Skip to content

grayscale: v0.11.0

Choose a tag to compare

@github-actions github-actions released this 28 Sep 21:48
92e0591

0.11.0 (2026-09-28)

Features

  • grayc: support TinyCC as a C backend (#2826) (54b6498)
  • parser: allow ';' as an optional statement and member separator (#2590) (a882151)

Bug Fixes

  • arrays: order u64 arrays as unsigned and return the element type from get_sum, get_min and get_max (#2783) (869bc73)
  • arrays: sort and is_sorted arrays of i8, i16, i32, u16, u32 and f32 at their real width (#2781) (a74484e)
  • binary: take and return the named type in the 8- to 64-bit encode and decode functions (#2867) (fafac6b)
  • codegen: check that char(n) and casts to char receive a valid codepoint (#2857) (7d99777)
  • codegen: compare unsigned and signed integers by value (#2858) (d267dfc)
  • codegen: default an enum with an explicit first value to its first variant (#2844) (06e9fed)
  • codegen: emit a module-level mut's constant initializer without a runtime range check (#2872) (f08a941)
  • codegen: give a fixed-size array field N slots in new() (#2893) (bc5a770)
  • codegen: overflow- and zero-check compound assignment to integer map values (#2871) (282b5ec)
  • codegen: overflow-check unary negation of i128 and i256 values (311fc26)
  • codegen: print float, f32, and f64 through one width-aware formatter (890d3f5)
  • codegen: range-check narrowing casts from u64 and float sources before converting through int64 (#2856) (dade0c5)
  • codegen: read and write [f32] elements as float in random.choice and arrays.fill (#2849) (5c08674)
  • codegen: resolve the sqlite Database type to GraySqlite * (#2885) (56580dc)
  • codegen: shift sized integers at the operand's width and bound the shift amount by it (#2855) (957a572)
  • codegen: sort arrays of bool at their 1-byte width in sort_asc and sort_desc (#2873) (985e7f4)
  • codegen: stage [f32] values as float in arrays.insert_at and arrays.prepend (#2834) (1af5d92)
  • codegen: store fixed-size [f32, N] arrays as packed float (f7249ce)
  • codegen: support bitwise operators on i128, u128, i256 and u256 values (#2875) (ec9aa87)
  • codegen: zero-fill a fixed-size struct field given fewer elements than its length (1293138)
  • codegen: zero-filled [T, N] slots hold the element type's zero value (#2842) (4da468e)
  • math: panic on overflow in factorial, lcm, gcd, abs and neg (#2785) (3532e4c)
  • math: return the argument type from min, max and clamp (#2866) (a106897)
  • math: support i128, i256 and u128 arguments in abs and neg (#2874) (e279e66)
  • parser: accept a lowercase type alias as the first type in a multi-value return (#2850) (6821456)
  • parser: parse a fixed-size array of any element type (#2841) (cc44edc)
  • parser: stop an E3134 alias error from skipping the next declaration (#2897) (8d1aba8)
  • stdlib: make http and server share one HttpResponse type (#2886) (3c0331e)
  • tests: update codegen e2e tests to the one-signature error and threads.spawn (bc52de3)
  • typechecker: accept an array or map literal reassigned to a variable with sized elements (47bae94)
  • typechecker: check the length of a non-literal value stored into a fixed-size struct field at runtime (#2830) (64db2d2)
  • typechecker: enforce signedness for 128- and 256-bit integers in fmt %d and %u directives (6d582ff)
  • typechecker: give a narrow operand combined with a wide integer the wide type (73c6160)
  • typechecker: print the whole fixed-size array type in E3054 (#2894) (0ebcb1a)
  • typechecker: reject a one-argument random.rand_f64 or random.rand_char call with E5008 (#2882) (0285832)
  • typechecker: reject an alias whose target is not a type (#2891) (3255738)
  • typechecker: reject returning a wide integer as a narrower or differently signed integer (ea44ea3)
  • typechecker: report an f64 strings.repeat count or strings.slice bound once (#2878) (2dec0a7)
  • typechecker: report an undefined parameter or return type inside a func(...) type with E4016 (#2854) (68f5f66)
  • typechecker: require a ^i64 pointer argument for the 64-bit atomic functions (#2868) (82d6311)

Performance Improvements

  • typechecker: index file-scope const integers in a hash table instead of scanning them linearly (5836cd1)