Swiftpack.co - Swift Packages by pointfreeco

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. ## What's Changed * Fixed: Re-entrant actions from store publisher subscriptions no longer run with the previous state (thanks @diederich, #1360). * Fixed: The test store is now slightly more forgiving when it comes to Swift's async runtime. If you are having issues with async effects and the test store, let us know and we'll do our best to accommodate! * Fixed: The SPM package no longer publicly vends its internal schemes. - Fixed: Removed the `@StateObject` shim added in 0.40.0 to avoid a release build regression in some applications. We'll investigate on bringing it back in the future! This release brings many performance improvements to the library. See our [blog post](https://www.pointfree.co/blog/posts/80-improving-composable-architecture-performance) for information about those changes. - Added: `WithViewStore(_:observe:…)`, a new initializer that makes state observation explicit. We hope these initializers will help users avoid performance pitfalls and identify areas of their application where performance can be improved. The existing initializers have been soft-deprecated and will surface warnings in a future release. - Performance: `Effect` performance has been improved, especially with regard to merging effects (a common operation that occurs when reducers are combined), and async effects (which previously often incurred overhead from Combine). - Performance: `Store.scope` performance has been improved for nested scopes, which no longer incur a penalty for each layer of scoping (thanks @iampatbrown). - Performance: `WithViewStore` now uses a state object under the hood, which should tie the lifecycle of a view store to a view (thanks @tgrapperon). - Performance: `ForEachStore` performance has been improved for state updates in which element state has changed but no elements were added, removed, or moved (thanks @tgrapperon). - Fix: `Effect.run`'s `Send` closure is now explicitly `@MainActor`. - Fix: Long-living effects no longer accumulate `Task` values in memory for each action sent back into the store, saving memory over time. - Fix: macOS 11 compilation with the Xcode 14 release candidate. - Fix: Actions that re-enter the store at the end of processing are now processed immediately. - Infrastructure: README fixes (thanks @Jager-yoo). - Infrastructure: Added benchmarks for effects, store scoping. - Infrastructure: Fixed Todos demo's incrementing UUID helper (thanks @pitt500). - Infrastructure: Fixed Voice Memos AV settings (thanks @entaku0818). - Fixed: Pinned Combine Schedulers to [0.7.3](https://github.com/pointfreeco/combine-schedulers/releases/0.7.3), which removes an errant `@_exported import Foundation`. > **Note**: This update addresses a breaking change introduced to Combine Schedulers (see the [release notes](https://github.com/pointfreeco/combine-schedulers/releases/0.7.3) for more). As a consequence, you may need to add explicit `import Foundation`s to your code base after updating to this version. - Changed: Updated `Global`/`Local` generics to more relative `Parent`/`Child`, or more domain-specific names. - Infrastructure: Added Spanish language translation of the README (thanks @pitt500). - Infrastructure: README and documentation improvements (thanks @ole, @SevioCorrea, @kalupas226) - Infrastructure: Demo cleanup. - Infrastructure: Made test classes `final` (thanks @Jager-yoo). - Added: This release includes many new tools to better support Swift concurrency. See [this discussion](https://github.com/pointfreeco/swift-composable-architecture/discussions/1186) and [this pull request](https://github.com/pointfreeco/swift-composable-architecture/pull/1189) for more information. > **Note**: Many of the old Combine-centric interfaces have been [soft-deprecated](https://pointfreeco.github.io/swift-composable-architecture/0.39.0/documentation/composablearchitecture/effectdeprecations). Prefer using `Effect.task`, `Effect.run`, and `Effect.fireAndForget` as described in [the documentation](https://pointfreeco.github.io/swift-composable-architecture/0.39.0/documentation/composablearchitecture/effect). - Infrastructure: README, documentation, and demo app fixes (thanks @filblue, @Jager-yoo, @konomae, @yimajo). * Deprecated: `Store.unchecked` has been deprecated. As the Composable Architecture migrates to modern Swift concurrency tools, its async endpoints must run on a consistent actor to ensure safety from data races. Because the primary use case of the Composable Architecture is driving UIs, we have chosen the `@MainActor` for these endpoints, which makes them unsafe to call from unchecked stores. We believe the number of unchecked stores out in the wild is small, but in the meantime are soliciting feedback from the community with this deprecation warning. See #1206 for more information and migration paths. - Changed: Bumped dependencies to include Combine Schedulers 0.6.0 and XCTest Dynamic Overlay 0.3.0, which includes the renamed `UnimplementedScheduler`, and `XCTUnimplemented`, respectively. - Infrastructure: Cleaned up and simplified some demo apps and case studies. - Fixed: 0.38.0 introduced a regression that could cause test stores that receive `BindingAction`s from effects to cause test failures. This has been fixed. - Added: A new runtime warning for when a `BindingAction` is sent to the store but `Reducer.binding()` wasn't applied to the reducer. This should help diagnose missing integrations faster. - Fixed: A regression for `TestStore.send` and `receive` was introduced in 0.37.0 that failed to perform assertions against state when the trailing closure was omitted. This has been fixed (thanks @umbertovolta). - Added: `Effect.animation()`, which adds animation to an effect without the need of a scheduler. - Changed: `TestStore` state modification failures are less noisy. - Renamed: `ViewStore.suspend(while:)` has been renamed to `ViewStore.yield(while:)`. - Bug fixed: A bug was introduced to the deprecated `TestStore.assert` APIs in 0.36.0, which caused test failures when a sent/received test action left state unchanged. This has been fixed. - Infrastructure: Worked around a SwiftUI bug/warning in the Case Studies demo; cleaned up some of the library's generic signatures.

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. ## Changes that may affect your build requirements - The package name has been changed from `SnapshotTesting` to `swift-snapshot-testing`, to better reflect community naming conventions (https://github.com/pointfreeco/swift-snapshot-testing/pull/555). This may require you to update any `Package.swift` files that depend on Snapshot Testing. - Support for CocoaPods and Carthage has been obsoleted. No new releases will be available on those platforms, starting with 1.10.0. Please use the Swift Package Manager to depend on Snapshot Testing. - Snapshot Testing now requires Swift 5.5+, as well at iOS 13+, macOS 10.15+, tvOS 13+, and watchOS 6+. ## Changes that may affect snapshot format/output > **Warning**: The following updates change how a strategy may output a snapshot, causing existing snapshot suites to fail. When upgrading, please rerecord your failing snapshots and audit the changes. - Fixed: A UIViewController strategy was not passing along the correct trait collection (thanks @arnopoulos, https://github.com/pointfreeco/swift-snapshot-testing/pull/554). - Fixed: iPhone XS Max portrait configuration was not applying base traits (thanks @imvm, https://github.com/pointfreeco/swift-snapshot-testing/pull/417) - Fixed: URLRequest strategy now sorts query items for more consistent diffing (thanks @mihai8804858, https://github.com/pointfreeco/swift-snapshot-testing/pull/491). - Removed: GLKView is no longer supported (thanks @jpsim, https://github.com/pointfreeco/swift-snapshot-testing/pull/507). ## Everything else - Added: Image-based strategies have a new `perceptualPrecision` option, which can be used to support snapshot tests across Intel and M1 devices. Along for the ride is a 90+% speed improvement to the strategies (thanks @ejensen, https://github.com/pointfreeco/swift-snapshot-testing/pull/628). - Added: Windows support, CI (thanks @MaxDesiatov, https://github.com/pointfreeco/swift-snapshot-testing/pull/532). - Added: Failure messages now include the snapshot name, if given (thanks @calda, https://github.com/pointfreeco/swift-snapshot-testing/pull/547). - Added: iPhone 13 configs (thanks @luispadron, https://github.com/pointfreeco/swift-snapshot-testing/pull/603). - Added: iPad 9.7" and iPad 10.2" configs (thanks @skols85, https://github.com/pointfreeco/swift-snapshot-testing/pull/405). - Added: iPhone 12 and iPhone 12 Pro Max configs(thanks @imvm, https://github.com/pointfreeco/swift-snapshot-testing/pull/418). - Added: `Snapshotting.json` for `Any` value (thanks @NachoSoto, https://github.com/pointfreeco/swift-snapshot-testing/pull/552). - Added: Newly-recorded snapshots are now attached to the test run (thanks @marcelofabri, https://github.com/pointfreeco/swift-snapshot-testing/pull/586). - Fixed: the UIImage strategy was using the device scale instead of the scale of the given images (thanks @codeman9, https://github.com/pointfreeco/swift-snapshot-testing/pull/472). - Fixed: the UIImage strategy now compares image contexts using the same colorspace (thanks @dflems, https://github.com/pointfreeco/swift-snapshot-testing/pull/446). - Fixed: the WebView strategy no longer overrides the delegate (thanks @teameh in https://github.com/pointfreeco/swift-snapshot-testing/pull/443). - Fixed: Patched a leak when running tests in a host application (thanks @llinardos, https://github.com/pointfreeco/swift-snapshot-testing/pull/511). - Fixed: False positive when asserting reference image with empty image (thanks @nuno-vieira, https://github.com/pointfreeco/swift-snapshot-testing/pull/453). - Fixed: watchOS compilation (thanks @aydegee, https://github.com/pointfreeco/swift-snapshot-testing/pull/579). - Fixed: Support for repeatedly running tests (thanks @krzysztofpawski, https://github.com/pointfreeco/swift-snapshot-testing/pull/585). - Infrastructure: Add `swift-snapshot-testing-stitch` to plugins list (thanks @Sherlouk, https://github.com/pointfreeco/swift-snapshot-testing/pull/483) - Infrastructure: Document diff tool configuration. - Infrastructure: README updates (thanks @MaatheusGois, @gohanlon), documentation fixes (thanks @heckj, @valeriyvan). - Infrastructure: Added GitHub issue templates (https://github.com/pointfreeco/swift-snapshot-testing/pull/556). - Infrastructure: Add SnapshotTestingHEIC to plugins list (thanks @alexey1312, https://github.com/pointfreeco/swift-snapshot-testing/pull/561). ## New Contributors * @arietis made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/468 * @codeman9 made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/472 * @Nikoloutsos made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/482 * @mihai8804858 made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/491 * @dflems made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/446 * @skols85 made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/405 * @imvm made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/418 * @teameh made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/443 * @MaatheusGois made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/496 * @calda made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/547 * @MaxDesiatov made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/532 * @jpsim made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/507 * @arnopoulos made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/554 * @llinardos made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/511 * @heckj made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/601 * @aydegee made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/579 * @valeriyvan made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/630 * @ejensen made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/628 * @luispadron made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/603 * @alexey1312 made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/561 * @gohanlon made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/592 * @marcelofabri made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/586 * @nuno-vieira made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/453 * @krzysztofpawski made their first contribution in https://github.com/pointfreeco/swift-snapshot-testing/pull/585 **Full Changelog**: https://github.com/pointfreeco/swift-snapshot-testing/compare/1.9.0...1.10.0 - Added: publicized `data` snapshotting strategy (thanks @regexident). - Speed up `NSImage` comparison (thanks @finestructure, @JaapWijnen). - Fixed: enabled testing search paths (thanks @thedavidharris). - Infrastructure: links for AccessibilitySnapshot and AccessibilitySnapshotColorBlindness (thanks @Sherlouk). - Rename `SnapshotTesting.record` to `SnapshotTesting.isRecording` to avoid clash with XCTestCase's new `record` method (thanks @xavierLowmiller). - Fix UIApplication selector to return the application as expected (thanks @mstultz). - Fix key window warning (thanks @tinder-maxwellelliott) - Change minimum iOS version from iOS 10 to 11 to fix SwiftUI linking problems (thanks @thedavidharris). - Added: SwiftUI support (thanks @natemann, @regexident). - Added: Apple TV 4K support (thanks @reez). - Added: Mac Catalyst support (thanks @rjchatfield). - Added: UIBezierPath, NSBezierPath, CGPath strategies (thanks @regexident). - Improved: don't crash on empty images, instead produce error artifact (thanks @mackoj). - Bug fixed: trait collection now passed to view controller's recursive description strategy (thanks @erikpoort). - Bug fixed: will no longer crash on hidden/unloaded web views (thanks @stuaustin). - Bug fixed: don't re-add view controller to window when already added (thanks @hassfers). - More test coverage (thanks @SatoTakeshiX, @sidepelican). - Typos fixed (thanks @freak4pc, @jstart). - Other improvements: timeout error messaging, installation instructions, troubleshooting instructions, Linux CI. This release fixes a bug introduced in the Xcode 11.4 beta with SPM and Swift 5.2. This minor release introduces backwards compatibility for SwiftPM Tools @ 5.0. ### What's new? - New snapshot strategy: `wait` (#268) ### What's improved? - Performance improvement: don't diff during recording (#250) - Documentation improvement: Xcode 11 installation instructions (#256) - NSView rendering improvements (#260) - Documentation improvement: SPM test dependency instructions (#265) - cURL rendering improvements (#273) Thanks to @mr-v, @f-meloni, @schwmi, @freak4pc, and @ldstreet for contributing to this release! ### What's changed since last time? - Add device sizes for split view variants of iPads (#209) - Add recording to inline snapshotting (#212) - Sort cURL strategy headers (#214) - Add iOS minimum required deployment target to Package.swift (#215) - Allow dynamic size of views based on trait collection content sizes (#217) - Disable bitcode (#221) - Improve `_assertInlineSnapshot` ergonomics and tests (#231 and #232) - Use `URL.init(fileURLWithPath:isDirectory:)` to avoid file IO (#236) - Speed up image diffing (#248) - Improve image diff drawing performance (#249) Thanks to @Sherlouk, @crayment, @jayhickey, @MarianaMeireles, @pavel-y-ivanov, @ferranpujolcamins, @kirillyakimovich, and @mr-v for making this release happen! ### What's new * @rjchatfield has introduced (#199) a brand new way of snapshotting textual information, called "inline snapshot testing," which automatically inlines snapshots directly in your test file. The 1.5.0 release includes a preview of this functionality! You can use the `_assertInlineSnapshot` helper to try it out. Thanks to @rjchatfield for the PR, and find out more [here](https://medium.com/@rjchatfield/inline-snapshot-testing-e5e237fc47f3).  * Two new `URLRequest` snapshot strategies have been included. First, @Sherlouk has included a `curl` snapshotting strategy (#203), which snapshots your prepared requests in the cURL format. Also we now have a snapshot strategy for capturing a `URLRequest` with its body pretty printed, when possible (#157). This helps make `POST` requests more easily inspectable in the reference file.

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. ### What's new? - Not much! We build for Swift 5 now! ## What's new * Add "combining" and "their" (#33) * Unzurry (#15) … * Add `update` (#39) ## What's fixed * Update to Swift 4.2 (#29) * Info.plist for targets (#35) * Make Overture app extensions friendly (#37) * Fix ambiguity with `with` and `mver` (#39) Thanks to https://github.com/pointfreeco/swift-overture/pull/26 Overture 0.3.0 includes a handful of handy `zip` functions! This includes: - N-ary overloads of `zip` on sequences, zipping up to 10 sequences at once - `zip` on optionals, zipping up to 10 optionals at once - `zip(with:)` on both sequences _and_ optionals, for transforming zipped tuples into more interesting values See more about the announcement on our blog: [Overture: Now with Zip](https://www.pointfree.co/blog/ https://www.pointfree.co/blog/posts/15-overture-0-3-0-now-with-zip). Overture 0.2.0! This release is all about setters: functions that allow us to build complex transformations out of smaller units. See more about the announcement on our blog: [Overture: Now with Functional Setters](https://www.pointfree.co/blog/posts/4-overture-now-with-functional-setters). The first release of Overture! Overture is a function composition toolchain for Swift.

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. - Added: Conditional conformances to `ExpressibleByArrayLiteral` and `ExpressibleByDictionaryLiteral`. - Added: Conditional conformance to `Strideable` (thanks @TheOtherDave). - Added: Conditional conformance to `Sendable` (thanks @ericlewis). - Fixed: `Encodable` will now try to encode using the wrapped value's decoding initializer if the single value container fails to decode. - Infrastructure: Update README (thanks @volkdmitri, @mattyoung). - Added: dynamic member lookup support for tagged values (thanks @Dragna). - Added: `ExpressibleByStringInterpolation` support (thanks @mdarnall). - Changed: SPM's package name has changed from `Tagged` to `swift-tagged` in the spirit of Apple conventions, and conventions set in other @pointfreeco packages). Migration strategy: remove any instances of `name: "Tagged"` from package files depending on Tagged). ### What's new? In Tagged: - Fix Date decoding with custom strategy (#24) - LocalizedError conformance (#33) - Identifiable protocol support (#35) - Add Collection and Sequence conditional conformances (#36) - Swift 5.1 support (#38) In TaggedTime: - Add DispatchTimeInterval methods to TaggedTime (#25) Thanks to @buscarini, @pteasima, @kaunamohammed, and @tonyarnold for contributing! This release brings Tagged up to date with Swift 5! ## What's new? - Added `TaggedTime` and `TaggedMoney` modules for safer time and safer money (#10). - Added `coerce` (#15). ## What's fixed? - Codability should now work on non-single value containers by delegating more directly to the underlying type (#12). - Fixed Carthage (#17). * Removed `ExpressibleByNilLiteral` conformance: https://github.com/pointfreeco/swift-tagged/pull/4 * Added `CustomPlaygroundDisplayConvertible` conformance: https://github.com/pointfreeco/swift-tagged/pull/5 * Added `map` to `Tagged`: https://github.com/pointfreeco/swift-tagged/pull/6 * Fixed deployment targets: https://github.com/pointfreeco/swift-tagged/pull/8 This preliminary release has the basic `Tagged` type implemented with a few starting conformances to common Swift protocols.

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. - Added: `Binding.removeDuplicates()`, for ignoring the upstream binding's `set` when the wrapped value remains unchanged. - Fixed: `IfCaseLet`'s else branch will now render (thanks @junebash). - Infratructure: README fixes (thanks @AnasAlhasani), and case study improvements. - Initial release.

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. - Bug fixed: `NonEmpty` now manages its own synthesized iterator in order to avoid potential stack overflows when calling to APIs on the iterator of the collection it wraps. - Changed: `NonEmpty` no longer conditionally conforms to `StringProtocol`. Apple documents against this conformance, and though engineers say there should be no issue with doing so, ABI-wise, we encountered a compiler crash when fixing the bug in the previous bullet point. - Fixed: `+` operator between two `NonEmpty`s was ambiguous (thanks @buscarini). - Changed: `NonEmpty` now wraps a base collection rather than a single element and collection as separate fields. This improves the performance and reliability of working with `NonEmpty` collections. This does unfortunately mean that invariants can compile and make their way into the system at run time, but we do our best to ensure they cannot be constructed. - Changed: the SPM package name has been changed from `NonEmpty` to `swift-nonempty` (thanks @maximkrouk). - Fixed: range subscripts no longer cause runtime stack overflows (thanks @clang13). - Removed: CocoaPods and Carthage support has been deprecated in favor of SPM. - NonEmpty now correctly encodes and decodes structures as their more primitive collection types (#14) ### What's fixed: - Fix compiler error (#22) - Application extension support (#16, #18) This release brings Swift 5 support, including an optimization around the mutation of non-empty collections! Swift 4.2 on Linux has a regression that prevents `NonEmpty` from conditionally conforming to `RandomAccessCollection`. We're temporarily removing support for the sake of compatibility. This release updates NonEmpty to be compatible with Swift 4.2's randomness APIs. This preliminary release has the basic `NonEmpty` type implemented with a few starting conformances to common Swift protocols.

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. - Added: Support for `viewport-fit` (thanks @hallee). - Added: Support for the `accept` attribute (thanks @xavierLowmiller). - Added: an experimental rendered for XML, `_xmlRender`. It is prefixed with an underscore for now. It renders valid XML and avoids rendering "void" (non-closing) HTML tags. - Changed: the package has been renamed to `swift-html` to match Apple conventions. - Fixed: `srcset` is now rendered in a stable order. - Removed: support for Carthage and CocoaPods. If you depend on these package managers, you must pin to an older version than 0.4.0. This minor release restores support for Swift 4.2! ### What's new? - Swift 5 support (#51). - The element and attribute APIs have been completely revamped to use `static`s (#39). - The underlying `Node` API has been improved! - Arrays of nodes are now contained within a `.fragment` case, and `.element` has been updated to take a single child node, which may be a fragment (#23). - `Node` is now `ExpressibleByArrayLiteral` (#23). - `Node` is now `Equatable` and `Hashable` (#48). - `Node` now supports Swift 5 `ExpressibleByStringInterpolation` (#48). - The `mailto` API has been improved to take `subject` and other parameters (#41). - There's a new `template` tag helper (#43). The cumulative changes make for a big difference in API, but we think it's worth it! ``` diff -let link: Node = a([href("/")], ["Hello, ", .text(name), "!"]) +let link: Node = .a(attributes: [.href("/")], "Hello, \(name)!") ``` ### What's fixed? - Both `summary`/`details` and `fieldset`/`legend` have been fixed to compile correctly (#42). - The `HtmlSnapshotTesting` module has been fixed on Carthage by disabling bit-code (#49). Fixed a typo where `onkeydown` was mistakenly rendered as `onkeyup`. https://github.com/pointfreeco/swift-html/pull/35 ## What's new? - Support for Carthage and CocoaPods - A brand new module, `HtmlSnapshotTesting` for testing your HTML using [SnapshotTesting](https://github.com/pointfreeco/swift-snapshot-testing) - A new `debugRender` function, which renders HTML with human-readable indentation - Support for iOS, tvOS, and watchOS - Added a missing `style(unsafe:)` attribute helper that takes non-static strings - Added a missing `enctype` form attribute helper ## What's fixed? - Fixed `thead` and `tbody` element helpers to require nesting in a `table` This release fixes how empty, non-void elements are closed. Today we are open sourcing a new library for building HTML documents in Swift. It’s extensible, transformable, type-safe, and provides many benefits over templating languages. Read more [here](https://www.pointfree.co/blog/posts/16-open-sourcing-swift-html-a-type-safe-alternative-to-templating-languages-in-swift)!

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. - Added: `StartsWith`, `PrefixThrough`, and `PrefixUpTo` overloads to aid in type inference for substring and UTF-8 parsers (thanks @tgrapperon). - Changed: Parsing now has Apple platform requirements equivalent to SwiftUI (iOS 13+, macOS 10.15+, tvOS 13+, watchOS 6+). If these minimum requirements don't fit your needs, [let us know](https://github.com/pointfreeco/swift-parsing/discussions/new). - Fixed: `PipeEnd` now conforms to `ParserPrinter`, fixing printability of piped parsers (thanks @JaapWijnen). - Deprecated: the `_URLRouting` module is now deprecated. Use the [URL Routing](https://github.com/pointfreeco/swift-url-routing) package instead. - Removed: `Conversions.Parsing`, which was not reachable through any static member on `Conversion`, and confused some users. If you have a use case motivating this conversion, [let us know](https://github.com/pointfreeco/swift-parsing/discussions/new). - Fixed: Added a missing `ParserPrinter` conformance to `OptionalOneOf`, which prevented parsers that used `if` statements in `@OneOfBuilder` blocks from being printers. The following changes have been made to the more experimental `_URLRouting` module: - Added: A `Body()` initializer that takes no arguments. It simply parses the entire body as `Data`. - Infrastructure: documentation changes. - Fixed: A `Double.parser()` overflow bug has been fixed (thanks @tgrapperon). The following changes have been made to the more experimental `_URLRouting` module: - Added: A `URLRoutingClient` for wrapping a router and URL session into an HTTP client. - Fixed: A few potential bugs around `Field` and `Body` being empty. Introduced the `ParserPrinter` protocol for invertible parsing. - Added: A case-iterable, raw-representable parser. Simply tack `.parser()` onto any conforming type: ```swift enum Role: String, CaseIterable { case admin case guest case member } try Role.parser().parse("admin") // Role.admin ``` - Fixed: An Xcode 13.3 compiler error has been fixed. - Fixed: `Optionally` will now backtrack if the parser fails (thanks @randomeizer). - Fixed: `Double.parser()` now parses as freely as `Double.init`'s `LosslessStringConvertible` functionality. - Optimized: `Peek` and `Not` will only backtrack when they are successful (_i.e._, if `Peek`'s upstream parser successfully parses a value, or if `Not`'s upstream parser fails to parse a value). Backtracking on failure is now delegated to any upstream `OneOf`s. - Optimized: `OneOfMany` no longer backtracks its final failure, bringing it in line with the behavior of the variadic `OneOf`s. - Breaking change: The non-`inout` overloads of `Parser.parse` now attempt to fully consume `Collection`-based inputs. ```swift // Before: try Int.parser().parse("42hello") // 42 // After: try Int.parser().parse("42hello") // error: unexpected input // --> input:1:13 // 1 | 42hello // | ^ expected end of input ``` This change makes parsing a bit more strict by default in order to catch potential issues with input. If you want to ignore trailing output, use the `inout` version of `parse`, or explicitly describe how the input should be ignored in the parser, for example using `Optionally { Rest() }.map { _ in () }`. - Breaking change: The `Rest` parser now fails when the rest of input is empty. ```swift // Before: try Rest().parse("") // "" // After: try Rest().parse("") /// error: unexpected input /// --> input:1:1 /// 1 | /// | ^ expected a non-empty input ``` If your use of `Rest` should not fail on empty input, wrap it explicitly in an `Optionally` parser, or use `replaceError(with:)` to provide a default value of `""`. - Breaking change: `Peek` is now a `Void` parser. It can be used to inspect a value in order to test that a parser should be successful, but capturing any data is now the responsible for the parsers that comes afterward (thanks @randomeizer). - Breaking change: The `isSigned` parameter of `Int.parser()` has been removed. `Int.parser()` will now always parse a sign if `FixedWidthInteger.isSigned` returns true (_e.g._, `Int.parser()` will parse a sign, `UInt.parser()` will not.). If you want to parse a number without a sign, use a more explicit parser, or test for the sign before using `Int.parser()`. _E.g._: ```swift let digits = Prefix { $0.isNumber }.compactMap(Int.init) // ...or... let digits = Parse { Not { OneOf { "-"; "+" } } Int.parser() } ``` - Updated: `Double.parser()` can now be used on any type that conforms to `BinaryFloatingPoint`, including `Float16`. - Updated: `Many`'s `updateAccumulatingResult` can now throw. - Updated: Documentation has been revamped, including a new DocC-based static site with articles that cover common topics. - Infrastructure: Documentation fixes (thanks @haikusw). - Improved: Error messages that occur at the same input range are now coalesced into a single error (thanks @mayoff). - Fixed: `Prefix` now eagerly consumes input. - Fixed: `Prefix` with a minimum length now throws an error at the correct input location, rather than the input location of that minimum length. - Infrastructure: README and documentation updates. * Breaking change: `Parser.parse`'s method signature has changed from optional-returning to throwing: ```diff -func parse(_ input: inout Input) -> Output? +func parse(_ input: inout Input) throws -> Output ``` All of the parsers the library ships with now include error messages and context alongside failure, which can be printed: ```swift do { let output = users.parse(&input) } catch { print(error) } // error: multiple failures occurred // // error: unexpected input // --> input:3:11 // 3 | 3,Blob Jr,tru // | ^ expected "true" or "false" // // error: unexpected input // --> input:2:16 // 2 | 2,Blob Sr,false // | ^ expected end of input" ``` This is unfortunately a breaking change. It does not seem possible to support both throwing and optional-returning at the same time without requiring alternate breaking changes, so we opted for a clean break that requires upgrading existing parsers to throwing parsers. For most of our users, it will hopefully be a matter of prepending `try?` to lines that call `parse` to get things building again, though we hope folks will take full advantage of the new error messages. For users with custom parser conformances, they will need to do a little more work to make these parsers throwing. For more information about the release and migration strategies, see [our announcement](https://www.pointfree.co/blog/posts/73-parser-errors). * Added: `Parser.replaceError(with:)`, a parser modifier that transforms a throwing parser into a non-throwing parser by providing a default output. ```swift let sign = OneOf { "+".map { 1 } "-".map { -1 } } .replaceError(with: 1) var input = "-123"[...] // No `try` required: sign.parse(&input) // -1 input // "123" // Simply returns the default when parsing fails: sign.parse(&input) // 1 ``` * Added: the `Fail` parser has a new `.init(throwing:)` initializer that takes an `Error` for custom error messaging. * Deprecated: `Conditional` as a top-level parser has been deprecated. Use `if`-`else` builder syntax instead, which uses `Conditional` parsers under the hood, or reference the nested `Parsers.Conditional` type. * Infrastructure: documentation has been rewritten and expanded to reflect throwing parsers. * Infrastructure: Swift Package Index configuration was fixed so that our supported platform are better communicated (thanks @finestructure). * Breaking change: many parsers that used to explicitly revert changes made to `input` when parsing fails (a process known as "backtracking") no longer do. See [our announcement](https://www.pointfree.co/blog/posts/72-backtracking-parsers) and pull request #108 for more information about the change and its motivation. `OneOf`, which previously did no backtracking whatsoever and trusted other parsers to do the work, is now the main entry point into backtracking behavior. Backtracking has been removed from `Double.parser()`, `Parser.filter`, `Parser.flatMap`, `Many` (with the exception of backtracking a trailing separator), `Parser.pipe`, `PrefixThrough`, and `PrefixUpTo`. Note: `Peek` and `Not` preserve their backtracking behavior, as they are designed to never consume any input. * Infrastructure: Add tests for `CharacterSet.parse` (thanks @andrewjl). - Added: New result builder syntax for creating parsers! Use the `Parse` and `OneOf` entry points into `@ParserBuilder` and `@OneOfBuilder` blocks. - Added: A `Peek` parser: for running a parser and checking its result _without_ consuming input (thanks @randomeizer). - Added: A `Not` parser: for running a parser and succeeding if it fails (or failing if it succeeds) (thanks @randomeizer). - Changed: Many parsers have been updated to work with builder syntax, including `Skip`, `Many`, etc. - Changed: `Int.parser()` now takes an `Int` (not `Self`) as its radix. - Deprecated: Older, non-parser builder interfaces have been soft-deprecated. - Deprecated: Swift 5.3 support has been removed. - Performance: `Many`'s performance has been improved. - Changed: `Many` and `OptionalParser`'s `Upstream` generic has been deprecated and renamed to `Element` and `Wrapped`. - Infrastructure: benchmarks and documentation cleanup.

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. Fixed: The swift-case-paths package no longer publicly vends its internal schemes. - Fixed: The `CasePath.init` overloads introduced in 0.9.0 now support the same implicit optional promotion behavior of `/`. - Added: `CasePath.init(Enum.case)` alternative to `/Enum.case`. Xcode 14 beta 1 puts projects using Case Paths in a very broken state right now. This initializer is a workaround that can be used, instead, to avoid breaking things like syntax highlighting and file-wide indentation. - Fixed: case path composition operator associativity should now behave as expected (thanks @nonameplum). - Bug fix: A Linux bug that could prevent multiple associated values from being extracted has been fixed. - Changed: `CasePath.modify` now returns the value returned by its closure. - Fixed: runtime extraction of an `Error` existential should no longer crash (thanks @mayoff). - Fixed: runtime extraction of an object-bound protocol existential should no longer crash (thanks @iampatbrown). - Infrastructure: corrected README typos (thanks @ferologics) and example code, fixed CI and auto-format. - Added: `CasePath.modify`, which allows you to mutate the associated value of an enum in-line: ```swift try (/Result<Int, Error>.success).modify(&result) { $0 *= 2 } ``` - Bug fixed: a regression in which case path expressions to enum cases with no associated values, where the `Root` is optional-promoted by the compiler, extraction wasn't working correctly. - Bug fixed: failed extractions via the optional demotion handled in [0.6.0](https://github.com/pointfreeco/swift-case-paths/releases/0.6.0) would infinitely recurse. - Fixed regression: Swift can optional-promote the `Root` of an embed function `(Value) -> Root` to `(Value) -> Root?` and yield `CasePath<Root?, Value>` instead of `CasePath<Root, Value>. Extraction was still handled successfully in 0.3.0 and earlier, but broke in 0.4.0. Thanks @mayoff for the fix! - Fixed regression: Swift can optional-demote the `Value` of an embed function `(Value?) -> Root` to `(Value) -> Root` and yield `CasePath<Root, Value>` instead of `CasePath<Root, Value?>. Extraction was still handled successfully in 0.3.0 and earlier, but broke in 0.4.0. Thanks @mayoff for this fix as well! - A family of helpers functions around reflection/runtime extraction has been deprecated, this includes module-level `extract` functions and `CasePath.case`. The library now always favors using case path syntax (`/Root.caseName`) or the `CasePath.extract(from:)` method. - Reflection code is now 100% powered by the runtime metadata and no longer relies on `Mirror`! Thanks @mayoff for diving _even deeper_ into the problem, surfacing with a nice and succinct solution.

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. - Added: Objects that are printed more than once per dump are now identified, _e.g._ `MyObject#1(…)` instead of `MyObject(…)` (thanks @tahirmt). - Fixed: Superclass fields are now correctly dumped alongside fields defined on the subclass (thanks @tahirmt). * Add workaround for Xcode 14.1 beta 1, which fails to build iOS packages that import GameKit. - Added: SwiftWASM support (thanks @inamiy). - Fixed: The swift-custom-dump package no longer publicly vends its internal schemes. - Fixed: diffs now differentiate when value dumps the same output, but the type differs. ```swift let xs: [Any] = [Float(42)] let ys: [Any] = [Double(42)] diff(xs, ys) // [ // - [0]: 42 as Float // + [0]: 42 as Double // ] ``` ## What's Changed * Update README.md by @iampatbrown in https://github.com/pointfreeco/swift-custom-dump/pull/34 * Remove platform requirements by @stephencelis in https://github.com/pointfreeco/swift-custom-dump/pull/37 * Run 13.2 CI on experimental build system by @stephencelis in https://github.com/pointfreeco/swift-custom-dump/pull/38 * Add CI labels by @stephencelis in https://github.com/pointfreeco/swift-custom-dump/pull/39 * Add issue templates and CoC by @stephencelis in https://github.com/pointfreeco/swift-custom-dump/pull/41 * Revert removal of platform requirements by @stephencelis in https://github.com/pointfreeco/swift-custom-dump/pull/42 * Show label for visited items by @brentleyjones in https://github.com/pointfreeco/swift-custom-dump/pull/43 * fix: Resolve Xcode 13 project startup beachball due to this playground by @importRyan in https://github.com/pointfreeco/swift-custom-dump/pull/46 ## New Contributors * @iampatbrown made their first contribution in https://github.com/pointfreeco/swift-custom-dump/pull/34 * @brentleyjones made their first contribution in https://github.com/pointfreeco/swift-custom-dump/pull/43 * @importRyan made their first contribution in https://github.com/pointfreeco/swift-custom-dump/pull/46 **Full Changelog**: https://github.com/pointfreeco/swift-custom-dump/compare/0.3.0...0.3.1 - Fixed: differences that were previously undetected in the dump are now better surfaced, _e.g._ when object identity differs, it will be printed in the diff. - Fixed: Xcode 13.2's new experimental build system is now supported (thanks @jaanussiim). - Fixed: Null pointers no longer cause the dump function to crash, and are instead printed as `(null pointer)` - Updated: Improved dumping of errors bridged to `NSError` (thanks @aroben). - Updated: Inline types no longer include `(unknown context)` in their dumps. - Fixed: Swift 5.5 support on macCatalyst, Linux, and Windows (thanks @klundberg, @davdroman). - Infrastructure: CI improvements, including running on Windows and against more versions of Swift (thanks @Jeehut, @davdroman). - Added: a bunch of new conformances (thanks @noppefoxwolf, @konomae, @Bellaposa). - Fixed: macOS builds on Big Sur using Xcode 13 RC no longer fail. - Added: more conformances for Photos, StoreKit, and UIKit (thanks @larryonoff). - Fixed: a build failure for watchOS in Xcode 13. - Infrastructure: continuous integration now run tests and builds for release on all Apple platforms over a number of recent Xcode versions.

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. - Added: `AnySchedulerOf<UIScheduler>.shared` and `UIScheduler.test` endpoints have been added, making it friendlier to interact with UI schedulers in a testable way (thanks @emixb). - Fixed: The combine-schedulers package no longer publicly vends its internal schemes. - Changed: The `TestScheduler` will now yield more time to the cooperative thread pool to better allow asynchronous work to be scheduled and picked up before assertions are run. * Fixed: Combine Schedulers no longer causes DocC builds to fail in projects that depend on Combine Schedulers. This includes the Composable Architecture and applications/libraries that depend on the Composable Architecture. > **Warning**: This is a breaking change that may affect your application when upgrading. > > Prior to this update, Combine Schedulers automatically imported the Foundation framework alongside itself (due to an `@_exported import Foundation` declaration that was included in a file vendored from an Apple code base). > > Removing this has the following ramifications: > > * If you use [the Composable Architecture](https://github.com/pointfreeco/swift-composable-architecture) and are pinned to a version before [0.39.1](https://github.com/pointfreeco/swift-composable-architecture/releases/0.39.1), you must also explicitly pin Combine Schedulers to 0.7.2 or lower for previous versions of the Composable Architecture to build: > > ```swift > .package(url: "https://github.com/pointfreeco/combine-schedulers", .exact("0.7.2")), > ``` > > ...or you must update to the Composable Architecture 0.39.1 or higher. > > * If your application relied on ambient access to Foundation from importing Combine Schedulers (or the Composable Architecture), you must explicitly `import Foundation` in those files after upgrading, or you must pin Combine Schedulers to 0.7.2 or lower. ## What's Changed * Remove the SchedulerOf typealias. by @mbrandonw in https://github.com/pointfreeco/combine-schedulers/pull/64 This is needed due to the typealias causing Xcode editor stability problems. A Swift bug has been filed [here](https://github.com/apple/swift/issues/60550), and an Xcode feedback has been filed [here](https://gist.github.com/tgrapperon/977f0406f67c417a4730a8640ab50526). **Full Changelog**: https://github.com/pointfreeco/combine-schedulers/compare/0.7.1...0.7.2 - Added: Build support for non-Apple platforms. Allows libraries to depend on `CombineSchedulers` but still build for other platforms, like Linux. - Added: `TestScheduler.advance(to:)`, for advancing to a specific moment in time (thanks @ytyubox). - Added: Async-friendly versions of `TestScheduler`'s `advance(by:)`, `advance(to:)`, and `run()` methods, which yield time to cooperative concurrency while scheduling work. - Added: Async-friendly functions for sleeping and running timers: `Scheduler.sleep(for:)` and `Scheduler.timer(interval:)`. - Added: `AnySchedulerOf<DispatchQueue>.global(qos:)` (thanks @iampatbrown). - Renamed: `FailingScheduler` is now `UnimplementedScheduler`. Type aliases to `FailingScheduler` and `failing` statics are soft-deprecated and will become warnings in a future version of Combine Schedulers. - Infrastructure: README fix (thanks @junpluse). * Updated xctest-dynamic-overlay to 0.2.1 (#42) (thanks @mRs- !) - Fixed: restore platform requirements to platforms with Combine. Due to a possible but in Xcode 13, Combine Schedulers would fail to build in release mode otherwise. - Infrastructure: fixed a README/documentation typo (thanks @kalupas226). - Changed: relaxed platform dependencies to be platform independent (thanks @maximkrouk). - Infrastructure: updated documentation.

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days.

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. This locks Validated to NonEmpty 0.2.1, which fixes a compiler error from previous versions. This release supports Swift 5 and NonEmpty 0.3! This (very) minor update ensures that CocoaPods users can merely `import Validated`. Validated relies on [NonEmpty](https://github.com/pointfreeco/swift-nonempty), which was affected by a regression in the release of Swift 4.2. This version of Validated updates this dependency in order to support Swift 4.2. The first release of Validated: a result type that accumulates multiple errors.

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. - Added: public APIs for `Xoshiro` generator (thanks @broomburgo). - Changed: the package name has changed from `Gen` to `swift-gen` for better SPM compatibility. - Added: an implementation of the [xoshiro256**](http://prng.di.unimi.it) PRNG for seedable randomness. - Changed: the package name has been renamed from `Gen` to `swift-gen` to match the community and other Point-Free projects. - Deprecated: the LCRNG is not stable across Swift versions and thus has been deprecated. It will be obsoleted and removed in future releases. - Bug fixed: the `Float80` generator is now properly scoped to architectures that support it. ### What's new? - Swift 5 support - A generator for the `Result` type Gen now correctly builds for iOS. https://github.com/pointfreeco/swift-gen/pull/6 Introducing Gen, a library for composable, transformable, controllable randomness!

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. - Fixed: The swift-identified-collections package no longer publicly vends its internal schemes. - Added: `IdentifiedArray.append(contentsOf:)`, which appends elements to an array iff they are not already present (thanks @jeffersonsetiawan). - Updated: `swift-collections` dependency has been updated to 1.0.2 (thanks @KaiOelfke). - Updated: Identified Collections platform requirements have been relaxed. - Improved: performance of `IdentifiedArray.==` (thanks @mihaho). - Fixed: Identified arrays that specify a key path should no longer fail equatable checks when elements are equal (thanks @sroche27r). - Changed: Swift Identified Collections now depends on [Swift Collections 1.0.0](https://github.com/apple/swift-collections/releases/tag/1.0.0). - Added: `IdentifiedArray.elements`, an array view of the identified array (thanks @p4checo). - Changed: `IdentifiedArray.move(fromOffsets:toOffset:)` and `IdentifiedArray.remove(atOffsets:)` no longer requires SwiftUI (thanks @gonzalolarralde). - Improved: `IdentifiedArray.move(fromOffsets:toOffset:)` and `IdentifiedArray.remove(atOffsets:)` are now much more performant. - Fixed: `IdentifiedArray.partition` now partitioning keys along with their values by calling down to the correct underlying implementation (thanks @tgt). - Fixed: stopped using a few APIs from Swift Collections that have been deprecated (thanks @diederich). - Fixed: `removeLast(n:)` renamed to `removeLast(_:)` (thanks @konomae). - Fixed: dependence on Swift Collections relaxed to 0.0.1. - Initial release! What are the main differences between this library and the `IdentifiedArray` [that shipped](https://github.com/pointfreeco/swift-composable-architecture/blob/0.20.0/Sources/ComposableArchitecture/SwiftUI/IdentifiedArray.swift) in [swift-composable-architecture 0.20.0](https://github.com/pointfreeco/swift-composable-architecture/releases/tag/0.20.0)? - It is now a wrapper around [`OrderedDictionary`](https://github.com/apple/swift-collections/blob/main/Documentation/OrderedDictionary.md) from [Swift Collections](https://github.com/apple/swift-collections). This means we can lean on the performance and correctness of Apple's library and avoid many of the bugs discovered in the original implementation. - It no longer conforms to `MutableCollection` and `RangeReplaceableCollection`. We take the same position Swift Collections [takes](https://github.com/apple/swift-collections/blob/main/Documentation/OrderedSet.md#sequence-and-collection-operations) with `OrderedSet` and `OrderedDictionary`. All elements in an identified array should have unique identity, operations such as `MutableCollection`'s subscript setter or `RangeReplaceableCollection`'s `replaceSubrange` assume the ability to insert/replace arbitrary elements in the collection, but allowing that could lead to duplicate values. Methods have been deprecated or made unavailable with suggested migration paths.

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. * Fixed: [A regression](https://github.com/apple/swift/issues/61192) introduced to Swift in Xcode 14.1 broke compilation. This release includes a workaround to get things building again. - Added: `Fragment` parser, for parsing a URL fragment (thanks @ryanbooker). - Added: `Host` parser, for parsing a URL host (thanks @ryanbooker). - Changed: URL Routing's platform requirements have been bumped to match Parsing's requirements, equivalent to SwiftUI (iOS 13+, macOS 10.15+, tvOS 13+, watchOS 6+). If these minimum requirements don't fit your needs, [let us know](https://github.com/pointfreeco/swift-url-routing/discussions/new). - Infrastructure: Fixed documentation typos (thanks @fonkadelic, @volkdmitri). - Added: `URLRoutingClient` can now be configured with a JSONDecoder for global response decoding (thanks @jtouzy). - Added: a `Scheme` router (thanks @ryanbooker). - Added: `URLRoutingClient.data(for:)`. - Updated: `URLRoutingClient.request(_:as:decoder:)` has been renamed to `URLRoutingClient.decodedResponse(for:as:decoder:)`. - Changed: the package name has been changed to `swift-url-routing` to match its repo name. - Optimized: added inlining to the path component router. - Infrastructure: documentation updates. - Infrastructure: added SPI badges to README (thanks @finestructure). Our first release!

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. This is the first release of `generate-enum-properties`, which makes working with Swift enums like working with Swift structs!

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. - Fixed: XCTestDynamicOverlay now builds for SwiftWASM environments (thanks @inamiy). - Changed: `XCTUnimplemented` is now overloaded to take a non-optional `placeholder`, or nothing at all, to prevent ambiguity in situations where you pass a `.none` value. - Fixed: Linux now builds correctly when statically linking against the standard library. - Infrastructure: Cleaned up DocC. - Changed: `XCTUnimplemented` errors now print more context, including the values passed to the function. - Fixed: the underscored `_XCTIsTesting` property no longer returns `false` when tests run in app hosts. - Changed: `XCTUnimplemented` will no longer fatal error when range replaceable values are to be returned. Instead, it will return an empty value. - Changed: `XCTUnimplemented` overloads have been reduced, where `Void`-returning logic is built into the generic `Result`-returning functions. - Fixed: 0.3.0 introduced a regression in which calling `XCTFail` outside of tests from a debug build could cause a recursive loop. - Infrastructure: DocC documentation has been added. - Added `XCTUnimplemented`, a function for constructing "unimplemented" closures that call `XCTFail` when invoked. Helpful for stubbing dependencies held in value types. For example, a common use of XCTest Dynamic Overlay: ```swift struct FactClient { var fetch: (Int) async throws -> String } extension FactClient { struct Unimplemented: Error { let endpoint: String } static let unimplemented = Self( fetch: { _ in XCTFail("Unimplemented: \(Self.self).fetch") throw Unimplemented(endpoint: "\(Self.self).fetch") } ) } ``` Can be simplified using `XCTUnimplemented`: ```swift extension FactClient { static let unimplemented = Self( fetch: XCTUnimplemented("\(Self.self).fetch") ) } ``` - Changed: `XCTFail` now causes a runtime warning when called outside of tests. - Changed: `XCTFail` functions are now no-ops in non-DEBUG builds to avoid leaking symbols into RELEASE builds. - Changed: `XCTFail` is no longer restricted to DEBUG builds. - Includes MIT LICENSE. * Initial release!

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. - Updated: Vapor 4 support (thanks @JaapWijnen) - Updated: swift-html 0.4.0 support Synchronizes the Swift 4.2-compatible release of swift-html 0.3.1. This release brings in support for Html 0.3.0 and Swift 5!

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. - Fixed: Errors thrown in the handler should now correctly propagate in Vapor applications as expected. - Infrastructure: README updates and fixes (thanks @gohanlon, @finestructure) Our first release!

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. > Note: This release is breaking! The APIs have changed significantly. - Breaking change: Clients have been simplified so that endpoints no longer require ids. This means the client lifecycle is now internal to the live instance and managing explicitly via `create` and `destroy` is no longer necessary. As such, `create` was also renamed to `delegate`. - Changed: Pinned to TCA 0.27.1 (for failing effects) (thanks @andreyz). - Added: Monitor significant changes API (thanks @joeblau). - Added: `requestTemporaryFullAccuracyAuthorization` (thanks @tomhut). - Fixed: Publicized initializers for `Region` (thanks @ferologics). - Fixed: Preserve `Visit`'s `CLVisit` raw value (thanks @andreyz). - Fixed: `requestLocationWhileInUse` is now available on macOS (thanks @fire-at-will). - Infrastructure: README fixes (thanks @hanneskaeufler). First release since being extracted from https://github.com/pointfreeco/swift-composable-architecture - Added: support for headings APIs (thanks @dannyhertz). - Added: support for accuracy authorization (thanks @mackoj).

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. - Fixed: removed broken protocol conformance (thanks @andreyz). - Fixed: corrected README typo (thanks @barksten). First release since being extracted from https://github.com/pointfreeco/swift-composable-architecture - Added: support for headphone motion manager.

The markdown parsing is broken/disabled for release notes. Sorry about that, I'm chasing the source of a crash that's been bringing this website down for the last couple of days. This release brings in support for Html 0.3.0 and Swift 5!