0.14 --- a keystroke the environment keeps for itself, and the sense in which zero is the weaker claim - #38
Merged
Merged
Conversation
… keystrokes, with the cross-repo plan A design proposal, no implementation. Full design and task graph in .agents/docs/2026-09-18-kal-term-control-design.md; arising from consumer report mcpplibs/openkal-musl#36 (a program cannot enter raw mode or keep Ctrl+C from killing it). * The position pair. KAL_TERM_CONTROL (mode word bit 2): 1 = the environment consumes reserved control keystrokes, 0 = every keystroke is a byte, including 0x03. KAL_TERM_PROP_CONTROL (props word bit 2): the implementation distinguishes the position. * Why a third bit and not a fold into LINE_EDIT. readline turns line editing off and keeps control-key interruption on every platform (ICANON off + ISIG on; ENABLE_LINE_INPUT off + ENABLE_PROCESSED_INPUT on); the orthogonal bit is the native shape and the only 7.1-conformant one. * Why the props bit is not optional. For this position alone, clause 6.2's unassigned-reads-zero does not match released behaviour (openkal-linux preserves ISIG): a program that needs the keystrokes-are-data guarantee shall enquire kal_terminal_props first. Recorded as the normative exception, closed by the claim-is-a-claim principle of clause 7.11. * The dependency graph. openkal 0.14 (hard first) -> mcpp-index registration -> openkal-linux/windows/macos in parallel -> openkal-musl last (largest change: the ioctl dispatcher routes to kal_terminal_* and cfmakeraw becomes correct for free). Documents and comments carry no emoji.
…in which zero is the weaker claim openkal.terminal gains a third position in the mode word, and clause 6.2 gains the rule that decides how a position is spelled. * KAL_TERM_PASS_CONTROL. Set, every keystroke reaches the program as the bytes it produces, including the ones an environment reserves for itself; clear, the environment may reserve a set of its own choosing. Which keystrokes those are belongs to the environment: an interface enumerating them would require of every environment what one of them spells. The position covers every mechanism by which the environment keeps a keystroke and not the interrupt alone, so that one mode word does not mean two things on two environments. * The position is spelled so that ZERO IS THE WEAKER CLAIM, and clause 6.2 now states that as a rule rather than leaving it to each assignment. An implementation released before 0.14 reads as zero and thereby says something true about itself. The arrangement in which zero would have been the guarantee needed an exception to clause 6.2 and a property position beside the mode position, to repair a hazard the spelling itself removes; it is recorded in SPEC.md entry 20 as considered and not adopted. * Two rules beside set_mode, both of which a multi-mechanism position makes visible. A position whose requested value is the one in effect is not written, so a program turning the echo off does not restore keystrokes its user had released. And a mode is not a way to end the input: an implementation that turns line assembly off shall not make a read report zero while input has not ended, which clause 7.4 says denotes the end of it. A bound upon waiting is stated by openkal.timeout. * KAL_VERSION_MINOR was 11 against a package at 0.13, and every implementation answers kal_version with that constant --- so the comparison clause 6.2 gives a consumer bound at load was between two equal numbers on every implementation there is. Corrected to 0.14.0 and checked by tools/check-version.sh, which CI also proves rejects a drift. * The suite observes the position in the only terms it has: asking for it is not an error, what is read back is either the position or zero, and the terminal is left as it was found. Whether the keystroke then arrives as a byte is measured above a C environment, upon a pseudo-terminal. Arising from consumer report mcpplibs/openkal-musl#36. Design, execution plan and the review of the earlier proposal are in .agents/docs/2026-09-20-*. Documents and comments carry no emoji.
This was referenced Sep 20, 2026
Merged
Merged
Merged
…ms reads differently `sed -i` takes the next argument as a backup suffix there, so the probe edited a file named by the expression and the check it was proving reported success. Written beside and moved over instead.
…his wave The Web target produces openkal-conformance.js beside a .wasm, and the selection named only the extensionless and the .exe forms --- so a build that had just succeeded was reported as 'expected exactly one suite to have been produced, found 0', which names neither the cause nor anything a reader can act upon. Measured on mcpp 2026.9.18.3, where the engine's Web output has that shape.
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
1 participant
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.
openkal.terminal的模式字加第三个位，条款 6.2 加一条决定「位怎么拼」的规则。起因是消费者报告 mcpplibs/openkal-musl#36：程序无法让每个按键以字节到达，因而被中断键结束。本 PR 取代 #37 的方案（#37 的设计稿作为第一个提交保留在历史里），差异在于位的极性。
方案
KAL_TERM_PASS_CONTROL（模式字位 2）：1 = 环境不为自己保留任何按键，每个键以字节到达；0 = 环境可保留一组由自己定义的按键。1 方向是保证，0 方向是许可；保留哪些键属于环境，规范不列举（条款 7.1）。^S冻终端而 Windows 不冻。set_mode旁两条规则：请求值与当前值相同的位不写（多机制的位上，这条决定了「只改回显的程序不会恢复用户已释放的按键」）；模式不是结束输入的方式（关行组装不得让读返回 0，条款 7.4）。要「会放弃的读」用openkal.timeout。KAL_VERSION_MINOR从 11 校正到 14：每个实现的kal_version就是return KAL_VERSION;，所以停滞的数字是整个生态对外的自述，条款 6.2 给 load 期绑定消费者的那个比较一直是两个相等的数。新增tools/check-version.sh，CI 同时证明它会拒绝漂移。验证
tools/check-version.sh、check-declarations.sh（104 names）、check-types.sh（175 declarations）、gen-macros.sh重生成无 diff、mcpp build全绿。kal_version = 0xe0000。^C以 0x03 到达）在 openkal-musl 侧以 pty 探针与宿主 C 库对照，见该仓库同名分支。生态
规范先行，随后 mcpp-index 注册；四个实现（linux / macos / windows / emscripten）并行适配；openkal-uefi、openkal-opensbi 跟随；openkal-musl 与 openkal-llvm-runtime 收尾。分支名在各仓库一致：
openkal-0.14。设计稿、执行计划与对 #37 的 review 在
.agents/docs/2026-09-20-*。