONQUER.INI is now created with our [Paths]/[Intro] plus the engine's own persisted [Mouse]/[Video] settings (the merge-don't-clobber path holding), and syncdawn boots to the main menu. The REDALERT.INI previously left in a tibdawn/ home is inert -- TD never read it -- and can be deleted.
--- SBBSecho 3.37-Linux
* Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)

https://gitlab.synchro.net/main/sbbs/-/commit/736866fe916b7f150db638f8
Modified Files:
src/doors/syncconquer/door/door_io.c
Log Message:
syncconquer: scale the image to the exact canvas on any terminal

The sixel/image size clamp only trusted an XTSMGRAPHICS-derived ceiling (g_canvas_is_gfx), falling back to a conservative 1000px "geometry-less
xterm" floor otherwise -- so Windows Terminal, which reports an exact pixel canvas via ESC[4;h;wt but doesn't answer XTSMGRAPHICS, had its image
clamped to 1000px and rendered small in a large window. Trust the exact
canvas (g_px_exact) whenever any terminal reports it -- WT, SyncTERM, any modern sixel terminal -- keeping the 1000px floor only for genuinely geometry-less terminals (xterm with window-ops off never sends ESC[4t, so g_px_exact stays false and it's still protected). Large pixel canvases
(e.g. 2700x1440) only ever come from non-SyncTERM terminals anyway, so the prior effectively-SyncTERM-only gating was exactly backwards.

Also add a geometry-provenance tag to the Ctrl-S stats bar (S/X/G =
SyncTERM detected / exact-px canvas / XTSMGRAPHICS gfx ceiling), which made this diagnosable at a glance.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
--- SBBSecho 3.37-Linux
* Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)

https://gitlab.synchro.net/main/sbbs/-/commit/d9d8d36916532df76f9616c2
Modified Files:
src/doors/syncconquer/door/door_io.c
Log Message:
syncconquer: Phase-2 dirty-rect JXL rendering (~2.5x bandwidth cut)

On the JXL tier, only re-send the parts of the map that actually changed between frames instead of re-encoding the whole 640x400 canvas every time.

Each changed frame is diffed against the previous one at 16x16-tile granularity; the dirty tiles are coalesced into a small set of tight
bounding boxes (4-connected component labeling plus a short-gap merge
pass), and each box is shipped as its own pixel-positioned JXL blob (DrawJXLBlob) that composites in place over the retained frame. A frame
that changes too much (>= DR_FALLBACK_PCT of tiles) or produces too many
boxes falls back to a normal full-frame encode, as does the first frame,
a palette change, or any fit/geometry change.

Live-measured on real play this drops a moving map from 50+ KB/frame to
under 20 KB/frame. The path is gated by a data/<door>/dirtyrect flag that
is re-checked about every 1.5s, so it can be toggled on and off mid-
session without relaunching the door.

Stats bar (Ctrl-S) tweaks that go with it: KB/f now reports the recent
windowed per-frame cost (from the ~1s bits/s + fps windows) instead of
the lifetime cumulative average, so toggling dirty-rect shows an
immediate drop; and the redundant cell-size field was removed and KB/f
moved to the front of the row so it no longer gets chopped.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
--- SBBSecho 3.37-Linux
* Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)

https://gitlab.synchro.net/main/sbbs/-/commit/44e5ea5042f6098243ea2a83
Modified Files:
src/doors/syncconquer/door/door_io.c
Log Message:
syncconquer: sixel-tier dirty-rect rendering (partial-sixel backend)

Extend the Phase-2 dirty-rect path to the sixel tier -- the non-JXL
fallback population: xterm / mlterm / wezterm / Windows Terminal, plus a
no-JXL SyncTERM. The framebuffer differ + coalescer are hoisted out of
the WITH_JXL guard into a shared dr_diff_coalesce() that both backends
call; only the per-tier emit differs.

Sixel can only be placed at a character cell (unlike a pixel-exact JXL
blob), so each coalesced box is snapped OUT to cell boundaries and its
top-left cell is addressed with the cursor, then re-emitted as its own
partial sixel that composites over the frame the client still holds. The
box height is rounded up to a 6-px sixel band to dodge SF #258 (SyncTERM mis-decodes a partial last band). On SyncTERM the palette registers
persist from the prior full frame so boxes ship palette-less; other sixel terminals reset registers per image, so each box re-sends the palette.

Same data/<door>/dirtyrect flag and ~1.5s live re-check as the JXL path,
and the same full-frame fallback on a large or over-fragmented change --
so it is never worse than a full frame. The saving scales with the static (unchanged) fraction: largest on a built-up map, near zero on a sparse animating scene. The ratio is lower than JXL's because sixel is
cell-snapped and compresses locally rather than globally, and it is
smallest on a huge canvas (Windows Terminal) where every box inflates and
the palette re-sends per box. Live-tested in Windows Terminal (sixel
tier): composites cleanly, frame size swings down with map development.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
--- SBBSecho 3.37-Linux
* Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)

https://gitlab.synchro.net/main/sbbs/-/commit/c59d086d6d5b7c5cd4270cce
Modified Files:
src/doors/syncconquer/door/door_io.c
Log Message:
syncconquer: omit the charset field from the stats bar in graphics tiers

The Ctrl-S debug bar's utf8/cp437 indicator only matters in the text
tiers (half/quadrant/sextant), where the glyph encoding is actually used
to render the frame. In the image tiers (sixel/JXL/PPM) it is just noise
on an already-long row, so show it only for the text tiers. The leading
space is folded into the field (as blob/fill already do) so suppressing
it leaves no gap.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
--- SBBSecho 3.37-Linux
* Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)

https://gitlab.synchro.net/main/sbbs/-/commit/b3ce4522b8b3e4474a4d840b
Modified Files:
src/doors/syncconquer/door/door_io.c
Log Message:
syncconquer: default dirty-rect on, sysop-toggle via <door>.ini [video]

Dirty-rect was gated behind a data/<door>/dirtyrect touch-file as A/B-test scaffolding -- off unless the file existed, re-checked live so it could be toggled mid-session. Now that both backends (JXL + sixel) are proven, make
it the default and retire the dev flag.

It is ON unless a sysop sets "[video] dirty_rect = false" in the door's
config .ini (syncalert.ini / syncdawn.ini beside the binary) -- the escape hatch for a terminal that mis-composites the partial updates, needing no rebuild. The .ini path + bool read is factored into a shared
door_config_bool() (also used by the existing [game] movies setting), and
the value is read once and cached since it is queried every present.

Validated the [video] dirty_rect read (absent file -> default on; = false

off; = true -> on; key absent -> default) against xpdev's iniReadBool.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
--- SBBSecho 3.37-Linux
* Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)

https://gitlab.synchro.net/main/sbbs/-/commit/bb8096e4d445077c979888b5
Modified Files:
src/doors/syncconquer/door/door_io.c
Log Message:
syncconquer: use termgfx's shared mouse handshake

Delegate the DEC 1003/1006/1016 enable/restore, the SGR-Pixels latch and
the DECRPM reply to termgfx/mouse instead of open-coding them here. The
emitted bytes and the coordinate mapping are unchanged -- door_input.c
and the door_io_* accessors keep their behaviour; only their backing
state moves into a termgfx_mouse_t. First consumer of the extraction.
--- SBBSecho 3.37-Linux
* Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)

https://gitlab.synchro.net/main/sbbs/-/commit/8fd7b6f2ca768062cab36f9e
Modified Files:
src/doors/syncconquer/door/door_io.c
Log Message:
syncconquer: subset the sixel palette on register-resetting terminals

Select SIXEL_PAL_USED for the full-frame and dirty-box sixel paths on non-SyncTERM clients; SyncTERM keeps FULL-on-change / NONE-reuse
unchanged. Matches the termgfx present path.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
--- SBBSecho 3.37-Linux
* Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)

https://gitlab.synchro.net/main/sbbs/-/commit/380a07db7630ebed03b09572
Modified Files:
src/doors/syncconquer/door/door_io.c
Log Message:
syncconquer: use the shared dirty-rect diff

The third and last copy of it. This door marked a 16px tile grid from two
index planes, labelled 4-connected components, merged boxes within a gap
and gave up past a coverage or fragmentation budget -- the same code as syncretro and termgfx_termio, agreeing with the latter on every constant
that mattered (128 components, gap 2, 45%) and differing only in what a
tile spans.

That is what ../../termgfx/dirty.h parameterises, so the diff half goes
and dr_diff_coalesce becomes a call. Its own post-merge cap of 16 boxes survives as cfg.max_boxes; nothing else about the behaviour changes,
including which frames fall back to a full one.

What stays here is what the shared module deliberately does not take: the mapping from tile boxes to something drawable, which in this door means
two backends -- JXL blobs placed pixel-exact, sixel snapped to character
cells.

Both binaries build; neither has been run. This door needs game data and a terminal to exercise, so syncalert and syncdawn want a play-test before
this is trusted.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
--- SBBSecho 3.37-Linux
* Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)

https://gitlab.synchro.net/main/sbbs/-/commit/f8589e01f493bd50dec0ce90
Modified Files:
src/doors/syncconquer/door/door_io.c
Log Message:
syncconquer: patch a palette change instead of repainting the frame

The same gate the sibling doors carried: any change to the palette forced
a whole frame, on both backends. The reasons were real -- the JXL backend
bakes the palette into each frame's RGB pixels, so a fade moves every
affected pixel's colour while leaving the index buffer the diff looks at untouched; and sixel colour registers are shared with what is already
drawn, so redefining one recolours the screen under the parts a patch
would not repaint.

Both are about which pixels a moved entry actually affects, and that is computable. Record WHICH entries moved and let the shared diff treat a
tile drawn with one of them as dirty even when its indices are identical (dirty.h's `stale`); the gate then has nothing left to protect. A change
that reaches no displayed pixel at all -- an engine rewrites its palette
freely and a scene draws with a fraction of it -- now de-dupes instead,
and is deferred rather than dropped: that path updates no state, so the
entries stay pending for the first frame that draws with them.

A SyncTERM box carries no palette by design, so the first box of such a
frame carries exactly the entries that moved. Every other terminal resets registers per image and its used-colour subset already covers it; the JXL backend carries pixels and needs nothing.

Measured on the sibling that could be measured -- Flight of the Amazon
Queen over Windows Terminal -- this gate was 267 frames carrying 77% of everything the door sent, after an earlier fix had already halved the
total. This door's content is different (a strategy map with a static
sidebar, not an adventure scene) so the share will differ, but the gate
is the same and so is the reasoning.

UNVERIFIED beyond compiling: syncalert and syncdawn need game data and a terminal, and neither was run. Worth a play-test on both, watching for a
region that keeps a stale colour -- that is what this class of change
breaks when it is wrong.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
--- SBBSecho 3.37-Linux
* Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)

https://gitlab.synchro.net/main/sbbs/-/commit/7ee248d3313b59a9f9970a60
Modified Files:
src/doors/syncconquer/door/door_io.c
Log Message:
syncconquer: map the mouse against where the sixel actually lands

Clicking the top of a sidebar button did nothing on a sixel terminal;
the click had to land noticeably lower to register. SyncTERM looked fine,
which pointed suspicion at mouse granularity -- pixel reports there, cell reports elsewhere -- and that was the wrong trail.

A sixel is placed by CURSOR ADDRESS, so its true top-left is the addressed cell's pixel: the centering offset rounded DOWN to the cell grid. The mouse mapper inverted the offset the centering math ASKED for instead, so every
click was biased by the remainder. On a 1330x1480 canvas with 20px cells a 1330x831 image is centred at y=314 but drawn at y=300, and those 14 pixels
are about 7 game pixels of upward bias -- comfortably the top of a button.
JXL is placed by pixel offset and needs no correction, which is the real
reason SyncTERM escaped: it draws on the JXL tier, and at 640x400 its sixel tier happens to round to zero anyway.

The mapper now derives its origin from the addressed cell on the sixel
tier, and keeps the pixel offset on JXL. Measured against the door's own geometry across three cell sizes -- SyncTERM 8x16, Windows Terminal 20x20
and foot's 13px cell -- the error is now zero on both axes in each.

The geometry line also reports where the mapper believes the image starts.
It already reported where the image was PUT; the two being different was
the entire bug, and nothing on that line could have shown it.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
--- SBBSecho 3.37-Linux
* Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)