core: Buffer<Item> as the heap storage primitive #54

New issue

Closed

opened 2026-10-05 21:07:14 +00:00 by orion · 11 comments

orion commented 2026-10-05 21:07:14 +00:00

Collaborator

Copy link

Context

Buffer<Item> joins Array in std/core: a growable heap sequence with an in-place promise and swapRemove. ECS columns (#92), the rebuilt collections (#55) and tensors stand on it. This issue is commit 2 of 2: the std/core type, its prelude export, its conformance programs and its docs. The compiler side is #299 (commit 1) and has to be merged and in a published seed first.

Source

• COLLECTIONS.md gap 6
• COLLECTIONS.md table
• ECS.md "What the design needs from Buffer<Item>" (promises 1 and 3; promise 2, the windows, is std/parallel's Window, already built over ArrayList and out of scope here)

Decisions (made in refinement; the design record stays the specification)

• Where: std/core/src/buffer.trb, re-exported from std/core/src/lib.trb and std/prelude like Array (public use Array, ... from "std/core"). std/core cannot depend on std/collections, so Buffer implements only what core owns: Iterate<Item>, Length, MutableIndex<Int, Item>, MutableSlice, plus Equals, Show, Hash as extends (items in brackets, ordered hash with combineHashes), written like the ones at the end of array.trb. Joining the List/Collection family and Bytes becoming Buffer<UInt8> is the follow-up #300, not this issue.
• Representation: public native type Buffer<Item> over the same runtime list as ArrayList (RuntimeKind.ListStorage; #299 registers it). No new C, no new torb_ symbol, nothing in the VM beyond the shared table.
• The in-place promise is the existing runtime behaviour made a documented contract: buffer[i] = x, set, append, insert, removeAt, swapRemove, replace, reverse, clear change the storage in place whenever it has one owner; they copy once (copy on write) only when the Buffer value has been copied (const b = a, passed or captured by value) and is then written. The doc comment states exactly this and nothing stronger.
• slice copies. ArrayList.slice shares storage, so a live slice silently turns the next write into a copy - the trap a frame budget cannot have. Buffer.slice is TorbScript (withCapacity, then get/append over the checked range) and the result owns its own storage. Hence no sliceBetween native for Buffer. Borrowed views are Window's job.
• swapRemove is TorbScript over the natives, O(1), no new native.

API (exact; doc comments with # Examples like Array)

Member	Kind	Behaviour
static fn withCapacity(capacity: Int): Self	native	empty; the capacity is a hint (as ArrayList)
static fn from(items: Iterate<Item>): Self	TorbScript	with From<Iterate<Item>>; so Buffer.of-style construction needs nothing else
fn length(): Int	native
fn get(index: Int): Item?	native	None out of range, negative included
fn at(index: Int): Item	TorbScript	panics with the ArrayList.at message (index <i> is out of bounds for a length of <n>)
var fn set(index: Int, value: Item)	native	in place when unique; panics out of range like ArrayList.set
var fn update(index: Int, change: (var element: Item) => Void)	TorbScript	change self[index], same line as ArrayList.update
var fn append(value: Item)	native
var fn insert(index: Int, value: Item)	native	shifts the tail
var fn removeAt(index: Int): Item?	native	shifts the tail; None out of range
var fn swapRemove(index: Int): Item?	TorbScript	answers the item at index, moves the last item into its slot, shrinks by one; order is not kept; None (and no change) out of range; on the last index it is removeAt(last). Never shifts.
fn slice(range: Bounds<Int>): Self	TorbScript	copy, see above; panics like ArrayList.slice for a bad range
var fn replace(range: Bounds<Int>, values: Self)	TorbScript over private native replaceBetween	as ArrayList.replace
var fn reverse(), var fn clear()	native
fn iterate(): Iterator<Item>, fn keys(): Iterate<Int>	TorbScript	BufferIterator holds the buffer by value (as ArrayIterator); keys is 0..length()

Not on Buffer in this issue: compact (moves to Collection, COLLECTIONS 3.2), list-literal adaptation (const b: Buffer<Int> = [1, 2]), windows (stays on ArrayList until #55), capacity()/reserve.

Acceptance criteria

• std/core/src/buffer.trb declares Buffer<Item> with exactly the API above; natives are only those of the "native" rows; exported from std/core and std/prelude; torb check on std clean.
• swapRemove removes in O(1): on [10, 20, 30, 40], swapRemove(1) answers Some(20) and leaves [10, 40, 30]; swapRemove(3) on that three-item buffer answers None and changes nothing; the last index just drops the last item.
• Value semantics: var a: Buffer<Int> = ...; const b = a; a[0] = 9 leaves b unchanged; the same through a field, a var parameter and a closure capture.
• In place when unique, shown by cost and not by a hook: a program writes buffer[i] for 200 000 indices of a 200 000-item Buffer<Int> and a loop of swapRemove/append over the same size, and finishes within the suite's per-program time; the PR states the measured time and the time of the same program with const keep = buffer held (every write copying) so the margin is on record. Plus a micro-benchmark under benchmarks/ (same shape, ArrayList beside it) with the numbers in the PR description.
• No leak: a Buffer<String> and a Buffer<Buffer<String>> through append, set, removeAt, swapRemove, replace, clear and going out of scope pass the zero-live-blocks gate.
• Both back ends: every conformance program below runs under the native build and under the VM (the conformance runner already does both) with identical .expected.
• Conformance programs in tests/conformance/ (each with .expected, .exit): buffer-basics (append/get/at/set/insert/removeAt/length/iterate/keys, out-of-range get/removeAt/swapRemove answer None), buffer-swap-remove, buffer-value-semantics, buffer-slice-copies (a slice taken, the original then written, neither sees the other), buffer-equality-show-hash, buffer-index-panics (.exit 101 style as array-index-bounds, with .stderr), buffer-in-place-scale, buffer-strings-leaks.
• Docs: the COLLECTIONS.md table row and gap 6 lose "(planned)" and say Built , the standard-library page for std/core lists Buffer; then build/release/torb.exe docs skill docs skills run and both committed (--check is a tier A gate).
• Tier A green. No change to ArrayList, Bytes, std/parallel or the compiler (that is #299).

Dependencies

• Depends on #299 (compiler natives for Buffer, commit 1 of 2), merged and its commit in a published seed (seeds.txt on the seeds release lists a seed at or after the #299 merge commit; the forge's Actions -> seed -> Run workflow publishes one at once). Without it CI cannot bootstrap std/core.
• Blocks #55, #92 and #300.

Notes

Not ready until the seed condition above holds. Commit 2 of the two-commit rule (CLAUDE.md "Seed and breaking changes"): after #299, no further compiler change is needed.

Moved from the planning lists of the docs; the design record stays the specification.

## Context `Buffer<Item>` joins `Array` in `std/core`: a growable heap sequence with an in-place promise and `swapRemove`. ECS columns (#92), the rebuilt collections (#55) and tensors stand on it. This issue is **commit 2 of 2**: the `std/core` type, its prelude export, its conformance programs and its docs. The compiler side is #299 (commit 1) and has to be merged **and in a published seed** first. ## Source - [COLLECTIONS.md gap 6](https://git.torb.dev/torbscript/language/src/commit/393ef2608e4a60b8ff76d8b6b408951b822d887d/docs/design/COLLECTIONS.md#L925) - [COLLECTIONS.md table](https://git.torb.dev/torbscript/language/src/commit/393ef2608e4a60b8ff76d8b6b408951b822d887d/docs/design/COLLECTIONS.md#L646) - [ECS.md "What the design needs from `Buffer<Item>`"](https://git.torb.dev/torbscript/language/src/branch/main/docs/design/ECS.md#what-the-design-needs-from-bufferitem) (promises 1 and 3; promise 2, the windows, is `std/parallel`'s `Window`, already built over `ArrayList` and out of scope here) ## Decisions (made in refinement; the design record stays the specification) - **Where:** `std/core/src/buffer.trb`, re-exported from `std/core/src/lib.trb` and `std/prelude` like `Array` (`public use Array, ... from "std/core"`). `std/core` cannot depend on `std/collections`, so `Buffer` implements only what core owns: `Iterate<Item>`, `Length`, `MutableIndex<Int, Item>`, `MutableSlice`, plus `Equals`, `Show`, `Hash` as `extend`s (items in brackets, ordered hash with `combineHashes`), written like the ones at the end of `array.trb`. Joining the `List`/`Collection` family and `Bytes` becoming `Buffer<UInt8>` is the follow-up #300, not this issue. - **Representation:** `public native type Buffer<Item>` over the **same runtime list** as `ArrayList` (`RuntimeKind.ListStorage`; #299 registers it). No new C, no new `torb_` symbol, nothing in the VM beyond the shared table. - **The in-place promise** is the existing runtime behaviour made a documented contract: `buffer[i] = x`, `set`, `append`, `insert`, `removeAt`, `swapRemove`, `replace`, `reverse`, `clear` change the storage in place whenever it has one owner; they copy once (copy on write) only when the `Buffer` value has been copied (`const b = a`, passed or captured by value) and is then written. The doc comment states exactly this and nothing stronger. - **`slice` copies.** `ArrayList.slice` shares storage, so a live slice silently turns the next write into a copy - the trap a frame budget cannot have. `Buffer.slice` is TorbScript (`withCapacity`, then `get`/`append` over the checked range) and the result owns its own storage. Hence no `sliceBetween` native for `Buffer`. Borrowed views are `Window`'s job. - **`swapRemove` is TorbScript over the natives, O(1), no new native.** ## API (exact; doc comments with `# Examples` like `Array`) | Member | Kind | Behaviour | |---|---|---| | `static fn withCapacity(capacity: Int): Self` | native | empty; the capacity is a hint (as `ArrayList`) | | `static fn from(items: Iterate<Item>): Self` | TorbScript | `with From<Iterate<Item>>`; so `Buffer.of`-style construction needs nothing else | | `fn length(): Int` | native | | | `fn get(index: Int): Item?` | native | `None` out of range, negative included | | `fn at(index: Int): Item` | TorbScript | panics with the `ArrayList.at` message (`index <i> is out of bounds for a length of <n>`) | | `var fn set(index: Int, value: Item)` | native | in place when unique; panics out of range like `ArrayList.set` | | `var fn update(index: Int, change: (var element: Item) => Void)` | TorbScript | `change self[index]`, same line as `ArrayList.update` | | `var fn append(value: Item)` | native | | | `var fn insert(index: Int, value: Item)` | native | shifts the tail | | `var fn removeAt(index: Int): Item?` | native | shifts the tail; `None` out of range | | **`var fn swapRemove(index: Int): Item?`** | TorbScript | answers the item at `index`, moves the **last** item into its slot, shrinks by one; order is not kept; `None` (and no change) out of range; on the last index it is `removeAt(last)`. Never shifts. | | `fn slice(range: Bounds<Int>): Self` | TorbScript | copy, see above; panics like `ArrayList.slice` for a bad range | | `var fn replace(range: Bounds<Int>, values: Self)` | TorbScript over private native `replaceBetween` | as `ArrayList.replace` | | `var fn reverse()`, `var fn clear()` | native | | | `fn iterate(): Iterator<Item>`, `fn keys(): Iterate<Int>` | TorbScript | `BufferIterator` holds the buffer by value (as `ArrayIterator`); `keys` is `0..length()` | Not on `Buffer` in this issue: `compact` (moves to `Collection`, COLLECTIONS 3.2), list-literal adaptation (`const b: Buffer<Int> = [1, 2]`), `windows` (stays on `ArrayList` until #55), `capacity()`/`reserve`. ## Acceptance criteria - [ ] `std/core/src/buffer.trb` declares `Buffer<Item>` with exactly the API above; natives are only those of the "native" rows; exported from `std/core` and `std/prelude`; `torb check` on `std` clean. - [ ] `swapRemove` removes in O(1): on `[10, 20, 30, 40]`, `swapRemove(1)` answers `Some(20)` and leaves `[10, 40, 30]`; `swapRemove(3)` on that three-item buffer answers `None` and changes nothing; the last index just drops the last item. - [ ] Value semantics: `var a: Buffer<Int> = ...; const b = a; a[0] = 9` leaves `b` unchanged; the same through a field, a `var` parameter and a closure capture. - [ ] In place when unique, shown by cost and not by a hook: a program writes `buffer[i]` for 200 000 indices of a 200 000-item `Buffer<Int>` and a loop of `swapRemove`/`append` over the same size, and finishes within the suite's per-program time; the PR states the measured time and the time of the same program with `const keep = buffer` held (every write copying) so the margin is on record. Plus a micro-benchmark under `benchmarks/` (same shape, `ArrayList` beside it) with the numbers in the PR description. - [ ] No leak: a `Buffer<String>` and a `Buffer<Buffer<String>>` through `append`, `set`, `removeAt`, `swapRemove`, `replace`, `clear` and going out of scope pass the zero-live-blocks gate. - [ ] Both back ends: every conformance program below runs under the native build and under the VM (the conformance runner already does both) with identical `.expected`. - [ ] Conformance programs in `tests/conformance/` (each with `.expected`, `.exit`): `buffer-basics` (append/get/at/set/insert/removeAt/length/iterate/keys, out-of-range `get`/`removeAt`/`swapRemove` answer `None`), `buffer-swap-remove`, `buffer-value-semantics`, `buffer-slice-copies` (a slice taken, the original then written, neither sees the other), `buffer-equality-show-hash`, `buffer-index-panics` (`.exit` 101 style as `array-index-bounds`, with `.stderr`), `buffer-in-place-scale`, `buffer-strings-leaks`. - [ ] Docs: the COLLECTIONS.md table row and gap 6 lose "(planned)" and say *Built <date>*, the standard-library page for `std/core` lists `Buffer`; then `build/release/torb.exe docs skill docs skills` run and both committed (`--check` is a tier A gate). - [ ] Tier A green. No change to `ArrayList`, `Bytes`, `std/parallel` or the compiler (that is #299). ## Dependencies - Depends on #299 (compiler natives for `Buffer`, commit 1 of 2), **merged and its commit in a published seed** (`seeds.txt` on the `seeds` release lists a seed at or after the #299 merge commit; the forge's Actions -> `seed` -> Run workflow publishes one at once). Without it CI cannot bootstrap `std/core`. - Blocks #55, #92 and #300. ## Notes Not `ready` until the seed condition above holds. Commit 2 of the two-commit rule (`CLAUDE.md` "Seed and breaking changes"): after #299, no further compiler change is needed. _Moved from the planning lists of the docs; the design record stays the specification._

orion added this to the Kernel of natives milestone 2026-10-05 21:07:14 +00:00

orion added the

loop

label 2026-10-05 21:07:14 +00:00

orion referenced this issue 2026-10-05 21:07:15 +00:00

collections: rebuild ArrayList, the hash tables and the tries on Buffer #55

orion referenced this issue 2026-10-05 21:07:18 +00:00

compute: std/tensor on the CPU #82

orion referenced this issue 2026-10-05 21:07:19 +00:00

ecs: columns on Buffer and parallel systems (ECS slices 6 and 9) #92

orion referenced this issue 2026-10-07 15:42:01 +00:00

parallel: Window and windows (CONCURRENCY slice F, gap 6) #4

orion referenced this issue 2026-10-07 15:42:01 +00:00

parallel: Window and windows (CONCURRENCY slice F, gap 6) #4

orion referenced this issue 2026-10-07 15:42:07 +00:00

parallel: Window and windows (CONCURRENCY slice F, gap 6) #4

orion referenced this issue 2026-10-09 00:21:21 +00:00

parallel: Window and windows (CONCURRENCY slice F, gap 6) #4

orion referenced this issue 2026-10-09 04:37:23 +00:00

feat(parallel): windows and Window #248

orion commented 2026-10-09 04:38:54 +00:00

Author

Collaborator

Copy link

[loop:torbscript] intake: in scope (author orion, label loop), not a duplicate, no collision with a design record (COLLECTIONS.md gap 6 is the source). Likely adds a native, so it follows the two-commit rule; refine decides. next: waiting (ready buffer full: #4, #49, #163, #164, #223 are ready) until 2026-10-09T10:00:00Z

[loop:torbscript] intake: in scope (author orion, label loop), not a duplicate, no collision with a design record (COLLECTIONS.md gap 6 is the source). Likely adds a native, so it follows the two-commit rule; refine decides. next: waiting (ready buffer full: #4, #49, #163, #164, #223 are ready) until 2026-10-09T10:00:00Z

orion referenced this issue 2026-10-09 05:16:18 +00:00

collections: rebuild ArrayList, the hash tables and the tries on Buffer #55

orion referenced this issue 2026-10-09 05:32:19 +00:00

compute: std/tensor on the CPU #82

orion referenced this issue 2026-10-09 05:36:18 +00:00

ecs: columns on Buffer and parallel systems (ECS slices 6 and 9) #92

orion referenced this issue 2026-10-09 06:35:38 +00:00

parallel: Window and windows (CONCURRENCY slice F, gap 6) #4

orion commented 2026-10-10 01:48:44 +00:00

Author

Collaborator

Copy link

[loop:torbscript] claim refine by session:loop-torbscript-20261010 until 2026-10-10T03:48Z

[loop:torbscript] claim refine by session:loop-torbscript-20261010 until 2026-10-10T03:48Z

orion referenced this issue 2026-10-10 01:51:58 +00:00

core: teach the compiler the Buffer natives (commit 1 of 2 for #54) #299

orion referenced this issue 2026-10-10 01:52:31 +00:00

collections: Buffer<Item> joins the List family and Bytes becomes Buffer<UInt8> #300

orion commented 2026-10-10 01:52:38 +00:00

Author

Collaborator

Copy link

[loop:torbscript] refine: split into #299 (compiler natives for Buffer, commit 1 of 2, ready) and this issue (the std/core type with exact API, swapRemove in TorbScript, slice copies, conformance programs; commit 2), plus follow-up #300 (List family and Bytes), next: waiting (needs a published seed containing the #299 merge; not ready until then; #299 goes first)

[loop:torbscript] refine: split into #299 (compiler natives for `Buffer`, commit 1 of 2, `ready`) and this issue (the `std/core` type with exact API, swapRemove in TorbScript, `slice` copies, conformance programs; commit 2), plus follow-up #300 (`List` family and `Bytes`), next: waiting (needs a published seed containing the #299 merge; not `ready` until then; #299 goes first)

orion referenced this issue from a commit 2026-10-10 01:57:10 +00:00

feat(natives): teach the compiler the Buffer natives

orion referenced this issue 2026-10-10 01:57:17 +00:00

feat(natives): teach the compiler the Buffer natives #301

orion referenced this issue from a commit 2026-10-10 08:03:45 +00:00

feat(natives): teach the compiler the Buffer natives

orion referenced this issue from a commit 2026-10-10 09:00:35 +00:00

feat(natives): teach the compiler the Buffer natives

orion referenced this issue 2026-10-10 09:38:22 +00:00

core: teach the compiler the Buffer natives (commit 1 of 2 for #54) #299

orion referenced this issue from a commit 2026-10-10 09:44:09 +00:00

feat(natives): teach the compiler the Buffer natives

argus referenced this issue from a commit 2026-10-10 10:37:59 +00:00

feat(natives): teach the compiler the Buffer natives

orion referenced this issue 2026-10-10 10:38:50 +00:00

core: teach the compiler the Buffer natives (commit 1 of 2 for #54) #299

orion added the

ready

label 2026-10-10 13:48:41 +00:00

orion commented 2026-10-10 13:48:41 +00:00

Author

Collaborator

Copy link

[loop:torbscript] refine: unblocked, the published seed torb-seed-1724b82c2d87 (seeds release) contains the #299 merge; ready, next: implement

[loop:torbscript] refine: unblocked, the published seed torb-seed-1724b82c2d87 (seeds release) contains the #299 merge; `ready`, next: implement

orion commented 2026-10-10 13:48:41 +00:00

Author

Collaborator

Copy link

[loop:torbscript] claim implement by session:loop-torbscript-20261010 until 2026-10-10T19:50Z

[loop:torbscript] claim implement by session:loop-torbscript-20261010 until 2026-10-10T19:50Z

orion commented 2026-10-12 06:53:16 +00:00

Author

Collaborator

Copy link

[loop:torbscript] release implement: the claim of session:loop-torbscript-20261010 expired without a result (no branch, no PR, no live worker).

[loop:torbscript] release implement: the claim of session:loop-torbscript-20261010 expired without a result (no branch, no PR, no live worker).

orion commented 2026-10-12 06:53:17 +00:00

Author

Collaborator

Copy link

[loop:torbscript] claim implement by session:loop-torbscript-20261012 until 2026-10-12T12:53Z

[loop:torbscript] claim implement by session:loop-torbscript-20261012 until 2026-10-12T12:53Z

orion commented 2026-10-12 06:58:36 +00:00

Author

Collaborator

Copy link

[loop:torbscript] release: the worker on anthropic:theorben died on its weekly limit before its first reply; respawning on the next profile.

[loop:torbscript] release: the worker on anthropic:theorben died on its weekly limit before its first reply; respawning on the next profile.

orion commented 2026-10-12 06:58:37 +00:00

Author

Collaborator

Copy link

[loop:torbscript] claim implement by session:loop-torbscript-20261012 until 2026-10-12T12:58Z

[loop:torbscript] claim implement by session:loop-torbscript-20261012 until 2026-10-12T12:58Z

orion referenced this issue from a pull request that will close it, 2026-10-12 07:00:48 +00:00

feat(core): Buffer<Item>, the heap storage primitive with swapRemove #310

orion commented 2026-10-12 08:04:31 +00:00

Author

Collaborator

Copy link

[loop:torbscript] implement: Buffer<Item> in std/core (natives plus TorbScript from/at/update/swapRemove/slice copy/replace/iterate/keys, Equals/Show/Hash), 10 conformance programs with allocation budgets, ArrayList benchmark twins, docs and skill regenerated. CI run 865 green on e65fc831 (tier A, gates, bootstrap, same-C; tier B skipped by path filter, the buffer programs pass native + VM locally against nightly-20261011). PR #310, next: review (merge gate argus). Tokens: ~230k.

[loop:torbscript] implement: `Buffer<Item>` in std/core (natives plus TorbScript `from`/`at`/`update`/`swapRemove`/`slice` copy/`replace`/`iterate`/`keys`, Equals/Show/Hash), 10 conformance programs with allocation budgets, ArrayList benchmark twins, docs and skill regenerated. CI run 865 green on e65fc831 (tier A, gates, bootstrap, same-C; tier B skipped by path filter, the buffer programs pass native + VM locally against nightly-20261011). PR https://git.torb.dev/torbscript/language/pulls/310, next: review (merge gate argus). Tokens: ~230k.

orion commented 2026-10-12 08:04:42 +00:00

Author

Collaborator

Copy link

[loop:torbscript] implement: a second worker was driving the same branch; both pushes landed on loop/54-buffer-std-core (mine: the format fixes, native markers in the std/core page, the swapRemove Panics section, pdisc stamps), CI green on e65fc831 (run 865), PR #310, next: review (merge gate argus). The benchmark margin numbers for the PR are still open (benchmarks are not a gate). Tokens: ~330k.

[loop:torbscript] implement: a second worker was driving the same branch; both pushes landed on loop/54-buffer-std-core (mine: the format fixes, native markers in the std/core page, the swapRemove Panics section, pdisc stamps), CI green on e65fc831 (run 865), PR https://git.torb.dev/torbscript/language/pulls/310, next: review (merge gate argus). The benchmark margin numbers for the PR are still open (benchmarks are not a gate). Tokens: ~330k.

argus closed this issue 2026-10-12 08:04:53 +00:00

orion referenced this issue 2026-10-13 22:53:26 +00:00

text: normalization data and natives, teach the compiler (commit 1 of 2 for NFC and NFKC) #362

orion referenced this issue 2026-10-13 22:53:26 +00:00

dns: IDNA, joining and bidi data and natives, teach the compiler (commit 1 of 2 for UTS #46) #364

orion referenced this issue 2026-10-17 22:03:44 +00:00

collections: Buffer<Item> joins the List family and Bytes becomes Buffer<UInt8> #300

orion referenced this issue 2026-10-23 04:17:49 +00:00

collections: design record for ArrayList, the hash tables and the tries on Buffer (step 1 of #55) #413

orion referenced this issue 2026-10-23 04:17:56 +00:00

collections: rebuild ArrayList, the hash tables and the tries on Buffer #55

orion referenced this issue 2026-10-23 16:10:13 +00:00

compute: design record for std/tensor on the CPU (step 1 of #82) #438

orion referenced this issue 2026-10-23 16:10:42 +00:00

compute: std/tensor on the CPU #82

orion referenced this issue 2026-10-23 16:10:42 +00:00

compute: design record for std/tensor on the CPU (step 1 of #82) #438

orion referenced this issue 2026-10-30 09:05:33 +00:00

compute: package std/tensor, Tensor<Scalar> over a Buffer (step 2 of #82) #691

Sign in to join this conversation.

No Branch/Tag specified

Branches Tags

main

loop/705-default-shadow

loop/691-tensor-core

loop/705-fix-candidate

orion/456-tierb-probe

loop/164-measure2

loop/207-measure

loop/164-prototype

probe/207-count

loop/205-measure

loop/140-hide-staged

nightly-20261103

nightly-20261031

nightly-20261029

nightly-20261027

nightly-20261026

nightly-20261025

nightly-20261024

nightly-20261023

nightly-20261021

nightly-20261020

nightly-20261015

nightly-20261014

nightly-20261013

nightly-20261011

nightly-20261010

nightly-20261009

nightly-20261008

seeds

Labels

Clear labels   

loop

Backlog item of the project loop (Orion), see CONTRIBUTING

  

ready

Refined to the definition of ready

No labels

loop

ready

Milestone

Clear milestone

No items

No milestone

Kernel of natives

Projects

Clear projects

No items

No project

Assignees

Clear assignees

No assignees

1 participant

Notifications

Subscribe

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference

torbscript/language#54

No description provided.