v1.1.0

Collects v1.1.0

Diff since v1.0.0

• A bunch of inference and performance improvements.

• New feature: single argument method of collect_as, behaves like partial application. That is, collect_as(type) is usually equivalent with Base.Fix1(collect_as, type).

Merged pull requests:

• deduplicate IteratorSize(i) isa Union{Base.HasLength, Base.HasShape} (#34) (@nsajko)
• improve effects for iterators with unknown eltype but known length (#37) (@nsajko)
• outline throws and, in one instance, use a LazyString (#40) (@nsajko)
• build(deps): bump actions/checkout from 4 to 6 (#42) (@dependabot[bot])
• Fix word duplication in readme (#43) (@MagicMuscleMan)
• use Dependabot instead of CompatHelper (#45) (@nsajko)
• CI: test on Julia release channels v1.12 and nightly (#46) (@nsajko)
• make collect_as callable with a single positional argument (#47) (@nsajko)
• empty iterator error handling: improve error message, inlining hints (#49) (@nsajko)
• empty union error handling: improve inlining hinting (#50) (@nsajko)
• refactor: introduce push, used in the implementation of push!! (#51) (@nsajko)
• eliminate unnecessary copy when collecting to Vector (#52) (@nsajko)
• refactor: collect_as_vector_with_unknown_eltype_and_known_length (#53) (@nsajko)
• refactor: deduplicate methods of push!! (#54) (@nsajko)
• refactor: very slight simplification for Memory output (#57) (@nsajko)
• refactor: unify some logic for invariant types (#58) (@nsajko)
• Memory output: happy path for iterator with known length but not type (#59) (@nsajko)
• update readme for new method of collect_as, release v1.1 (#60) (@nsajko)

Closed issues:

• effect_free not inferred, unlike with Base.collect (#33)
• effect_free still not inferred for Memory (#38)
• Support piping (#44)

v1.0.0

Collects v1.0.0