3bd7ee8eac
feat: phase 7 — proxy daemon with HTTP login and WebSocket bridge
...
CI / fmt (push) Successful in 38s
CI / check (push) Successful in 1m20s
CI / clippy (push) Successful in 1m19s
ericrfb-proxy axum server with three endpoints:
POST /api/login:
- Proxies credentials to OmniView auth.asp
- Extracts session cookie, fetches title_app.asp
- Returns JSON with applet_id, port, protocol_version, board_name
GET /api/ws?applet_id=...&port=...:
- WebSocket upgrade, connects to OmniView via e-RIC RFB
- Bidirectional pump: OmniView frames → RGBA blits over WS,
browser input events → key/mouse/hotkey to OmniView
- Binary protocol: TAG_BLIT(0x01), TAG_RESIZE(0x03) server→client;
TAG_KEY_PRESS(0x10), TAG_KEY_RELEASE(0x11), TAG_POINTER(0x12),
TAG_CTRL_ALT_DEL(0x13) client→server
Static file fallback via tower-http ServeDir.
Config via config.toml or BLEKIN_HOST env var.
Tested against real OmniView:
- Login endpoint returns valid APPLET_ID
- WebSocket upgrade succeeds (HTTP 101)
- Session connects and pumps frames
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com >
2026-05-06 15:11:50 +03:00
ab74f607e8
feat: phase 6 — keyboard/mouse input + fix Tight zlib init
...
CI / fmt (push) Successful in 30s
CI / check (push) Successful in 1m9s
CI / clippy (push) Successful in 1m9s
input.rs:
- write_key_press/release/tap: scancode | 0x80 = press, bare = release
- JavaScript KeyboardEvent.code → e-RIC scancode mapping (104pc layout)
- Hotkey sequence sender (raw hex byte strings from applet params)
- write_ctrl_alt_del() using HOTKEYCODE_0 "36 f0 37 f0 4e"
- PointerEvent writer already in msg.rs (8 bytes, absolute mode)
- 5 unit tests
Tight zlib fix:
- Changed Decompress::new(true) for zlib-wrapped format (matching
Java's Inflater() which expects zlib header 78 9C)
- Added output length verification after decompression
- Tested: Tight-encoded frames now decode correctly from real device
37 tests passing.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com >
2026-05-06 14:51:50 +03:00
c8f981f045
feat: phase 5 — Tight decoder with zlib streams and sub-palettes
...
CI / fmt (push) Successful in 36s
CI / check (push) Successful in 1m1s
CI / clippy (push) Successful in 1m4s
codec/tight.rs:
- Full Tight (encoding 7) decoder per ByteColorRFBRenderer.a() line 324
- Control byte: bottom 4 bits = zlib stream reset flags,
top 4 bits = subencoding (0-15)
- Subencoding 8: solid fill (1-byte color)
- Subencoding 15: palette-indexed fill (selector + index)
- Subencodings 4-7: filtered data with optional 2-color palette
- Subencodings 10-13: reduced bit-depth packed (1/2/4 bpp)
- Subencodings 0-3: raw 8bpp data
- Data >= 12 bytes uses zlib compression with varint length
- 4 persistent zlib streams with reset-on-flag logic
- All 4 hardcoded sub-palettes ported as RGB332 indices:
PALETTE_BW (2), PALETTE_GRAY4 (4), PALETTE_GRAY16 (16),
PALETTE_COLOR16 (16 EGA-like colors)
- Bit-depth unpackers: 1bpp, 2bpp, 4bpp (MSB-first)
- 5 unit tests
Updated examples to request [7, 5, 1, 0, -250] (Tight preferred).
Tested against real OmniView: correct rendering with Tight encoding.
32 tests passing.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com >
2026-05-06 14:44:06 +03:00
21ed797302
feat: phase 4 — Hextile decoder and recording example
...
CI / fmt (push) Successful in 28s
CI / check (push) Successful in 1m12s
CI / clippy (push) Successful in 1m12s
codec/hextile.rs:
- Full Hextile (encoding 5) decoder per ByteColorRFBRenderer.int()
- Handles: Raw tiles, BackgroundSpecified, ForegroundSpecified,
AnySubrects, SubrectsColoured flags
- Background/foreground colors persist across tiles
- 4 unit tests covering all subencoding paths
framebuffer.rs:
- Added fill_rect() for Hextile background/subrect fills
session.rs:
- Wired Hextile encoding 5 into the rect dispatch
examples/record.rs:
- 30-second (configurable) recording session
- Saves 1 PNG per second to out/ directory
- Requests encodings [5, 1, 0] (Hextile, CopyRect, Raw)
- Tested against real OmniView: 10 frames in 10s, no errors
27 tests passing.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com >
2026-05-06 14:35:20 +03:00
1164ffdd98
fix: send SetPixelFormat to request 8bpp RGB332
...
CI / fmt (push) Successful in 35s
CI / clippy (push) Successful in 57s
CI / check (push) Successful in 1m0s
The OmniView defaults to 16bpp when no SetPixelFormat is sent.
The Java applet sends SetPixelFormat (msg type 0) at
ByteColorRFBRenderer.new() line 76 to request 8bpp RGB332
(bpp=8, depth=8, true_color, red_max=7, green_max=7, blue_max=3,
shifts 0/3/6). Without this, pixel data arrives as 16bpp pairs
that produce green stripe artifacts when interpreted as 8bpp.
Verified: snapshot now matches the reference screenshot from
http://10.3.0.130/screenshot.jpg (dark screen with "Free" label).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com >
2026-05-06 14:31:37 +03:00
e9823aff03
feat: phase 3 — framebuffer, raw decoder, session pump, snapshot
...
CI / fmt (push) Successful in 37s
CI / check (push) Successful in 1m0s
CI / clippy (push) Successful in 1m4s
framebuffer.rs:
- Framebuffer struct (8bpp RGB332, row-major)
- apply_raw() blit, copy_rect() with overlap-safe logic
- to_rgba() via compile-time RGB332 LUT
session.rs:
- ActiveSession: connect + SetEncodings + initial FBUpdateRequest
- Full server message dispatch loop (all 15 message types)
- Raw (encoding 0) and CopyRect (encoding 1) decoders
- Ping response, bandwidth probe bookends, mode change resize
examples/snapshot.rs:
- Connects, waits for first FramebufferUpdate, saves PNG
- Tested against real OmniView at 10.3.0.130:443
- Successfully captured 640x480 frame
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com >
2026-05-06 14:23:43 +03:00
1bd43fc1f9
feat: phase 2 — handshake, message writers, and server message dispatch
...
CI / fmt (push) Successful in 30s
CI / check (push) Successful in 1m1s
CI / clippy (push) Successful in 1m4s
handshake.rs:
- Config, PixelFormat, ServerInit, Session types
- connect() walks all 11 handshake steps per aw.g() line 226
- Auth error mapping from aw.a(int) line 350 (7 error codes)
msg.rs — client-to-server writers:
- SetEncodings (type 2), FramebufferUpdateRequest (type 3)
- KeyEvent (type 4), PointerEvent (type 5, 8 bytes)
- PingResponse (type 149), BandwidthMarker (type 151)
msg.rs — server message dispatch:
- ServerMsg enum covering all 15 message types
- Readers for ping, bandwidth probe, ack, debug string,
RFB command, server cut text, server name update,
layout/locale, RDP event, FB update header
examples/handshake.rs: connects and prints session info.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com >
2026-05-06 14:11:31 +03:00
07db90094d
style: apply rustfmt to proto.rs
...
CI / fmt (push) Successful in 32s
CI / check (push) Successful in 55s
CI / clippy (push) Successful in 59s
Fixes CI fmt check failure — rustfmt wants multi-line assert_eq! for
long struct literals and the varint roundtrip assertion.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com >
2026-05-06 13:59:54 +03:00
3db2927add
feat: phase 0+1 — workspace scaffold and protocol primitives
...
Phase 0: Cargo workspace with ericrfb (lib) and ericrfb-proxy (bin)
crates, .envrc for RUST_LOG, workspace dependency pins.
Phase 1: ericrfb/src/proto.rs implements all wire primitives:
- read/write helpers matching h.java (u8, i8, u16-BE, i16-BE, i32-BE)
- varint reader/writer matching aw.int() (1-3 bytes, Tight lengths)
- modified-UTF-8 string reader matching h.byte()
- RectHeader { x: u16, y: u16, w: u16, h: u16, encoding: i32 }
- RGB332 → RGBA compile-time LUT (256 entries)
20 tests including proptest varint roundtrip over [0, 2^22).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com >
2026-05-06 13:44:24 +03:00
