Repository navigation
docs: Zig ADR, restored dirty-NIF ADR, ACCOMPLISHMENTS.md, decision-log guidance - #228
Merged
Merged
Conversation
…og guidance - decisions/2026-05-09-zig-for-android-nif-and-native-build.md: retroactive record of why core Android native code is Zig (VM-crash risk, types, defer, comptime), what the build actually does (build.zig orchestration; Apple clang/swiftc and NDK clang remain), and the consequences: partial defer adoption, ReleaseSmall in every build, two live Zig pins, the d5038c5 bug class, agents defaulting to C. - decisions/2026-04-28-dirty-nifs.md: restores docs/decisions/0001-dirty-nifs.md, removed in 88532d9 while ios/mob_nif.m still cited it. Status marks the regular-scheduler list superseded by 2026-09-05-input-nifs-are-dirty-io.md. The iOS comment now cites the new path. - ACCOMPLISHMENTS.md: verified hard problems agents solved here, each with commit/PR/ADR evidence, for agents to read before declaring a task impossible. - AGENTS.md: 'Before you say it can't be done' section; pre-empt rule 20 (core android/jni is Zig); Decision log paragraph on recording default-contrary and retroactive decisions; Where-to-look rows; drop clang-format globs for ios/*.c and android/jni/*.c, which match nothing since the Zig port. Fact-checked by an adversarial review subagent before commit; its findings are applied. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
GenericJam
force-pushed
the
docs/decision-log-and-accomplishments
branch
from
October 24, 2026 07:46
ccd27ce to
2fcfbf1
Compare
The commit counts are pinned to 8b7f7aa: the HEAD form of the command drifts as soon as the doc lands. The torch entry was toggled over dist on Android but from an on-device button on iOS. Release screenshots are opt-in. The Zig driver tables came in b361383, and ab9ff59 deleted the C ones. The restored dirty-NIF ADR notes its updated See-also line. Rule 20 is scoped to core android/jni/.
Owner
Author
|
Review: Codex was rate-limited until 5:02 AM, so a reviewer subagent did it. It said MERGE with three should-fix and three nitpicks; all six were fixed in 081c957, and the re-review said MERGE. All 24 commit hashes, the 10 PR numbers, the NIF counts and the defer coverage were checked against the source. |
Merged
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Docs-only, plus one comment path in
ios/mob_nif.m. No code or build changes.What
decisions/2026-05-09-zig-for-android-nif-and-native-build.md: retroactive ADR for why core Android native code is Zig. Covers NIF crashes killing the VM, types,deferkeeping cleanup next to the acquire, andcomptime. Describes what the build actually does:build.zigorchestrates, while Apple clang/swiftcand NDK clang are still invoked. Consequences:deferadoption is partial (40/45 env frees, 7/22 unlocks, 2/6 JNI string releases deferred);.ReleaseSmallin every build; two live Zig pins (Android NIF builds on both Zig pins; pin Keyboard.dismiss and Android blur in tests #226); thed5038c5bug class; agents defaulting to C.decisions/2026-04-28-dirty-nifs.md: restoresdocs/decisions/0001-dirty-nifs.md, removed in the unrelated commit88532d9whileios/mob_nif.mstill cited it. The body is the original text. The Status line marks its regular-scheduler list superseded by2026-09-05-input-nifs-are-dirty-io.md. The iOS comment now cites the new path.ACCOMPLISHMENTS.md: verified hard problems agents solved in this repo, each with commit/PR/ADR evidence, to be read before declaring a task impossible. Includes the rules for adding entries.AGENTS.mdandroid/jni/is Zig.ios/*.candandroid/jni/*.c, which match nothing since the Zig port.Verification
origin/master, and every PR number, path, and count was checked by an adversarial fact-check subagent before commit. Its findings are applied: 74 NIFs (not 79), the audio-probe scope, the full battery range, the build-toolchain description,defercoverage, plugin NIF language policy, and Bluetooth extraction commits.future_developments.md:454citesdocs/decisions/0002-wiretap-poc-then-port.md, which never existed in this repo's history. Left alone rather than inventing its content.