All
Hardware
Maps
ML & AI
Music
Social
Tools
Writing
Apps & Games
Deciduous
Decision graph CLI for tracing how software decisions evolve
Deciduous Archaeology Demo
Demonstrations of decision archaeology on React and stacked git workflows
Phish Explorer
Jam analytics dashboard for Phish 3.0 era
Local LLM on MacBook
4-bit quantization, safetensors, and Bumblebee + EMLX for Apple Silicon
A Bot that posts like me
Porting the posting bot to Elixir using local LLM work
Receipt Printer Software Suite
A complete software suite for thermal receipt printers
Role Call
TV writer overlap explorer
Fill Your Sky
Interactive map of 418+ Bluesky communities with 545K+ people
Code Mirror
A live code mirror experiment
Pocket Pnin
A local LLM running on my iPhone, coming to the App Store for free
NYC Census Maps
Interactive census and PLUTO data visualization for New York City
MTA Bus Tracker
Real-time MTA bus and train tracking on an interactive map
Concert GIF Maker
Extract GIFs from concert videos with a retro Mac interface
Send a VERY direct message, to my receipt printer
A social project where friends send photos that print on my receipt printer
Archive TV
A real over-the-air TV channel from magnetic media archives
Losselot
Neural network loss function explorer
Todoinksies
A personal todo app
Ormery
An ORM written in Temper
Collage Maker
Upload photos and arrange them into grid collages
GenStage Tutorial 2025
A modern GenStage tutorial for the Elixir ecosystem
Temper Rust Bug
Found a bug in the Temper compiler and built a demo repo
300+ Years of Tree Law
A blog post that became its own LiveView application
HEEx in Other Languages
Experiments porting Phoenix HEEx templates to Rust, Lua, C#, Java, Python, and JS
Live Draft LSP
Live-stream blog drafts from Zed to Phoenix via a custom LSP
Bluesky Hoover Apps
Various apps that vacuum up and process the Bluesky firehose
Bobby Posts Bot
A bot that posts like me, in Python
Photo Booth Receipt Printer
A portable photo booth that prints on receipt paper
Nathan For Us
A Nathan For You social network with video search and GIF creation
Browser History Roast MCP
An MCP server that roasts you based on your browser history
GenStage Tutorial (Original)
The original GenStage tutorial
Bluesky Firehose Toys
Real-time firehose visualizations: emoji streams, jetstream comparisons, and more
+485-0 notactuallytreyanastasio/lumines-clone/main Mar 28 22:49
738e81b update claude with elixir stuff
+196-11 notactuallytreyanastasio/lumines-clone/main Mar 28 22:45
40b2e58 Merge pull request #2 from notactuallytreyanastasio/liveview-engine
Liveview engine
+196-11 notactuallytreyanastasio/lumines-clone/liveview-engine Mar 28 22:43
810967c docs: add zero-to-running README and update CLAUDE.md with architecture notes
README covers installing Erlang/Elixir/Postgres from scratch on
macOS, Ubuntu, and Windows, plus setup, controls, and project structure.
CLAUDE.md updated with architecture decisions, style rules, and notes
on what strategies worked during parallel development.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
eb63143 add decidous db
+1768-0 notactuallytreyanastasio/lumines-clone/main Mar 28 22:41
26a8986 Merge pull request #1 from notactuallytreyanastasio/liveview-engine
Lumines game engine + LiveView UI
8bb02d1 feat: implement Lumines engine in Phoenix LiveView with shared CSS
- 7 engine modules: Board (sparse map), Piece, Gravity, Scanner, Sweep, Scoring, Game
- LiveView at /game with keyboard input and dual timers
- Integrated shared CSS design system (lumines-* classes)
- 72 ExUnit tests, all passing
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
+442-44 notactuallytreyanastasio/lumines-clone/liveview-engine Mar 28 22:37
7a1b355 Address PR feedback + add property-based testing
- board.ex: replace nested ifs with multi-clause functions using guards
- board.ex: use `not is_nil()` instead of `!= nil`
- game.ex, piece.ex, sweep.ex, scoring.ex: convert to Ecto embedded_schema
- Add 27 property-based tests using stream_data covering:
- Board: put/get roundtrip, nil clearing, out-of-bounds, occupied? consistency
- Piece: 4-rotation identity, color preservation, 2x2 area, hard_drop validity
- Gravity: cell count preservation, color preservation, idempotency, no floating cells
- Scanner: only marks valid squares, preserves non-square cells
- Sweep: bounds, 16-tick wraparound
- Scoring: monotonic score, combo tracking, reset preserves score
- Game: valid init, game_over handling, input safety
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
+9346-78 notactuallytreyanastasio/blimp/main Mar 27 23:00
a496725 Fix type checker issues, remove --no-type-check escape hatch
types.zig: List → [Any], Map → %{Any=>Any}, ViewNode → Any
Previously these fell through to actor type, causing false errors
checker.zig: length() accepts list, map, tuple, string, nil — any collection
parser.zig: multi-line map literals — skipNewlines() after % and after {
allows %{\n key: val\n} syntax to parse correctly
case/situation as expression on RHS of assignment (x = case foo do...)
builtins.zig: set_at(list, index, value) — list element replacement
main.zig: remove --no-type-check flag entirely, type checker is mandatory
examples: add missing type annotations, use set_at instead of replace_at
683dfb7 examples: three nontrivial Hole operator demos
rate_limiter.blimp — adaptive token bucket with missing backoff strategy,
burst abuse detection, and priority request overdraft logic
content_pipeline.blimp — staged moderation pipeline: sanitize → classify →
score → publish/reject. Classification heuristics and scoring edge cases
are holes; the pipeline wiring and decision dispatch are implemented.
game_rules.blimp — tic-tac-toe engine with missing win detection, score
calculation, illegal move handling, and AI opponent move selection.
Good example of domain-specific reasoning Claude can fill in from context.
5238c23 docs: web framework handoff guide for next session
4a847ab Fix docs and hole prompt: lambdas/filter/reduce/map all exist and work
f0435a1 Typed lists and record types: [Item] and %{key: Type, key: Type}
types.zig: add record_type variant to the Type union
- RecordField struct: {name, ty}
- recordField(name) method for O(n) lookup by field name
- record_type and map are mutually compatible (subtyping both directions)
since records are just typed maps at runtime
- parseTypeName now handles %{key: Type, ...} format (record)
alongside the existing %{K => V} format (homogeneous map)
- parseRecordField helper splits "key: Type" into name + type
parser.zig: %{key: Type, ...} accepted as a type annotation
- detects homogeneous (%{K => V}) vs record (%{key: T}) by presence of =>
- handles multi-line record type annotations with newlines between fields
- trailing comma allowed
checker.zig: lookup() on record type returns specific field type
- if key is a string literal, resolves the declared field type
- lookup(profile: %{name: String}, "name") infers String return
- falls back to Any for non-literal keys or unknown fields
b4ae6b6 finish up some more typing concerns
fb5466b more stuff
b0bef0e web framework: TCP/HTTP builtins in Zig, server.blimp in Blimp
- tcp_listen/accept/read/write/close builtins (POSIX sockets, native only)
- to_html builtin: renders view_node trees to HTML strings
- Handles all 17 view primitives
- row gets data-row attr for flexbox direction
- button sends: attr becomes data-sends + onclick handler
- HTML-escapes string content
- web/server.blimp: HTTP/1.1 server written entirely in Blimp
- HTTP.Server actor: tcp_listen loop, accept, read, dispatch, respond
- Sessions actor: in-memory session store (map of id -> state)
- Pure render_app(data) -> view_node
- Pure handle_event(data, msg) -> data
- html_page wraps view HTML in a full document with styles + client JS
- Client JS: blimpSend() submits button clicks as form POSTs
- PRG pattern: POST handler -> redirect -> GET
20d8541 Merge blimp-core/view-dsl: View DSL actors render to DOM, interactive tutorial, mobile playground
Resolved conflict in builtins.zig by keeping both math/util builtins from main
and view primitive registrations from view-dsl. Took view-dsl WASM binaries.
a23d476 Hole: auto re-exec after patch + diff output
After patching the source file, execv() replaces the current process with
a fresh run of the patched file — zero user interaction required.
Diff output shows exactly what changed:
╔═══ Hole filled: file.blimp:42 ══════
║ - _ # Hole: directive text
║ + _ ->
║ + generated line 1
║ + generated line 2
╚═════════════════════════════════════
The full cycle: hit hole → call Claude → print diff → patch file → re-exec.
Second run uses the real implementation, no hole, no Claude call.
f9f865f web server working: escape sequences, auto-spawn actors, file evaluation
- main.zig: file mode now evaluates (was only printing AST); --ast flag for old behavior
- eval.zig: string literals process \r \n \t \\ escape sequences
- eval.zig: actor definitions auto-spawn a singleton + bind actor_ref in env
- value.zig: add ViewNode to typeName() exhaustive switch
- builtins.zig: ArrayListUnmanaged for Zig 0.15 compat in to_html
- .tool-versions: zig 0.15.2
- Makefile: build commands for macOS 26 (zig build runner broken)
- server.blimp: case (not situation) for complete dispatch
9c5c74d more chugging along with the hole operator
8acd934 Hole source patching: living codebase — holes fill themselves permanently
- lexer.zig: preserve last_comment across newline tokens so parser can read
directive text from the line AFTER the hole token is consumed
- parser.zig: save `_` token's line/col BEFORE advancing (correct loc for patch)
new parseHandlerBodyPublic() allows parsing handler-body statements in holes
- eval.zig: evalHole() now patches the source file after filling in
- finds the hole line by its stored loc
- detects original indentation from that line
- builds replacement: `_ ->\n <generated lines>`
- writes patched source back to disk
- subsequent runs hit the real implementation, not the Hole
- main.zig: set evaluator.source_path from resolved file path
--no-type-check flag to skip type checker for exploratory code
33ccd90 WIP
33b3690 Hole operator: situation _ # directive shells out to Claude
- lexer.zig: capture comment text in last_comment (not reset on newlines)
so parser can read directives like `_ # Hole: fill this in`
- parser.zig: hole branches now store directive text from the comment
body contains a hole node with directive instead of being empty
- eval.zig: evalHole() called when situation hits a directive-only _ branch
- builds prompt: directive + subject value + visible bindings + syntax notes
- shells out to `claude -p <prompt>` as subprocess
- strips markdown fences from response
- parses response with parseFilePublic, evaluates all nodes, returns last
- graceful nil fallback if claude not found or parse fails
- Works at REPL time and file evaluation time (both contexts)
Semantics: situation = incomplete code with Holes, case = complete dispatch
+3119-21 notactuallytreyanastasio/blimp/blimp-core/view-dsl Mar 27 15:52
7a33020 View DSL: actors render to DOM, interactive tutorial, mobile playground
- view_node JSON serialization in wasm_api.zig (blimp_has_view, blimp_get_view_ptr/len)
- JS DOM renderer maps view primitives to HTML elements
- Buttons send messages back to actors via <- syntax, triggering re-render
- Parser fix: skip newlines inside () argument lists for multi-line view calls
- dom_playground.html: mobile playground with live sim loop, 5-actor coffee shop demo
- tutorial.html: 14-section interactive tutorial from zero to coffee shop, all code blocks runnable
- docs/lang_design/view-engine.md: capabilities document
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
+182-24 notactuallytreyanastasio/blimp/main Mar 25 04:26
c557e98 LLVM O3 + native CPU, TCO, LTO, Elm-style compiler errors
- O3 + native CPU targeting: Blimp matches/beats C, Rust, Zig
- fib(40): 433ms (C: 444ms), btree(25): 162ms, KNN: 160ms
- Tail call optimization: recursive def calls annotated, tailcallelim pass
- LTO: --lto flag emits bitcode, links with -flto
- Rich error messages: title bars, source lines, region underlines, hints
- Type checker wired into blimp-compile
+89-9 notactuallytreyanastasio/blimp/main Mar 25 03:41
8cbb9b9 LLVM O2 optimization, type checker in compiler, actor-typed state fields
- LLVMRunPasses("default<O2>") eliminates all benchmark overhead
- fib(40): 434ms (was 579ms, now faster than C's 450ms)
- btree(25): 164ms (was 338ms, now tied with Rust)
- Type checker wired into blimp-compile (was skipped entirely)
- spawn/struct_lit infer actor types for message send validation
- Parser supports dotted type names (Marketplace.AccountHolder.Account)
- Wrong actor type in message send now caught at compile time
- try/catch removed from roadmap -- bubbles+orelse is the error model
+162-130 notactuallytreyanastasio/blimp/main Mar 25 03:15
4b6f560 more ownership and actors example
+170-138 notactuallytreyanastasio/blimp/main Mar 25 03:15
ea531f0 more ownership and actors example
+288-76 notactuallytreyanastasio/blimp/main Mar 25 03:03
dec79c4 readme refresh
+917-37 notactuallytreyanastasio/blimp/main Mar 25 02:43
40870d0 more marketplace configuration and changes with drawings to children
+488-0 notactuallytreyanastasio/blimp/main Mar 25 02:33
b2d542d add canvases of compiled programs as an option
+638-45 notactuallytreyanastasio/blimp/main Mar 25 02:22
ed203c1 Compiled benchmarks: fib, btree, knn vs C/Zig/Rust/Python/Ruby
- Forward-declaration pass for def functions (order-independent)
- Fibonacci fib(40): 579ms (1.28x C)
- Binary tree depth 25: 338ms (2x C)
- KNN 1000x1000 grid: 173ms (1.02x C)
- bench/run.sh runner with median-of-3 timing
- All benchmarks produce correct results across all 6 languages
+2395-122 notactuallytreyanastasio/blimp/main Mar 25 01:50
4578e7d Typed function parameters with runtime type checking
def add(a: Int, b: Int) -> Int do a + b end
fn(x: String) -> Int do length(x) end
Type annotations on params are now mandatory-compatible (optional
for backwards compat, but supported). Runtime type checking:
passing a String to an Int param gives TYPE MISMATCH.
Actor types work: def inc(c: Counter) -> Int accepts Counter refs.
Built-in types: Int, Float, String, Atom, Bool, Nil, List, Tuple,
Map, Function. List types: [Int], [String]. Actor types match by
template name.
Return type annotation: -> Type shown in function display.
Closure format shows types: fn(a: Int, b: Int) -> Int do ... end
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
120d063 String interpolation: "hello #{name}!"
Strings containing #{expr} are evaluated at runtime. The expression
inside #{...} is parsed and evaluated, then formatted and inserted.
Works with variables, arithmetic, function calls:
name = "world"
"hello #{name}!" # => "hello world!"
"math: #{1 + 2}" # => "math: 3"
"count: #{length([1,2])}" # => "count: 2"
Nested braces handled via depth counting. String values have
their quotes stripped in the output.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
c101ed4 Batch: rem/abs/nil?/elem/floor/ceil/round builtins, multi-line collections
7 new builtins (29 total):
rem(10, 3) => 1, abs(-5) => 5, nil?(nil) => true
elem({10,20,30}, 1) => 20 (works on tuples and lists)
floor(3.7) => 3, ceil(3.2) => 4, round(3.5) => 4
Multi-line collections: lists, tuples, and maps can span lines.
REPL depth counter now tracks brackets/parens/braces in addition
to do/end. Skips brackets inside strings and comments.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
4657560 Rebuild WASM: try/catch, string interp, 36 builtins, pipe into sends. 134KB
a00f270 Elm-quality error messages: Did you mean?, line numbers, better hints
Undefined variables now suggest closest match via edit distance:
"Did you mean \`x\`?" for \`xz\`
Unknown functions suggest closest builtin:
"Did you mean \`length\`?" for \`lenght\`
Error format now includes line number gutter (when available).
Source line highlighted in red. Improved type mismatch shows
compatible type combinations. No matching handler explains
what was attempted and suggests the handler definition.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
4f0c48a Builtin signatures in completions: reduce(list: List, init, fn: Function)
All 36 builtins now show full type signatures in autocomplete:
reduce(list: List, init, fn: Function)
map(list: List, fn: Function) -> List
split(str: String, sep: String) -> [String]
range(start: Int, end: Int) -> [Int]
etc.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
0b131f7 Debug: add console.log to autoComplete
d585702 Elm-inspired error messages: language refugee detection
Detects common patterns from other languages and explains the
Blimp way:
return 42 -> "Blimp doesn't use return. Last expression is the value."
class Foo -> "Use actors instead."
var x = 5 -> "Just assign directly: x = 5"
if true -> "Use situation for branching."
while -> "Use for to iterate."
function -> "Use def or fn."
1 === 2 -> "Use ==. Blimp does structural equality."
Also: "Did you mean?" for unknown functions now searches all 36
builtins with edit distance.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
5135afc Rebuild WASM with Elm-quality errors. 138KB
0655d28 Rebuild WASM: string interp, rem/abs/nil?/elem, multi-line collections, 29 builtins
31bd195 Fix diff viewer compactness and colors: nil pattern match bug, tighter spacing
- Fix line_class returning empty string when highlighted_at is nil (nil != false
in pattern match, fell through to catch-all)
- Tighten gutters, hunk spacing, line height for compact diff display
- Gate AutoScroll on follow mode to prevent scroll hijack
- Use custom CSS classes for diff line backgrounds instead of Tailwind utilities
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
be4d8db Add auto-complete to playground: shows as you type, Tab cycles, Enter accepts
47f7a97 Fix: atoms in branch body, random() builtin, marketplace demo
Branch body while loop was stopping at .atom tokens, preventing
branches like 1 -> :one from having a body. Removed .atom from
the outer while condition; atom-as-new-branch detection stays in
the newline check. Fixes for+situation, nested situations, and
any branch returning an atom value.
random(min, max) builtin using xorshift64 PRNG.
Marketplace simulation: Account, Network, Thief actors interacting
with randomness, bubbles, try/catch over 15 rounds.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
d80514d Pipe into message sends and closures
10 |> actor <- :set(_) substitutes the piped value for _ in args.
items |> double pipes into closures (not just builtins).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
d203de9 %Actor{fields} struct literal syntax for spawning actors
%Counter{count: 42} spawns a Counter with count overridden to 42.
%Greeter{} spawns with all defaults. Supports dot notation:
%Shop.Checkout{items: 0}. Sugar for spawn Actor(field: value).
Parses by dispatching % + UpperIdentifier to struct path,
% + { to map path. Evaluator looks up template and spawns
with field overrides like evalSpawnExpr.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
0294995 Add syntax highlighting to diff viewer via highlight.js
- Per-hunk highlight.js hook: highlights whole hunk for context, splits
back into per-line HTML preserving +/- prefixes
- Language detection from file extension (Language.detect/1)
- diff_content/1 helper builds spans without template whitespace
- GitHub-style syntax colors that work over green/red diff backgrounds
- 230 tests pass, zero warnings
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
c185968 Error showcase examples + top-level Makefile
12 error example files demonstrating all error message types:
typo suggestions, type mismatches, operator-specific advice,
language refugee detection, bubble/catch, no matching handler.
Makefile targets:
make build - build compiler + interpreter
make repl - launch terminal REPL
make compile - compile .blimp to native
make wasm - build WASM module
make web - start local playground server
make test - run all tests
make bench - run benchmarks
make errors - showcase error messages
make term-diff - launch Phoenix diff viewer
make deploy - deploy blog + REPL to production
make deploy-wasm - rebuild WASM and deploy
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
88cafc8 Elm-style error architecture: pre/post messages, region underlines, operator-specific errors
Error messages now show explanation BEFORE the code, then the code
with ^^^^ region underlines, then explanation AFTER the code, then
hints. Matches Elm's visual structure exactly.
Operator-specific type errors:
"hello" + 42 -> "I can't use + with String and Int.
The + operator works with Int and Float.
To join strings, use ++ or concat."
[1,2] + [3,4] -> "To join lists, use the ++ operator."
Value.typeName() returns human-readable type names for errors.
BlimpError now supports col_end for multi-character underlines.
post_message field for after-code explanations.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
d2e8d31 Auto-show completions as you type, Tab cycles, Enter accepts
d73d9dc try/catch for structured error handling
try do
risky_actor <- :dangerous
catch err do
"caught: #{err}"
end
Catches all errors: Bubble, TypeError, DivisionByZero, etc.
Error reason bound to catch variable as an atom (:bubble,
:type_error, :division_by_zero). Bubble reasons from
bubble :reason are passed through. Happy path returns try
body result. REPL depth counter handles catch as block closer.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
9e0c32e Debug autoComplete
717d8f1 History recall shows full multiline blocks, auto-sizes textarea
fc78832 Fix: rebuild WASM with completion exports, add error handling
5e0c04d Tab completion in browser REPL with type-aware auto-fill
Tab shows up to 3 completion candidates below the cursor.
Repeated Tab cycles through them. Enter accepts the highlighted
completion. Escape or any other key dismisses.
Completions include:
- Variables with their types
- Functions with full signatures: add(a: Int, b: Int) -> Int
- Type-aware argument filling: selects best-matching variable
from scope based on parameter types
- Builtins (36), keywords, actor templates
Click a completion to accept it. Shows "... N more" when
there are more than 3 candidates.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
d3771aa Fix completion visibility: explicit display:block, background, border
3879d3a Fix completion rendering: use .after(), better styling, scroll into view
80e7f35 Fix: dismissComp was clearing compItems before renderComp could use them
e37996f More debug logging for completion
a970ed5 When guards in situation branches, random(), comments in REPL, marketplace demo
situation n do
x when x > 0 -> :positive
x when x <= 0 -> :negative
_ -> :zero
end
Branch guard evaluated after pattern match, before body execution.
Fixed: atoms in branch body, comments skipped in piped REPL.
random(min, max) builtin with xorshift64 PRNG.
Marketplace demo: Account, Thief actors with deposits, withdrawals,
theft attempts, bubble on insufficient funds, orelse for recovery.
Shows: typed state, guards, situation, bubble/orelse, struct literals,
message passing, all working together.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
edb5f08 Type-aware completion engine for REPL and LSP
CompletionEngine queries evaluator state for intelligent completions:
- Variables in scope with their types
- Closures show full signatures: add(a: Int, b: Int) -> Int
- Type-aware argument filling: auto-fills best-matching variable
e.g., greet(c) when c is a Counter and greet expects Counter
- Builtins (36), keywords, actor templates
- Ranked by prefix match quality
WASM exports: blimp_complete(prefix_ptr, len) returns JSON array.
JS API: blimp.complete("prefix") returns completion candidates.
Each candidate has: label (display), insert (auto-fill), kind, score.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
a59f019 7 more builtins: not, size, empty?, flat, zip, uniq, sum (36 total)
not(true) => false
size([1,2]) => 2 (alias for length)
empty?([]) => true, empty?("") => true
flat([[1,2],[3,4]]) => [1,2,3,4]
zip([1,2], [:a,:b]) => [{1,:a},{2,:b}]
uniq([1,2,1,3]) => [1,2,3]
sum([10,20,30]) => 60
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
+2389-1 temperlang/vale-spec-experiment/add-deciduous-boilerplate Mar 24 20:44
7d46c3e add readme instructions for basic sync of deciduous dbs
Signed-off-by: Robert Grayson <bobbbygrayson+github@gmail.com>
2a047d8 initial deciduous bootstrap
Signed-off-by: Robert Grayson <bobbbygrayson+github@gmail.com>
+35703-7063 notactuallytreyanastasio/blimp/main Mar 24 14:22
926e41e start adding TUI stuff
f1b53f4 Perceus reference counting for BlimpVal
Every BlimpVal gets an rc field initialized to 1. RC operations:
- blimp_rc_inc: increment on share (list push, variable copy)
- blimp_rc_dec: decrement on scope exit, recursive free at zero
(strings freed, list items dec'd, map keys/vals dec'd, closure env dec'd)
- blimp_rc_reuse: returns same pointer if rc==1 for in-place mutation
Codegen inserts rc_dec after list element pushes (list_push already
does rc_inc, so the local ref is balanced). Foundation for become
in-place reuse and actor death O(1) cleanup.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
a9b5ca0 Sync tree-sitter grammar with Zig parser
Added: fn expressions, def statements, for/in loops, spread operators
(... and ..), self keyword, bubble statement, ++ concat operator,
spawn with actor_name. Updated conflicts and precedences. All new
constructs parse correctly. Known limitation: spawn with dot-notation
names (spawn Shop.Checkout) needs Zig parser, tree-sitter treats the
dot as dot_access.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
499085e Closures and higher-order functions
fn(x, y) do ... end creates closures that capture enclosing scope.
Closures are first-class values: store in variables, pass as arguments.
Named function calls check for closure bindings before builtins.
Higher-order builtins: map, filter, reduce, each. These take a list
and a closure, calling the closure for each element.
double = fn(x) do x * 2 end
map([1, 2, 3], double) # => [2, 4, 6]
filter([1,2,3,4], fn(x) do x > 2 end) # => [3, 4]
reduce([1,2,3], 0, fn(acc, x) do acc + x end) # => 6
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
b1258ab range, head, tail, sort builtins
range(1, 5) => [1, 2, 3, 4, 5] for integer sequences.
head/tail for list destructuring as functions.
sort for integer lists (insertion sort).
for i in range(1, 10) do i * i end now works naturally.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
040ac00 Codegen: lists, closures, for loops compile to native
Tagged value runtime (BlimpVal) with constructors, list operations,
and value extraction. Codegen emits blimp_val_int/blimp_val_list/
blimp_list_push for list literals, blimp_val_closure for fn expressions,
blimp_list_map for spread (...), and loop codegen with blimp_list_get/
blimp_val_to_int for unwrapping. New .tagged_val type tag for values
that live in the heap world. blimp_print_val for tagged output.
Full programs compile: actors + closures + lists + for loops + spread
all in native code. [21, 41, 61] from a program that spawns a Counter,
maps a list with a closure, and loops over the result.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
b28b738 Fix multiline situation branches: allow bare identifiers in body
The handler body parser rejected bare identifiers (like _ -> n)
as "probably a typo". This broke situation branches that return
variables. Removed the check. Now all situation patterns work:
situation n > 1 do
true -> n * 2
_ -> n
end
Works inside fn, def, standalone, with any expression in branches.
This was the root cause of all the multiline parsing failures.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
7d61d90 Add blog post #3: Actors Are Alive
New post covering diff viewer, multiplexer, actor runtime, evaluator,
Elm-style errors, LLVM compilation, REPL, and tooling feedback loop.
Split-pane layout for all code sections in posts 2 and 3. Screenshots
of diff viewer, multiplexer, REPL, line selection dialog. Image
lightbox modal on click. Index updated.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
0320969 Parser fixes, tree-sitter sync, benchmark suite
Fix parseStatementPublic to use parseTopLevel (handles def, situation,
case, for at top level). Remove def/situation/case/orelse from
parseError keyword reject list. Improve branch body termination to
detect new branch patterns at line start.
Tree-sitter grammar synced with all new constructs.
Benchmark suite: fib, actor spawn, ring in Blimp/Rust/Python/Elixir.
Known issue: multiline situation with 3+ branches inside def body
still fails in piped REPL. Needs deeper parser refactor for branch
boundary detection.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
530a306 Fix post 4 REPL embed: init widget, remove broken demo fn
d7c68bd fix canvas
f608d00 List concatenation with + operator
[1, 2] + [3, 4, 5] => [1, 2, 3, 4, 5]
Works alongside string + for concat and integer + for addition.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
b000853 Fix REPL depth counter for end), rebuild WASM, add playground
The REPL's do/end depth counter didn't recognize end) as closing
a block. Now allows ) and , after end keyword. Fixes inline
closures in function calls: map(list, fn(x) do x * 2 end)
Rebuild WASM with closures, spread operators, for/in, self.
Add playground.html with canvas + state panel + cheat sheet.
Rays draw on top of all shapes.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
f8fc2f9 Pattern matching: destructure maps, lists, tuples in situation/case
situation/case branches now do structural pattern matching:
- Variable binding: n -> n * 2
- Map destructuring: %{name: n, age: a} -> n
- List head/tail: [h | t] -> h
- Tuple destructuring: {a, b} -> a + b
- Literal matching: :ok, 42, "hello", true, nil
- Nested patterns compose recursively
- Holes always match (wildcard)
Patterns produce bindings that are scoped to the branch body.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
68a2c33 WIP
db61ab1 Fix 3+ branch situation parsing: don't skip newlines after body stmt
The branch body loop was skipping newlines after each statement,
preventing the branch boundary detection from seeing the newline
that separates branches. Now the newline check at the loop top
handles boundary detection, and looksLikeBranch scans for -> on
the next line.
Fixes: situation with 3+ branches, recursive fib definition.
fib(25) = 75025 in 187ms (tree-walking interpreter).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
e5b5078 Rebuild WASM with def, string ops, supervision. Add supervision example.
WASM now includes: def keyword, concat/split/contains/to_string/
to_int/slice/upcase/downcase builtins, bubble/orelse supervision,
CascadeBubble/SelfBubble restart strategies. 119KB module.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
7c5804d Embed WASM REPL in blog with canvas visualization
Inline REPL with canvas on the index page. "Try Blimp" button on all
three posts opens a modal REPL. Widget loads blimp.wasm (96KB),
blimp.js, and canvas.js from repl/ directory. Self-contained: one
script tag, no dependencies. Canvas shows generative hexagons for
actors, rays for message sends.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
c9ba44c Canvas: adaptive grid layout, squares for values, same-state same-art
Grid layout tries all column counts and picks the largest cell size.
Scales hexagons down when space is tight. Staggered odd rows for
hex packing. Raw values (non-actor variables) render as squares.
Same type + same state produces identical generative fill.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2517e17 Rebuild WASM: 123KB with def, string ops, supervision, ++, range, merge, type_of, print
3b9de49 WASM REPL: interpreter in the browser with generative canvas
Compile the tree-walking interpreter to wasm32-freestanding via
zig build wasm. 96KB blimp.wasm module. JS loader class with
eval/getState/reset API. Browser REPL with inline prompt, multiline
do/end buffering, command history, state sidebar showing variables
and actor instance state.
Canvas visualization: actors as hexagons with generative fills
seeded from state hash. Four pattern types (dots, rings, stripes,
blobs). Message sends animate as dashed rays from REPL blob to
target actor. State changes flash the border white. Breathing
green blob for the global runtime.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
e15ff3c Parser depth counter: use proper word boundaries for do/end detection
Word boundary check now uses isAlphanumeric instead of whitelist.
Fixes detection of 'do' after digits (e.g., 'situation n > 1 do').
Removed def/situation/case from parse error keyword reject list.
Improved branch body termination for new patterns at line start.
Known: 3+ branches in multiline situation still needs work.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
a987b2d term_diff: config-driven orchestrator, REPL explorer, cleanup
Orchestrator config for max concurrent slots and stall detection.
REPL LiveView consuming blimp --introspect JSON. Agent runner
session continuation. Commentary prompt refactor. Navigation and
diff parser cleanups. Test fixes for async coordination.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
9378be8 ++ operator for list and string concatenation
[1, 2] ++ [3, 4] => [1, 2, 3, 4]
"hello" ++ " world" => "hello world"
Separate from + (arithmetic/string). ++ is explicitly for
joining collections. Lexed as a two-character token.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
1262e6c Compile def to native LLVM functions with recursion
def name(params) do ... end compiles to a real LLVM function.
Forward-declared so recursive calls work. func_call finds it
via LLVMGetNamedFunction. situation branches compile to
conditional blocks.
fib(40) benchmark:
Blimp (compiled): 230ms
Rust: 190ms (1.2x faster)
Python: 8317ms (36x slower)
Blimp compiled is within 1.2x of Rust for pure recursion.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
45a5ada def keyword and string builtins
def name(params) do ... end creates named functions (sugar for
assigning a closure). String builtins: concat, split, contains,
to_string, to_int, slice, upcase, downcase. length() now works
on strings and maps too.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
365348e Loops, spread operators, and self keyword
Three ways to iterate:
...list, fn -> map (returns new list)
..list, fn -> each (side effects, returns :ok)
for x in list do ... end -> verbose loop (returns list)
Self keyword resolves to current actor's ref inside handlers:
on :double do
self <- :set(val * 2)
end
for/in keywords added. .. and ... are lexed as distinct tokens.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
7495ce3 Runtime message log for canvas rays, demo mode, mobile fixes
Evaluator tracks message sends in a log so canvas rays fire for
sends inside closures (map over actors). WASM state JSON includes
messages array. Canvas: ResizeObserver for mobile, re-layout on
resize, per-frame size check. Playground: Demo button types out
a full session at human speed. Root landing page links to playground.
Blog post #4 draft started.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
b31a8d1 Rebuild WASM with multiline fix
5a091ae Supervision: bubble propagation, orelse catches, CascadeBubble restarts siblings
bubble keyword signals actor failure. orelse catches Bubble errors
and executes fallback. Handler bubbles() annotation declares strategy:
- SelfBubble: restart just the failing actor
- CascadeBubble: restart all siblings under the supervisor
Dot notation defines the supervision tree: Shop.Checkout and
Shop.Inventory are siblings under Shop. When Checkout bubbles with
CascadeBubble, both siblings get restarted to their default state.
Also: spawn now handles dot-notation names (spawn Shop.Checkout).
Parent/child relationships derived from type name prefixes.
Registry.restartActor and restartChildren reset state to template defaults.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
b6384c8 Codegen: stub new features, preserve existing compilation
Add compileExpr cases for self_ref, for_expr, spread_map,
spread_each. Currently return UnsupportedNode (interpreter-only
features). Existing actor compilation (counter, bank, traffic_light)
unaffected. Full codegen for closures and list iteration needs
runtime support for heap-allocated values beyond i64.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
61421ba TUI REPL: live LLVM IR panel, Tab toggles STATE/IR view
Source accumulates across submissions, IR updates on each eval
with syntax highlighting. libvaxis TEA architecture.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
a081c6b merge, values, type_of, print builtins
merge(%{a: 1}, %{b: 2}) => %{a: 1, b: 2} with overwrite on conflict.
values(%{x: 10}) => [10] extracts map values as list.
type_of(42) => :integer for runtime type introspection.
print(val) outputs to stdout and returns val (identity).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
be54559 Async actor scheduler: run queue, round-robin dispatch, deadlock detection
blimp_send now enqueues messages and pumps a round-robin scheduler
instead of processing inline. FIFO run queue tracks runnable actors.
Self-send guard prevents deadlock when a handler sends to its own
actor. Deadlock detection aborts if scheduler makes no progress.
Actor status tracking (idle/runnable/running). Two new test programs:
cross_send.blimp and fairness.blimp.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
1b827fb Tagged value runtime: BlimpVal for heap-allocated data structures
Add BlimpVal tagged union to runtime.c alongside existing i64 path.
Supports: int, float, string, atom, bool, nil, list, map, actor_ref,
closure. Value constructors (blimp_val_int, blimp_val_list, etc.).
List operations: push, get, len, map, filter, reduce, each.
Tagged value printer (blimp_print_val).
This is the foundation for compiling closures, lists, maps, and
higher-order functions to native code. Existing i64-based actor
codegen is preserved and unaffected.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
f42e46b Rebuild WASM with 3+ branch fix, fib now works in browser REPL
Vibe Coding Rescue Missions
this mortal coil
are the buttons in the post editor just for show?
heya
it is katproto
Under-learning > Over-learning
Asking. Hacker News to Write For me
Anyone can post
Rat Kings Are Good Actually
Untitled
327 Years Of Tree Law In The Usa
A Genstage Tutorial And Reflection
Nathan Fielder
Whats My Schtick
Lets Write Letters
A Quick Typewriter Set Letter Project
I Pulled The Paul Tickets
A Fun Experiment
A Hilarious Shared Keyboard Key
A Me Museum
On Making A Link Blog
Building This Blog
Online (1)
Anonymous