v4.6.0

What's Changed

• chore: release v4.5.0 by @github-actions[bot] in #1474
• semantic: ask and answer in each language's own word (M2 sheet B8) by @codetalcott in #1475
• semantic: scroll … by with a direction in each language's own words (M2 sheet B9) by @codetalcott in #1476
• semantic: a handler's count reads in every language (M2 row 8) by @codetalcott in #1477
• semantic: a query of a pseudo-class alone reads in every language (M2 row 8) by @codetalcott in #1478
• scripts: vitest-run.sh gives each run its own TMPDIR and removes it by @codetalcott in #1479
• semantic: rare value spellings read in every language (M2 row 8) by @codetalcott in #1480
• semantic: collection expressions read in every language (M2 row 8, P57) by @codetalcott in #1481
• semantic: a query's scope inside a phrase reads (M2 row 8, P62) by @codetalcott in #1482
• semantic: handcrafted patterns read their schema's type lists (M3 a) by @codetalcott in #1483
• semantic: 17 more handcrafted copies read their schema's type list (M3 a) by @codetalcott in #1484
• semantic: the copies narrower or wider than their list, judged (M3 a) by @codetalcott in #1485
• semantic: qu's transformer until-heads are read-only (M3 b, R6) by @codetalcott in #1486
• docs: G1 measured — a coverage penalty changes no pattern choice (M3 c) by @codetalcott in #1487
• semantic: split the clause keywords and js blocks out of the parser (M3 d) by @codetalcott in #1488
• semantic: split the tail reclaims and parse diagnostics out of the parser (M3 d) by @codetalcott in #1489
• semantic: split the per-command rescues and event words out of the parser (M3 d) by @codetalcott in #1490
• semantic: split the trailing-role attachers out of the parser (M3 d) by @codetalcott in #1491
• semantic: split the stateless loop helpers out of the parser (M3 d) by @codetalcott in #1492
• semantic: split the SOV clause reader and event-head helpers out of the parser (M3 d) by @codetalcott in #1493
• semantic: the parser's state becomes one object; its leaves move out (M3 d) by @codetalcott in #1494
• semantic: split the clause walk and statement strategies out of the parser (M3 d) by @codetalcott in #1495
• semantic: split the event-handler builder and parseInternal out of the parser (M3 d) by @codetalcott in #1496
• docs: plan the domains rethink, one reliable core (M3 e) by @codetalcott in #1497
• semantic: the last five silent pairs read back as their source (Reliability row 1) by @codetalcott in #1498
• semantic: fetch takes a $ or : variable URL with as <type> (Reliability row 2) by @codetalcott in #1500
• semantic: toggle reads a possessive style property in every language (Reliability row 2) by @codetalcott in #1501
• semantic: a js block after then behind a fused handler head is read whole (Reliability row 2) by @codetalcott in #1502
• semantic: a snake_case name is one word in every language (Reliability row 2) by @codetalcott in #1503
• semantic: a custom event after a handler head's or is an additional event (Reliability row 2) by @codetalcott in #1504
• semantic: put … at end of reads back in vi and five SOV languages (Reliability row 2) by @codetalcott in #1505
• semantic: a set target no marker ends runs through .member, [index] and a spaced ( (Reliability row 2) by @codetalcott in #1506
• semantic: fetch's as Object and as Number read back behind a bn/hi/ko handler head (Reliability row 2) by @codetalcott in #1507
• semantic: a loop head's variable is never a handler's event (Reliability row 2) by @codetalcott in #1508
• semantic: a loop's bottom test reads back in de, fr and ko (Reliability row 2) by @codetalcott in #1509
• semantic: a ko second handler on a parameterized event, and pl's do before a throw (Reliability row 2) by @codetalcott in #1510
• semantic: bn, th and vi read their own word for checked (Reliability row 2) by @codetalcott in #1511
• docs: Reliability row 2 done; the plan for M3 C3, one reader for an end word by @codetalcott in #1512
• semantic: every end-word reader's definition in one module (M3 C3 phase 1) by @codetalcott in #1513
• semantic: one definition of an end word, endWordKind (M3 C3 phase 2) by @codetalcott in #1514
• semantic: "closes a block" through closesBlockAt, for the readers that asked it (M3 C3 phase 3a) by @codetalcott in #1515
• semantic: the block readers that never asked "is it a value" ask closesBlockAt (M3 C3 phase 3b) by @codetalcott in #1516
• semantic: "stops a clause" through endsClauseAt (M3 C3 phase 4) by @codetalcott in #1517
• semantic: an end word's place is read by its index, everywhere (M3 C3 phase 5) by @codetalcott in #1518
• docs: C3 done, one reader per question for an end word (M3 C3 phase 6) by @codetalcott in #1519
• docs: domains rethink decisions D1-D6 taken; Phase 1 is domains 4.0.0 by @codetalcott in #1520
• mcp-server: domains ^4.0.0, jsx tools retired (domains rethink Phase 1) by @codetalcott in #1521
• docs: domains rethink Phases 1 and 2 done; the domain gate's first numbers by @codetalcott in #1523
• framework: a domain DSL's parse refuses input it did not read (rethink Phase 3.1) by @codetalcott in #1524
• framework: translate and the domain tools refuse what does not read back (rethink Phase 3.2-3.3) by @codetalcott in #1525
• docs: domains rethink Phase 3 merged (unreleased); the gate's corrected numbers by @codetalcott in #1526
• mcp-server: execute_llm says to quote a question when it refuses one by @codetalcott in #1528
• docs(changelog): no line break inside a code span, so prettier leaves it alone by @codetalcott in #1527
• chore(release): prepare 4.6.0 by @codetalcott in #1529
• ci(release): the release workflows' test budget follows ci.yml's semantic step (600s) by @codetalcott in #1530

New Contributors

• @github-actions[bot] made their first contribution in #1474

Full Changelog: v4.5.0...v4.6.0