v1.0.0

What's Changed

• Clarify that .clear() works on arrays by @jviotti in #11
• Implement make_object and make_array constructors by @jviotti in #12
• Document how to convert existing instances to arrays/objects by @jviotti in #14
• Note functions that are useful for programatically creating containers by @jviotti in #15
• Test programatically creating objects and arrays by @jviotti in #13
• Implement prettify contrib/ program by @jviotti in #16
• Implement a contrib/ minify program by @jviotti in #17
• Do not build contrib/ programs by default by @jviotti in #19
• Add FetchContent usage instructions by @jviotti in #20
• Do not build tests by default by @jviotti in #21
• Reduce required C++ version to 17 by @jviotti in #23
• Improve contribution instructions by @jviotti in #26
• Vendor bootstrap by @jviotti in #25
• Minor fixes to support Windows by @jviotti in #28
• Support a custom backend path by @jviotti in #29
• Do not run stdin contrib tests on Windows by @jviotti in #30
• Extract contrib test scripts into variables by @jviotti in #31
• Specify --build-config when running ctest by @jviotti in #32
• Fix Windows stdin checks by @jviotti in #33
• Support finding rapidjson from vcpkg by @jviotti in #34
• Fix crash on .assign() with move semantics by @jviotti in #35
• Namespace JSON tests by @jviotti in #36
• Implement a metaschema() function by @jviotti in #37
• Refactor metaschema() to solve one-definition-rule problem by @jviotti in #38
• Specify --config when building using CMake by @jviotti in #39
• Simplify JSON unit tests by @jviotti in #40
• Upgrade vendorpull by @jviotti in #41
• Implement a JSON Schema id() function by @jviotti in #42
• Implement a function to return schema vocabularies by @jviotti in #43
• Ignore vendor/ in .ackrc by @jviotti in #44
• Implement a function to deep-copy a JSON document or value by @jviotti in #45
• Support from() when passing a zero literal by @jviotti in #46
• Add easy-to-use range-based for loop support by @jviotti in #50
• Implement support for iterating over subschemas by @jviotti in #51
• Increase granularity of GNU Make targets by @jviotti in #52
• Implement a default JSON Schema 2020-12 schema walker by @jviotti in #53
• Let vocabularies() take an optional default metaschema URI by @jviotti in #56
• Check vocabularies at every step in subschema_iterator() by @jviotti in #55
• Implement a flat subschema iterator by @jviotti in #57
• Change domain to jsontoolkit.org by @jviotti in #60
• Make JSON Toolkit build on Debian 11 (current Stable) by @jviotti in #58
• Enable pthreads on Linux by @jviotti in #61
• Test building on GNU/Linux using GCC by @jviotti in #59
• Add a contrib tool to analyze JSON Schema keywords by @jviotti in #47
• Implement a function to determine the schema's dialect URI by @jviotti in #62
• Assume core vocabularies if $vocabulary is unset by @jviotti in #63
• Take small enough std::uint64_t as input in from() by @jviotti in #64
• Fix flat_subschema_iterator() on schema without a metaschema by @jviotti in #65
• Define a schema walker that never walks by @jviotti in #66
• Implement erase_many() by @jviotti in #67
• Document GCC support by @jviotti in #68
• Minor misc fixes for erase_many() by @jviotti in #69
• Define defines_any() by @jviotti in #70
• Fix push_back() docs example by @jviotti in #71
• Implement is_number() by @jviotti in #72
• Support using from() with std::size_t by @jviotti in #73
• Implement an "add" function to increment/decrement numbers in-place by @jviotti in #74
• Document the contains() array function by @jviotti in #75
• Implement a "compare" function for standard algorithms by @jviotti in #76
• Generalize iterator-based erase_many to work on arrays by @jviotti in #77
• Implement a clear_except() function for objects by @jviotti in #78
• Make sure schema_walker_strategy_t is defined outside of GCC by @jviotti in #80
• Build website using CMake by @jviotti in #81
• Test CMake installation on CI by @jviotti in #82
• Make Makefile NMAKE compliant by @jviotti in #83
• Disable some JSONTestSuite tests for MSVC by @jviotti in #85
• Fix cast to make MSVC happy by @jviotti in #84
• Add basic Windows support by @jviotti in #86
• Create a new "test" target in the Makefile by @jviotti in #87
• Get rid of fail.sh contrib test helper script by @jviotti in #88
• Do not run tests twice on the Makefile by @jviotti in #89
• Bring back "contrib" tests on Windows by @jviotti in #27
• Run clang_format_test, not clang_format on CI by @jviotti in #92
• Implement a sample contrib cURL resolver by @jviotti in #91
• Disable min/max macros on MSVC by @jviotti in #94
• Implement a local-first resolver for 2020-12 official schemas by @jviotti in #93
• Split JSON and JSON Schema into its own modules within src/ by @jviotti in #95
• Merge the local resolver with the core JSON Schema module by @jviotti in #96
• Support JSON Schema 2019-09 on the default resolver by @jviotti in #97
• Throw if the core vocabulary is missing or set to optional by @jviotti in #98
• Better handle vocabularies for boolean schemas by @jviotti in #99
• Check current dialect when fetching vocabularies by @jviotti in #100
• Make vocabularies() dialect-aware by @jviotti in #101
• Add 2019-09 support in vocabularies() by @jviotti in #102
• Support 2020-12 and 2019-09 Hyper Schema by @jviotti in #104
• Note intention to support older JSON Schema dialects in .vocabularies() by @jviotti in #105
• Implement a copy() function for JSON containers by @jviotti in #106
• Add a copy() overload to duplicate entire containers by @jviotti in #107
• Ensure doubles are never NaN or infinity by @jviotti in #108
• Re-license the project as AGPL by @jviotti in #109
• Implement a to_vector() utility function by @jviotti in #110
• Implement an is_positive() utility function by @jviotti in #111
• Be explicit about JSON iterator begin/end types by @jviotti in #112
• Add iterator traits to JSON iterators by @jviotti in #113
• Declare iterator and const_iterator for SchemaWalker by @jviotti in #114
• Run CMake parallel builds by default by @jviotti in #115
• Handle "None" schema walker strategy on the switch case by @jviotti in #117
• Run tests in parallel by @jviotti in #118
• Properly allocate keys on object assignment by @jviotti in #119
• Implement a hash function for JSON documents by @jviotti in #120
• Rename "dialect()" function to "draft()" by @jviotti in #121
• Integrate with Cirrus CI for BSD runs by @jviotti in #122
• Mention FreeBSD support by @jviotti in #123
• Improve compiler warnings by @jviotti in #124
• Install README on the package docs directory by @jviotti in #125
• Define library aliases by @jviotti in #126
• Always perform component-based installation by @jviotti in #127
• Allow installing contrib programs by @jviotti in #128
• Use :: namespaced target aliases by @jviotti in #129
• Correctly define header sources in JSON interface library by @jviotti in #130
• Set correct output/export names for libraries by @jviotti in #131
• Only install README on top-level builds by @jviotti in #133
• Remove warnings macro in favor of global options by @jviotti in #134
• Ignore -Wdocumentation for RapidJSON by @jviotti in #135
• Exercise installation on local builds by @jviotti in #136
• Fix JSONTOOLKIT_BACKEND_PATH empty conditional logic by @jviotti in #137
• Declare top-level JSON header as part of the JSON library by @jviotti in #138
• Install library CMake exports by @jviotti in #139
• Turn rapidjson wrapper into an imported target by @jviotti in #140
• Make the project consumable by CMake find_package() by @jviotti in #132
• Better integrate with VCPKG on Windows by @jviotti in #141
• Add a Brewfile for macOS dependencies by @jviotti in #142
• Remove artificial JSON Schema resolver.h install by @jviotti in #144
• Implement a basic ParseError exception subclass by @jviotti in #145
• Introduce a general purpose SchemaError exception subclass by @jviotti in #146
• Implement a ResolutionError exception subclass by @jviotti in #147
• Support calling from() with a std::string_view by @jviotti in #148
• Add a Sourcemeta namespaces to libraries by @jviotti in #149
• Update #if guards to use SOURCEMETA_ by @jviotti in #150
• Re-organize installed CMake files by @jviotti in #151
• Import the Noa project by @jviotti in #152
• Use Noa's compiler variables by @jviotti in #153
• Simplify ISO C++ settings with Noa by @jviotti in #154
• Integrate with ClangTidy by @jviotti in #155
• Upgrade Noa to handle RPATHs by @jviotti in #156
• Address ClangTidy comments by @jviotti in #157
• Add a CHANGELOG template by @jviotti in #159
• Use underscores in contrib apps by @jviotti in #160
• Use underscores in library output names by @jviotti in #161
• Support shared libraries by @jviotti in #158
• Refactor JSON Schema library declaration by @jviotti in #162
• Rename main component to sourcemeta_jsontoolkit by @jviotti in #164
• Separate contrib tools into their own component by @jviotti in #165
• Create a separate sourcemeta_jsontoolkit_dev component by @jviotti in #166
• Set HOMEPAGE_URL in the CMake project declaration by @jviotti in #167
• Improve conditional with ClangTidy suggestions by @jviotti in #168
• Quiet argv pointer ClangTidy warnings until we are in C++20 by @jviotti in #169
• Quiet ref member ClangTidy warnings by @jviotti in #170
• Set JSONTOOLKIT_BACKEND_PATH to be of type PATH by @jviotti in #171
• Be explicit about CMake variable types by @jviotti in #172
• Improve "backend" option definition by @jviotti in #173
• Improve GoogleTest discovery defaults by @jviotti in #176
• Use typed CMake variables in packaging tests by @jviotti in #174
• Configure IDE folders in CMake by @jviotti in #177
• Do not hardcode warnings as errors by @jviotti in #178
• Move config.cmake.in to the root of the project by @jviotti in #180
• Remove support for specifying a backend path by @jviotti in #179
• Build against an address sanitizer on Linux by @jviotti in #181
• Fix several shared library generation warnings on Windows by @jviotti in #186
• Test including JSON Schema component in packaging tests by @jviotti in #185
• Rename export header by @jviotti in #188
• Unit test compare() with equal strings by @jviotti in #189
• Get rid of RapidJSON by @jviotti in #190
• Apply sanitizers per target by @jviotti in #193
• Use std::map instead of std::unordered_map in JSON Schema by @jviotti in #196
• Allow ParseError to be exported on Windows by @jviotti in #195
• Enable remaining JSONTestSuite cases by @jviotti in #197
• Use trailing return types by @jviotti in #198
• Add various grammar comments by @jviotti in #199
• Override operator<< for JSON by @jviotti in #200
• Vendor GoogleTest by @jviotti in #202
• Improve contrib target names by @jviotti in #203
• Fix various MSVC DLL warnings by @jviotti in #204
• Run with an undefined behavior sanitizer by @jviotti in #194
• Recommend C++20 on installation examples by @jviotti in #205
• Support initializing from string_view by @jviotti in #206
• Extract test scripts out of test/contrib by @jviotti in #208
• Support Emscripten by @jviotti in #207
• Remove obsolete Makefile ctest comment by @jviotti in #209
• Implement basic JSON Pointer support by @jviotti in #143
• Simplify JSON Pointer "get" branch by @jviotti in #210
• Test a pointer with a stringified index by @jviotti in #214
• Cover JSON Pointer RFC tests by @jviotti in #213
• Improve JSON Pointer "get" inline docs by @jviotti in #212
• Test a JSON string with an null byte inside by @jviotti in #215
• Test presence of null in an object property by @jviotti in #211
• Implement a "set" operator for JSON Pointer by @jviotti in #216
• Test JSON Pointer setting to a hyphen property by @jviotti in #217
• Support the concept of a JSON Pointer hyphen constant by @jviotti in #218
• Use iterators to traverse JSON Pointers by @jviotti in #219
• Declare JSON char and char_traits types by @jviotti in #220
• Implement a .back() method for JSON Pointer by @jviotti in #221
• Support setting to the hyphen JSON Pointer token in arrays by @jviotti in #222
• Add support for stringifying JSON Pointers by @jviotti in #223
• Correctly stringify control characters in JSON by @jviotti in #224
• Fix JSON Pointer stringify Doxygen group by @jviotti in #225
• No Emscripten on default Make builds by @jviotti in #226
• Rename parse to parse_json by @jviotti in #227
• Implement emplace_back for pointers by @jviotti in #229
• Do not include generated export file on Unikraft by @jviotti in #230
• Support C++17 in the JSON component by @jviotti in #231
• Add basic Unikraft support by @jviotti in #232
• Properly support C++20 in Unikraft by @jviotti in #234
• Allow object properties to be queries using integer indexes by @jviotti in #235
• Simplify syntax for accessing JSON Pointer iterators by @jviotti in #237
• Test pointers that traverse objects using integer indexes by @jviotti in #236
• Avoid exporting templated classes by @jviotti in #238
• Make Doxygen optional by @jviotti in #240
• Fix MSVC out of range debug warning when parsing unicode symbols by @jviotti in #239
• Match CMAKE_BUILD_TYPE on packaging tests by @jviotti in #241
• Upgrade Noa to get Windows library/runtime directory fixes by @jviotti in #242
• Test generation of shared libraries on Windows by @jviotti in #163
• Fix Traits::eof() on JSON parser in older GCC versions by @jviotti in #243
• Rename parse_json to parse by @jviotti in #244
• Don't test emscripten support in this project by @jviotti in #245
• Rename JSON parser test file to just "parse" by @jviotti in #247
• Implement a to_stringstream() method by @jviotti in #248
• Implement a PointerParseError class by @jviotti in #246
• Make JSON Pointer grammar header private by @jviotti in #249
• Add JSON Pointer parsing support by @jviotti in #228
• Refactor and simplify schema resolver functionality by @jviotti in #250
• Add Draft7 metaschemas by @jviotti in #251
• Include main JSON Schema metaschemas from all drafts by @jviotti in #252
• Add all missing meta-schemas to default resolver by @jviotti in #253
• Make schema resolver take a string view as identifier by @jviotti in #254
• Remove jsonschema_default_walker.h header by @jviotti in #255
• Simplify SchemaWalker class by @jviotti in #256
• Completely redo the schema walker interface by @jviotti in #257
• Test presence of all official metaschemas in default resolver by @jviotti in #258
• Rename draft() to dialect() by @jviotti in #259
• Simplify official schema resolver by @jviotti in #260
• Use string views in SchemaWalker by @jviotti in #261
• Fix 2019-09 "links" metaschema in default resolver by @jviotti in #263
• Rename dialect to "base_dialect" by @jviotti in #262
• Rename metaschema() to dialect() by @jviotti in #264
• Rename mentions of default metaschema to default dialect by @jviotti in #265
• Revise dialect() function to support all versions of JSON Schema by @jviotti in #266
• Rename ResolutionError to SchemaResolutionError by @jviotti in #267
• Split base dialect tests by dialect by @jviotti in #269
• Revisit official resolver tests by @jviotti in #270
• Split dialect tests by dialect by @jviotti in #271
• Set an explicit parallel build limit by @jviotti in #272
• Re-implement the vocabularies() function by @jviotti in #268
• Extract metaschemas file read into a separate file by @jviotti in #273
• Inline schema_walker_none by @jviotti in #275
• Rename default_resolver.cc to official_resolver.cc by @jviotti in #274
• Use C++20 map contains in default walker by @jviotti in #276
• Separately and properly test schema_walker_none by @jviotti in #277
• Revise the default walker tests by @jviotti in #278
• Revisit default walker 2020-12 tests by @jviotti in #279
• Support 2019-09 in the default walker by @jviotti in #280
• Add Draft7 support to the default walker by @jviotti in #281
• Refactor the definition of the default walker by @jviotti in #282
• Don't run every GoogleTest case as a separate process by @jviotti in #283
• Add Draft6 support to the default walker by @jviotti in #284
• Add Draft4 support to the default walker by @jviotti in #285
• Add Draft3 support to default walker by @jviotti in #286
• Add Draft2 support to the default walker by @jviotti in #287
• Add Draft1 support to the default walker by @jviotti in #288
• Add Draft0 support to the default walker by @jviotti in #289
• Add hyper-schema walker support to 2019-09 and 2020-12 by @jviotti in #290
• Test hyperschema support for Draft0 in default walker by @jviotti in #291
• Test hyperschema support for Draft1 in default walker by @jviotti in #292
• Add hyperschema support for Draft2 in default walker by @jviotti in #293
• Add hyperschema support for Draft3 and Draft4 in default walker by @jviotti in #294
• Add hyperschema support to default walker for Draft6 and Draft7 by @jviotti in #295
• Make id() taker a resolver and a default dialect by @jviotti in #297
• Return a string_view when fetching the id of a SchemaResolutionError by @jviotti in #296
• Properly implement the id based on dialects by @jviotti in #298
• Properly test id() across dialects by @jviotti in #299
• Remove unused VCPKG integration file by @jviotti in #301
• Implement a basic URI class based on uriparser by @jviotti in #300
• Document the new URI module by @jviotti in #302
• Implement a fragment method on the URI class by @jviotti in #303
• Implement URI escaping/unescaping by @jviotti in #304
• Add TODO for constructing URIs using input streams by @jviotti in #305
• Implement JSON Pointer equality support by @jviotti in #307
• Include JSON Pointer before JSON Schema by @jviotti in #308
• Report back only JSON Pointers when walking schemas by @jviotti in #310
• Take a default value in the dialect() function by @jviotti in #311
• Add basic copy/move semantics to URI by @jviotti in #312
• Add a pop_back() method to Pointer by @jviotti in #313
• Overload the < operator for Pointer by @jviotti in #314
• Implement an anchors() function similar to id() by @jviotti in #315
• Make Pointer's pop_back() throw on empty pointers by @jviotti in #316
• Implement a from_fragment URI static factory method by @jviotti in #317
• Support JSON Pointer and URI on Unikraft by @jviotti in #319
• Document how to use JSON Toolkit with Unikraft by @jviotti in #320
• Extract official_resolver interpolation to a separate script by @jviotti in #321
• Setup a Unikraft project for running JSON Toolkit tests by @jviotti in #322
• Support JSON Schema on Unikraft by @jviotti in #323
• Implement a subpointer walker by @jviotti in #324
• Fix Draft0, Draft1, and Draft2 hyper schema walkers by @jviotti in #325
• Implement JSON Schema framing by @jviotti in #306
• Support JSON Schema framing when mixing dialects by @jviotti in #326
• Revise frame() to ignore dynamic anchors for now by @jviotti in #327
• Use -latest runners on GitHub Actions by @jviotti in #329
• Report dialect URI on frame() by @jviotti in #328
• Store the top-level schema id on frame() entries by @jviotti in #330
• Fix documentation example for frame() by @jviotti in #331
• Ensure uriparser doesn't affect BUILD_SHARED_LIBS by @jviotti in #332
• Attempt to workaround Clang Ubuntu issues on GitHub Actions by @jviotti in #338
• Fix JSON Pointer include guards by @jviotti in #335
• Fix README link by @jviotti in #337
• Run Unikraft build GitHub Actions by @jviotti in #318
• Run Unikraft build using QEMU on CI by @jviotti in #340
• Don't run kraft fetch on the Makefile for local development by @jviotti in #341
• Build packaging example with C++20 by @jviotti in #345
• Fix ctest comment on GitHub Actions by @jviotti in #344
• Add Doxygen to the Brewfile by @jviotti in #343
• Run Unikraft on GitHub Actions by disabling acceleration by @jviotti in #342
• Use Doxygen integration from Noa by @jviotti in #346
• Run apt-get update on GitHub Actions by @jviotti in #347
• Extend Config.uk to list all requirements by @jviotti in #348
• Run apt-get update on Unikraft pipelines by @jviotti in #350
• Add a Patron license by @jviotti in #349
• Test that backspace is not whitespace according to JSON by @jviotti in #351
• Extend the URI class to get the scheme by @jviotti in #352
• Extend the URI class to get port numbers by @jviotti in #353
• Implement a convenience function to parse JSON from a file by @jviotti in #355
• Gracefully take boolean schemas in anchors() by @jviotti in #356
• Use noa_library to define libraries by @jviotti in #357
• Let the id() function take a default id by @jviotti in #359
• Accept an optional default id in frame() by @jviotti in #358
• Optionally take read/write line/column on JSON parsing by @jviotti in #360
• Include JSON Pointer in CMake config much earlier on by @jviotti in #361
• Implement JSONL parsing support by @jviotti in #334
• Vendor the 2020-12 tests of the JSON Schema Referencing Suite by @jviotti in #362
• Test parsing a URI that consists of an invalid fragment by @jviotti in #364
• Run test workflow once every working day by @jviotti in #363
• Implement a method to get a URI path by @jviotti in #366
• Add a method to check whether a URI is a URN by @jviotti in #367
• Test current URI methods against URNs by @jviotti in #365
• Test URNs that contain fragments by @jviotti in #368
• Add tests exercising URIs with empty fragments by @jviotti in #369
• Add test exercising URI recomposing with trailing slash path by @jviotti in #370
• Simplify id() ternary logic in frame() by @jviotti in #373
• Turn ReferenceFrame into a proper class by @jviotti in #372
• Fix frame() Doxygen example by @jviotti in #375
• Exercise frame() when passing an explicit equal default id by @jviotti in #377
• Add a constant that represents the empty pointer by @jviotti in #374
• Extend .frame() to work on multiple URI bases by @jviotti in #376
• Clarify that in frame(), found identifiers only may be relative by @jviotti in #378
• Handle framing with explicit different IDs by @jviotti in #371
• Clarify that URI::recompose doesn't really normalize by @jviotti in #379
• Implement a basic method to get the query part of a URI by @jviotti in #380
• Implement a simple URI canonicalize method by @jviotti in #381
• Implement support for tag URIs by @jviotti in #382
• Canonicalize when storing URIs to reference frames by @jviotti in #383
• Add basic URI iterator support to ReferenceFrame by @jviotti in #384
• Rename JSON Pointer walker files to "subpointer walker" by @jviotti in #385
• Implement a general-purpose JSON walker with JSON Pointer by @jviotti in #386
• Add missing URI tests to Unikraft test runner app by @jviotti in #387
• Implement JSON Pointer relative resolution by @jviotti in #388
• Perform URI canonicalization on ReferenceFrame::defines by @jviotti in #390
• Refactor frame() find bases internal function by @jviotti in #391
• Add an assertion to confirm there is one base dialect only in frame() by @jviotti in #392
• Implement a convenient function to convert a Pointer to a string by @jviotti in #393
• Introduce an EXPECT_FRAME test macro by @jviotti in #394
• Extend frame() to consider every JSON Pointer to the schema by @jviotti in #389
• Upgrade the JSON Schema Referencing Suite to latest by @jviotti in #395
• Rename base() to root() in ReferenceFrame by @jviotti in #396
• Add a base to every ReferenceFrame entry by @jviotti in #397
• Implement a function to convert a JSON Pointer to a URI fragment by @jviotti in #398
• Add URI canonicalize examples around URI escaping by @jviotti in #399
• Integrate with the JSON Schema referencing suite by @jviotti in #354
• Run the JSON Schema Referencing Suite for 2019-09 by @jviotti in #400
• Run the JSON Schema Referencing Suite for Draft7 by @jviotti in #401
• Run the JSON Schema Referencing Suite for Draft6 by @jviotti in #402
• Run the JSON Schema Referencing Suite on Draft 4 by @jviotti in #403
• Run most of the JSON Schema Referencing Suite on Draft 3 by @jviotti in #404
• Enable every test from the JSON Schema Referencing Suite by @jviotti in #405
• Add a method to concatenate two JSON Pointers by @jviotti in #407
• Implement schema transformation functionality by @jviotti in #406
• Make schema transformation rules report the values they modified by @jviotti in #408
• Fix rule/bundle examples by @jviotti in #409
• Revert schema transform rules returning set of pointers by @jviotti in #410
• Introduce a schema transformer proxy for transform rules by @jviotti in #411
• Add method to check whether a Pointer starts with another Pointer by @jviotti in #412
• Implement JSON Pointer rebasing by @jviotti in #413
• Test URI resolution with fragments by @jviotti in #414
• Let JSON Pointer instances be constructed from standard C++ strings by @jviotti in #415
• Support framing references in anonymous schemas by @jviotti in #416
• Add TODO about anchors and anonymous frames by @jviotti in #417
• Implement an example "frame" JSON Schema CLI tool by @jviotti in #418
• Cleanup Frame CLI output by @jviotti in #419
• Add CMake option to control component installation by @jviotti in #420
• Extend framing to support pointers without a base by @jviotti in #421
• Support framing pointers to $anchor in anonymous schemas by @jviotti in #423
• Frame static anchors on anonymous schemas by @jviotti in #424
• Extend .frame() to receive ReferenceMap argument by @jviotti in #425
• Revise .frame() reference map basic tests by @jviotti in #426
• Remove incorrect TODOs by @jviotti in #427
• Extend framing to resolve static references by @jviotti in #422
• Print references in frame CLI tool by @jviotti in #428
• Document frame references by @jviotti in #429
• Register top-level static anchors without identifiers by @jviotti in #431
• Extend framing to record dynamic anchors by @jviotti in #430
• Dereference after framing by @jviotti in #432
• Extract ReferenceFrame into a separate unit by @jviotti in #433
• Remove invalid TODO in frame() by @jviotti in #434
• Revert changes that frame dynamic anchors by @jviotti in #435
• Don't record plain reference frames on schemas with ids by @jviotti in #436
• Explicitly include CompilerOptions by @jviotti in #437
• Upgrade Noa to 0862a96772362a6997ce2f5fda877b182a7b6f71 by @jviotti in #438
• Simplify ReferenceMap to only consider static references by @jviotti in #439
• Avoid passing a separate ReferenceFrame for dynamic frames by @jviotti in #440
• Add a ReferenceType qualifier to ReferenceMap by @jviotti in #441
• Implement $dynamicRef support in frame() by @jviotti in #442
• Implement a method to recompose a URI without its fragment by @jviotti in #443
• Make recompose_without_fragment() return an optional value by @jviotti in #445
• Add TODOs to URI about the benefits of supporting mutations by @jviotti in #446
• Add TODO for checking that static references exists in frames by @jviotti in #447
• Extend ReferenceMap to split URI by its potential fragment by @jviotti in #444
• Extract ReferenceType to its own header by @jviotti in #448
• Simplify TO_POINTER macro in JSON Schema test utils by @jviotti in #450
• Revise GitHub Actions Linux Clang CI workaround by @jviotti in #452
• Check for 2020-12 Core vocabulary before framing $dynamicRef by @jviotti in #453
• Store the reference type in ReferenceFrame by @jviotti in #449
• Do not frame relative anchor URIs for schemas with identifiers by @jviotti in #455
• Update frame() description docs by @jviotti in #456
• Set ReferenceType as part of the ReferenceFrame key by @jviotti in #457
• Correctly detect anchors that are both static and dynamic by @jviotti in #458
• Add an overload for anchors() that directly take vocabularies by @jviotti in #459
• Frame $dynamicAnchor for 2020-12 by @jviotti in #454
• Implement a convenience base dialect shortcut for id() by @jviotti in #460
• Add comment about extending SchemaIterator to return more information by @jviotti in #461
• Add a helper for checking if $ref overrides adjacent keywords by @jviotti in #462
• Upgrade KraftKit to v0.7.2 by @jviotti in #463
• Optimize base_dialect() to require less JSON parsing by @jviotti in #465
• Add performance-friendly vocabularies() overloads by @jviotti in #466
• Optimize vocabularies() for 2020-12 and 2019-09 by @jviotti in #467
• Extract URI parser into a separate function by @jviotti in #468
• Perform URI canonicalization in-place by @jviotti in #469
• Simplify ReferenceFrame::defines to not canonicalize by @jviotti in #472
• Get rid of ReferenceFrame wrapper class in favor of an alias by @jviotti in #471
• Simplify the ReferenceMap helper type by @jviotti in #473
• Turn URI::resolve_from into a method that mutates the instance by @jviotti in #470
• Upgrade Vendorpull by @jviotti in #474
• Upgrade Noa to latest by @jviotti in #475
• Add a default Vendorpull mask by @jviotti in #476
• Perform a single pass over subschemas in frame() by @jviotti in #477
• Test framing remote references by @jviotti in #479
• Extend SchemaIterator to pass metadata about each schema by @jviotti in #478
• Implement an assign_if_missing JSON helper method by @jviotti in #481
• Implement JSON Schema bundling by @jviotti in #480
• Extend SchemaTransformer to support the JSON BinPack Mapper by @jviotti in #482
• Generalize transformer .at() into a schema getter by @jviotti in #484
• Improve docs on how to define schema transformation methods by @jviotti in #485
• Generalize transformer into_object() as replace() by @jviotti in #486
• Generalize transformer .replace() to take pointers by @jviotti in #487
• Generalize transformer .assign() to take base pointers by @jviotti in #488
• Generalize transformer erase to take a pointer path by @jviotti in #489
• Fix JSON erase iterator docs by @jviotti in #490
• Implement erase_keys at SchemaTransformer by @jviotti in #491
• Revise transformer erase_keys to take iterators by @jviotti in #492
• De-duplicate noa_sanitizer commands by @jviotti in #493
• Generate CMake config even on non-top level builds by @jviotti in #494
• Support component-based find_package by @jviotti in #495
• Get rid of Cirrus CI by @jviotti in #496
• Guard package find scripts by @jviotti in #497
• Fix GCC taking initializer_list precedence over copy-constructor by @jviotti in #498
• Upgrade Noa to latest by @jviotti in #499
• Simplify test Kraftfile by @jviotti in #500
• Upgrade Noa to the latest version by @jviotti in #502
• Implement a JSON method to estimate memory usage by @jviotti in #501

Full Changelog: https://github.com/sourcemeta/jsontoolkit/commits/v1.0.0