Owner

tokuhirom

13 indexed repositories · View on GitHub

java-handbook

Java · 146 stars
jawiki-kana-kanji-dict
Generate SKK/MeCab dictionary from Wikipedia(Japanese edition)

Python · 65 stars
mutsu

Rust · 23 stars
OrePAN2

Perl · 22 stars
KeyScripten
Keyboard automation for Mac

Rust · 14 stars
mindcheese
Editable mindmap library written in TypeScript

TypeScript · 13 stars
blog4

Go · 5 stars
ashron

Go · 4 stars
assertj-postfix-plugin
https://plugins.jetbrains.com/plugin/8093?pr=idea

Java · 4 stars
bsky-to-markdown

TypeScript · 3 stars
mieta

Go · 1 star
sact

Go · 1 star
historylens

TypeScript · 1 star

Beginner friendly Everything

Open All

perf: `@a.elems` is 8,314 instructions and decodes the receiver's NaN-box tag 57 times Open

perf todo:deep

Difficulty 5/5 Over a week Newbie friendliness 48/100

tokuhirom/mutsu#8888 · 2 comments · Opened Sep 20, 2026
perf: a method call costs 22,528 instructions because it re-resolves the method by name every time Open

perf todo:deep

Difficulty 5/5 Over a week Newbie friendliness 38/100

tokuhirom/mutsu#8880 · 2 comments · Opened Sep 20, 2026
perf: a typed lexical's declared constraint is an env probe per store, and belongs on BindingDesc (ADR-0097) Open

perf todo:deep

Difficulty 5/5 Over a week Newbie friendliness 35/100

tokuhirom/mutsu#8877 · Opened Sep 20, 2026
A role body's assignment to an outer lexical is lost on the way out of composition Open

todo:deep working

Difficulty 5/5 Over a week Newbie friendliness 30/100

tokuhirom/mutsu#8862 · 1 comment · Opened Sep 20, 2026
Coercing into a class with a built-in parent (`class C is Str {}; C('x')`) has no instance representation to build Open

todo:deep

Difficulty 5/5 Over a week Newbie friendliness 35/100

tokuhirom/mutsu#8856 · Opened Sep 20, 2026
ecosystem sweep: sandboxed network tests diverge on bundled optional modules Open

todo:ticket

Difficulty 4/5 3-5 days Newbie friendliness 65/100

tokuhirom/mutsu#8844 · Opened Sep 20, 2026
Lower the native-typed nqp:: ops to real JIT IR on unboxed registers — the JIT is call-threading, not codegen Open

todo:deep

Difficulty 5/5 Over a week Newbie friendliness 35/100

tokuhirom/mutsu#8831 · 5 comments · Opened Sep 19, 2026
perf: runtime string-keyed name/type resolution is the JSON::Fast gap now (#8686 Phase 3 re-profile) Open

todo:perf working

Difficulty 5/5 Over a week Newbie friendliness 30/100

tokuhirom/mutsu#8830 · 10 comments · Opened Sep 19, 2026
Repeated debugger stack dumps stall after resume Open

todo:deep

Difficulty 4/5 3-5 days Newbie friendliness 35/100

tokuhirom/mutsu#8826 · Opened Sep 19, 2026
Nested async Supply plans lose callback topics Open

todo:deep

Difficulty 4/5 3-5 days Newbie friendliness 35/100

tokuhirom/mutsu#8825 · Opened Sep 19, 2026
Register IR + unboxed native locals + a JIT that emits real code (the last ~3x, needs an ADR — do NOT start before #8819/#8817/#8822) Open

todo:deep

Difficulty 5/5 Over a week Newbie friendliness 18/100

tokuhirom/mutsu#8823 · Opened Sep 19, 2026
36% of executed opcodes are argument packaging (`ContainerizePair` + `WrapVarRef`) that a statically-known callee does not need Open

todo:deep

Difficulty 5/5 Over a week Newbie friendliness 42/100

tokuhirom/mutsu#8822 · Opened Sep 19, 2026
GC safepoint / profiler poll / JIT-entry probe run per opcode (~5.6% of a tight loop) Open

todo:deep

Difficulty 5/5 Over a week Newbie friendliness 45/100

tokuhirom/mutsu#8821 · 3 comments · Opened Sep 19, 2026
mutsu retires ~860 machine instructions per bytecode opcode (rakudo parity needs ~30) Open

todo:deep

Difficulty 5/5 Over a week Newbie friendliness 25/100

tokuhirom/mutsu#8819 · Opened Sep 19, 2026
`cmp` between a Range and an `is Range` instance answers `More` instead of comparing structurally Open

Difficulty 4/5 3-5 days Newbie friendliness 45/100

tokuhirom/mutsu#8814 · Opened Sep 19, 2026
Apriori collection pipeline is correct but too slow on repeated Titanic queries Open

todo:perf

Difficulty 4/5 3-5 days Newbie friendliness 45/100

tokuhirom/mutsu#8777 · 3 comments · Opened Sep 19, 2026
perf: push_routine_with_location re-interns a qualified call's short name every call Open

todo:perf

Difficulty 4/5 3-5 days Newbie friendliness 55/100

tokuhirom/mutsu#8776 · Opened Sep 19, 2026
A closure defined inside `sub EXPORT`'s own `use`-containing block loses the nested import once that block's dynamic scope exits Open

todo:deep

Difficulty 5/5 Over a week Newbie friendliness 35/100

tokuhirom/mutsu#8751 · 3 comments · Opened Sep 18, 2026
profiler: on a multi-threaded run `sampled_ns` sums past `wall_ns` and the two cannot be compared Open

todo:deep

Difficulty 5/5 Over a week Newbie friendliness 35/100

tokuhirom/mutsu#8739 · Opened Sep 18, 2026
profiler: ADR-0106 Slice 6 — the optional output formats (moarperf export, HTML, compile-phase, Raku API) Open

todo:deep

Difficulty 5/5 Over a week Newbie friendliness 25/100

tokuhirom/mutsu#8738 · Opened Sep 18, 2026
profiler: `hits` is not a trip count when a loop body occupies one line Open

todo:deep

Difficulty 5/5 Over a week Newbie friendliness 35/100

tokuhirom/mutsu#8737 · Opened Sep 18, 2026
MoarVM bytecode identity lookup has no backend under mutsu Open

todo:deep

Difficulty 5/5 Over a week Newbie friendliness 35/100

tokuhirom/mutsu#8733 · Opened Sep 18, 2026
perf: parsing large MoarVM bytecode files is too slow Open

todo:perf

Difficulty 4/5 3-5 days Newbie friendliness 45/100

tokuhirom/mutsu#8732 · Opened Sep 18, 2026
Vendored JSON::Fast is far too slow parsing a real-world ~300KB document (scales to a timeout) Open

todo:perf

Difficulty 5/5 Over a week Newbie friendliness 35/100

tokuhirom/mutsu#8673 · 9 comments · Opened Sep 17, 2026
OrePAN2::Indexer::add_index fills disk on full reindex Open

Difficulty 2/5 1-3 hours Newbie friendliness 78/100

tokuhirom/OrePAN2#105 · Opened Sep 15, 2026
Re-whitelist `Digest ripemd.t` once it has real margin under the batteries gate's 120s budget Open

todo:perf

Difficulty 4/5 3-5 days Newbie friendliness 45/100

tokuhirom/mutsu#8471 · Opened Sep 15, 2026
deep: default worker pool deadlocks virtual scheduler continuations Open

todo:deep

Difficulty 5/5 Over a week Newbie friendliness 35/100

tokuhirom/mutsu#8380 · 10 comments · Opened Sep 14, 2026
perf: a closure call has no light path — it pays the general binder and an env frame where a named sub pays neither Open

todo:perf

Difficulty 5/5 Over a week Newbie friendliness 35/100

tokuhirom/mutsu#8335 · 2 comments · Opened Sep 13, 2026
perf: resolve_lexical_type_key scans every registered type for a prefix that is almost never there Open

todo:perf

Difficulty 4/5 3-5 days Newbie friendliness 48/100

tokuhirom/mutsu#8315 · 2 comments · Opened Sep 13, 2026
RakuAST: no call records whether it was written without parentheses, so `Call::Name::WithoutParentheses` is only ever produced for the four IO listops Open

todo:deep

Difficulty 3/5 1-2 days Newbie friendliness 78/100

tokuhirom/mutsu#8125 · 3 comments · Opened Sep 12, 2026
`format!("__mutsu_...::{name}")`-built env keys: 286 sites, the same hotspot found in 4 separate perf campaigns, and 96 of them have a memoized replacement that already exists Open

todo:deep

Difficulty 5/5 Over a week Newbie friendliness 35/100

tokuhirom/mutsu#8087 · 9 comments · Opened Sep 12, 2026
bench-threads is the only row slower than rakudo, and the cause is `@a[$i] = $v`: 14k instructions, 22 interns and 8 allocations per element store Open

todo:deep

Difficulty 5/5 Over a week Newbie friendliness 35/100

tokuhirom/mutsu#8069 · 13 comments · Opened Sep 12, 2026
RakuAST: no regex node tree — `QuotedRegex` and grammar `token`/`rule`/`regex` declarations cannot cross the boundary Open

todo:deep

Difficulty 5/5 Over a week Newbie friendliness 35/100

tokuhirom/mutsu#8033 · 146 comments · Opened Sep 12, 2026
ecosystem parity: 20 distributions time out under mutsu on files rakudo passes (4427 assertions) Open

todo:perf

Difficulty 4/5 3-5 days Newbie friendliness 45/100

tokuhirom/mutsu#7995 · 17 comments · Opened Sep 11, 2026
ecosystem parity: 99 distributions blocked by parse gaps that share only their error message Open

todo:deep

Difficulty 5/5 Over a week Newbie friendliness 35/100

tokuhirom/mutsu#7988 · 56 comments · Opened Sep 11, 2026
Parse-failure index: 56 `blocked_load` distributions, located to a file and line Open

todo:deep

Difficulty 5/5 Over a week Newbie friendliness 35/100

tokuhirom/mutsu#7954 · 19 comments · Opened Sep 11, 2026
Ecosystem distribution lock board — claim a dist here before working it Open

ecosystem:lock

Difficulty 1/5 Under an hour Newbie friendliness 10/100

tokuhirom/mutsu#7884 · 278 comments · Opened Sep 11, 2026
Move type/package names and internal markers out of the per-frame Env (ADR-0084) Open

tier:B todo:deep

Difficulty 5/5 Over a week Newbie friendliness 42/100

tokuhirom/mutsu#7817 · 1 comment · Opened Sep 9, 2026
perf: the last ~35 `Symbol::intern` calls per Test assertion — the binder's baked param syms, dispatch names, and one missing pre-interned key Open

todo:perf

Difficulty 5/5 Over a week Newbie friendliness 45/100

tokuhirom/mutsu#7766 · 14 comments · Opened Sep 9, 2026
positional-light call: 35% of bench-fib is per-call frame bookkeeping spread across a dozen sites, not one hotspot Open

tier:B todo:perf

Difficulty 5/5 Over a week Newbie friendliness 30/100

tokuhirom/mutsu#7738 · 2 comments · Opened Sep 9, 2026
Unexplained one-off SIGSEGV in `roast/S17-procasync/stress.t` (CI run 30590633128) Open

tier:S todo:deep

Difficulty 5/5 Over a week Newbie friendliness 35/100

tokuhirom/mutsu#7609 · 6 comments · Opened Sep 8, 2026
Tracking: the late-August call-path slowdown Open

icebox:record tier:icebox todo:perf

Difficulty 5/5 Over a week Newbie friendliness 15/100

tokuhirom/mutsu#7579 · 3 comments · Opened Sep 8, 2026
Parsing YAML with the bundled `YAMLish`: now at/near raku, deeper documents still ~2x Open

todo:perf

Difficulty 5/5 Over a week Newbie friendliness 20/100

tokuhirom/mutsu#7576 · 38 comments · Opened Sep 8, 2026
The interpreter function-call path is the one place mutsu is slower than raku Open

tier:B todo:perf

Difficulty 4/5 3-5 days Newbie friendliness 58/100

tokuhirom/mutsu#7573 · 4 comments · Opened Sep 8, 2026
The hash benchmarks have no single hot spot: the cost is spread across GC, allocation, NaN-box decode and key hashing Open

icebox:record tier:icebox todo:perf

Difficulty 5/5 Over a week Newbie friendliness 35/100

tokuhirom/mutsu#7570 · 1 comment · Opened Sep 8, 2026
bench-ctor: close the remaining gap vs rakudo (construction residual costs) Open

icebox:record tier:icebox todo:perf

Difficulty 5/5 Over a week Newbie friendliness 35/100

tokuhirom/mutsu#7568 · 3 comments · Opened Sep 8, 2026
`use Test` taxes every hot loop in the file that loads it Open

tier:B todo:perf

Difficulty 5/5 Over a week Newbie friendliness 35/100

tokuhirom/mutsu#7565 · 20 comments · Opened Sep 8, 2026
RakuAST remaining work Open

tier:B todo:deep

Difficulty 5/5 Over a week Newbie friendliness 20/100

tokuhirom/mutsu#7564 · 7 comments · Opened Sep 8, 2026
The per-call locals frame is a pooled `Vec`, not a register window Open

tier:B todo:perf

Difficulty 5/5 Over a week Newbie friendliness 25/100

tokuhirom/mutsu#7562 · 7 comments · Opened Sep 8, 2026
ADR-0019 G3: diffuse allocation/hashing cost in attribute-heavy construction (bench-ctor/bench-class) Open

tier:N todo:perf

Difficulty 5/5 Over a week Newbie friendliness 30/100

tokuhirom/mutsu#7561 · 3 comments · Opened Sep 8, 2026

Showing the newest 100

This page lists what was indexed most recently. Advanced filter has the whole inventory, narrowed by language, difficulty and how long a task takes.

Open advanced filter

tokuhirom

Showing the newest 100

Get new issues in your inbox