# Changelog # 0.19.1 26 May 2017 * fix(typescript): capture static members correctly 0e41fae7 # 0.19.0 3 May 2017 - feat(typescript): upgrade to TypeScript 2.2.1 0f48e866 ## Breaking Changes The upgrade to TypeScript v2 may well break projects that are generating docs based on a TypeScript v1 application. # 0.18.0 30 April 2017 - fix(typescript): remove `index` from the end of module ids and names ([e6109f57](https://github.com/angular/dgeni-packages/commit/e6109f57266df1edde37071e985d58cda76c1bf1)) ## Breaking Changes This has a possible breaking change if users were depending upon a module at `/some/folder/index.ts` having a name of `index` and an id of `some/folder/index`. # 0.17.2 26 April 2017 ## Bug Fixes - **jsdoc:** reset backTickParserAdapter correctly ([a82ea9ab](https://github.com/angular/dgeni-packages/commit/a82ea9ab24d718d7c90b4d8766f3308e28019fec)) # 0.17.1 17 March 2017 ## Bug Fixes - **typescript:** ensure all modules have relative filepaths ([434233](https://github.com/angular/angular.js/commit/43423313e246e07a002d2410182d1157b206dbc7)) # 0.17.0 17 March 2017 ## Bug Fixes - **jsdoc:** do not parse for tags within HTML blocks ([451d84](https://github.com/angular/angular.js/commit/451d84a0d0e4bdfd675bced61dd3f2e9cef533a0)) ## Breaking Changes ### **jsdoc** due to: - **[451d84](https://github.com/angular/angular.js/commit/451d84a0d0e4bdfd675bced61dd3f2e9cef533a0)**: do not parse for tags within HTML blocks Tags inside HTML blocks are no longer parsed by default. If you wish this to enable this then you can modify the `parseTagsProcessor.parserAdapters` array from a config block: ``` somePackage.config(function(parseTagsProcessor, backtickParserAdapter) { parseTagsProcessor.parserAdapters = [backtickParserAdapter]; }); ``` # 0.16.10 7 March 2017 * REVERT: feat(typescript): link `class` docs to inherited class docs cf8b7444 # 0.16.9 6 March 2017 * feat(typescript): link `class` docs to inherited class docs cf8b7444 # 0.16.8 20 February 2017 * fix(typescript): correctly extract the type definition from type aliases 573111ea # 0.16.7 15 February 2017 Upgrades to dgeni 0.4.7 which correctly exports the mock logger for testing. * fix(typescript): the processors actually rely upon jsdoc pseudo-processors f5162131 # 0.16.6 13 February 2017 * fix(typescript): support symbols of type ValueModule ffce6227 # 0.16.5 14 January 2017 * chore(package): relax node version dependency 8ad91e57 # 0.16.4 23 December 2016 * chore(package): add missing mkdirp dependency 0a2c2e9f # 0.16.3 21 December 2016 * chore(package): relax yarn version dependency 0f875b0d # 0.16.2 16 December 2016 * refactor(*): remove Q and use native Promise d782698d This fix removes the q and q-io dependencies, since q-io was causing Array.find to be shimmed incorrectly. # 0.16.1 13 December 2016 * fix(base/services): split node-html-encoder require and instantiation over two lines 96cb8139e8876150b5a6ce08f2c5eef7fbd791f6 ## v0.16.0 13 September 2016 * fix(readTypeScriptModules): don't concatenate multiple declarations fb22bac5 ### BREAKING CHANGES Previously, if there were multiple declarations for an exported symbol, then the docs for those declarations were simply joined together with newlines. Now only the first declaration (or the valueDeclaration if there is one) is used in an export doc. If you needed the content for the other declarations then you can still access them via a call like `getContent(doc.additionalDeclarations[i])`. ## v0.15.2 9 September 2016 * feat(typescript/readTypeScriptModules): hide private Typescript members 8a1277de ## v0.15.1 6 September 2016 * fix(git): allow for git URLs that start with `git+https:` c27a324d * doc(readme_fr_FR) : update readme in french da7f6b3c * fix(ngdocs): use `import` syntax for macros 9a69a096 * fix(nunjucks): explicitly set autoescaping to off 54e20018 * fix(nunjucks): warnings for upgraded API, add error check e339b7d8 * chore(package.json) update nunjucks to 2.4.2 01108e4e ### BREAKING CHANGES The update to `nunjucks` 2.4.2 has changed how macros are used. Now you must import them rather than just including them in a template. See http://mozilla.github.io/nunjucks/templating.html#import for more detail on the syntax. Before: ``` {% include "lib/macros.html" -%} {$ functionSyntax(doc) $} ``` Now: ``` {% import "lib/macros.html" as lib -%} {$ lib.functionSyntax(doc) $} ``` The templates provided by this project (in particular those in the `ngdocs` package) have been updated but if your project uses macros in its own templates then these will need updating too. ## v0.14.0 4 July 2016 Major update of dependencies to latest. * chore(package.json): update lodash to 4.13.1 f082551 * chore(package.json): update lodash to 3.x a37d110 * chore(package.json): update q-io to 1.13.2 79819fa * chore(package.json): update q to 1.4.1 9bb7490 * chore(package.json): update semver to 5.2.0 5bfdbba * chore(package.json): update shelljs to 0.7.0 5480c55 * chore(package.json): removed unused packages 1546b8e * chore(package.json): update glob to 7.0.5 5208f8a * chore(package.json): update change-case to 3.0.0 7fcc9c7 * chore(package.json): update minimatch to 3.0.2 8f57d2c ### BREAKING CHANGES The `nunjucks` package rendering filter `sentenceCase` now capitalizes the first letter of the sentence. If you relied on the previous behaviour then you should use the `noCase` filter instead. ## v0.13.1 1 July 2016 * fix(typescript/readTypeScriptModules): `getReturnType` should cope with computed initializers 823d3f9e * fix(ngdoc/macros): use `type.name` to display the directive parameter types 18193fee ## v0.13.0 12 May 2016 * fix(jsdoc/access): refactor access tags and transforms to allow more configuration 70cd0bea * fix(typescript/createCompilerHost): fix syntax errors ae47de91 * refactor(jsdoc/codename): add service, made matchers pluggable 28dc7919 * fix(restrict): Change default from 'A' to 'EA'. e0e6a501 * fix(jsdoc/extract-tags): don't write properties if tag values are `undefined` 3eade6bc ### BREAKING CHANGES * Due to 3eade6bc Tags that had a value of `undefined` are no longer written to the document. If you relied upon such falsy values forcing keys onto the doc, then change your tagDefs to return `null` instead. * Due to e0e6a501 If you don't provide a value for the `@restrict` tag it now defaults to `EA` rather than just `A`. If you want your directives to be only `A` then you must provide a `@restrict` tag for them. * Due to 70cd0bea The way that you configure `extractAccessTransform` has changed. If you were creating your own access tags then you must use the new API. Previously: ``` module.exports = function(accessTagTransform) { var name = 'private'; accessTagTransform.addTag(name); accessTagTransform.addValue(name); function getValue () { return name; } return { name: name, docProperty: 'access', transforms: [getValue, accessTagTransform] }; }; ``` Now: ``` module.exports = function(extractTypeTransform, extractAccessTransform) { extractAccessTransform.allowedTags.set('private'); return { name: 'private', transforms: [extractTypeTransform, extractAccessTransform] }; }; ``` ## v0.12.0 3 April 2016 ### Features * feat(typescript): add package to parse exports from TypeScript modules 3e07adee * feat(jsdoc): add support for @access, @private, @protected, @public tags e87f61f3 * feat(ngdoc): add knownIssues tag and processing b590d058 ### Fixes * fix(jsdoc): only declare private tagDef once 04b6e08e * fix(git): allow non-github repository URLs afc213cc * fix(examples): don't error if the doc has no content 2944e203 * fix(ngdoc/api.template): use link filter for module link 4f4ab7ef * fix(code-name): add support for MethodDefinition, ArrowFunctionExpression nodes cf2c28bf * fix(git/versionInfo): don't throw error if there is no valid codename d749314c ## v0.11.1 7 November 2015 * fix(jsdoc/inline-tags): don't conflate successive inline tags when matching 67c2abf5 * feat(jsdoc): add license tag 6f4bb5f7 * feat(docs): improve templates and add npm script for generating docs 3c75243e * feat(dgeni): compute the full pipeline of processors for each package and add index page c12cd9ec ## v0.11.0 31 October 2015 **ES6 Support:** * fix(code-name): update espree, estraverse and add support for ExportDefaultDeclaration Yair Leviel 080620c2 * fix(jsdoc/jsParser): support parsing ES6 features Peter Bacon Darwin c14c18ce **Dgeni self-documentation:** * feat(docs): initial dgeni doc-gen package Peter Bacon Darwin 603e333e * feat(docs): initial dgeni doc-gen package Peter Bacon Darwin 8c176e53 * feat(dgeni): add dgeni configuration documentation processing Peter Bacon Darwin 0f4f5dd4 **Links package:** * feat(links): initial links package, abstracted from the ngdoc package Peter Bacon Darwin 94271a1a **Start-End Inline Tags** * feat(inline-tags): allow start-end tags Peter Bacon Darwin 362a3495 **Bug Fixes** * fix(versionInfo): don't crash if no codename match is found Peter Bacon Darwin 069c7762 * fix(jsdoc) file-readers assignment via concat Matt Fairbrass c6af7204 ## v0.10.19 23 August 2015 * fix(ngdoc/templates): move module installation above description 6f7b0fc3 ## v0.10.18 10 July 2015 * fix(jsdoc/private tagdef): assign true to private property if tag is found 640d3da6 ## v0.10.17 23 June 2015 * fix(git/api.template): remove weird character in folder name 3ad11ff1 ## v0.10.16 22 June 2015 * fix(ngdoc): remove `git` package as a dependency f0647968 * fix(git/packageInfo): ensure that we return consistent paths on all OSes 4e29bb62 ## v0.10.15 21 June 2015 * feat(git): add git information package 81944685 * fix(ngdoc/templates/lib/events): add event params 860dce98 * fix(extractTypeTransform): support record types 32b269aa * feat(ngdoc/multiElement tag): directives can declare that they support multiElement 41d583f8 ## v0.10.14 11th June 2015 * fix(checkAnchorLinks): add decodeURIComponent call when checking for dangling links d5ed516b * feat(ngdoc/getLinkInfo): add option to make links relative 7178e9eb ## v0.10.13 2rd April 2015 * fix(ngdoc/macros.html): use optional property from param rather than type ## v0.10.12 13th March 2015 * chore(packages): update nunjucks to 1.2.0 ## v0.10.11 13th March 2015 * fix(jsdoc/extract-type): improve error message e56ea300 ## v0.10.10 15th February 2015 * fix(jsdoc/parseTagsProcessor): cope with single line back-ticked code blocks 0d5892d0 ## v0.10.9 15th Februrary 2015 * fix(base/trimIndentation): cope with empty text blocks 8cb1bcb8 ## v0.10.8 15th December 2014 * fix(base/checkAnchorLinksProcessor): ignore "chrome://..." links 3d5dca46 * fix(nunjucks/firstLine): don't break inline tags when extracting first line 19a642f1 ## v0.10.7 18th November 2014 * fix(base/trimIndentation): ensure that lines that start at index zero kill the indentation 434a9b4a * fix(example/template.js): add functional form of 'use strict' f909b9c7 * feat(base/readFilesProcessor): accept arrays of include and exclude patterns 696c4c68 * fix(example/template.js): add 'use strict' a0ceb1e * fix(ngdoc/$moduleDocs): throw an error if a module is improperly tagged d6edbffd ## v0.10.6 10th November 2014 * fix(base/resolveUrl) : Handle the path c:\ under Windows ef23d430 * fix(jsdoc/extractTagsProcessor): remove parameter that was hiding the real variable ba34f921 ## v0.10.5 15th October 2014 * feat(examples/generateProtractorTestsProcessor): allow basePath of test to be specified 2bfb6666 ## v0.10.4 15th October 2014 * fix(base/checkAnchorLinks): match links using pathVariants ff903657 * fix(jsdoc/jsdocFileReader): provide file information in JS parse error 69680361 ## v0.10.3 9th October 2014 * fix(ngdoc/directive.template): add `id` to the `Animations` heading 6e9253d6 ## v0.10.2 8th October 2014 * feat(examples): allow external example dependencies ae900e77 ## v0.10.1 6th October 2014 Mostly minor refactoring and the new `checkAnchorLinksProcessor` Thanks to Lucas, Thor and Stéphane. * refact(base/checkAnchorLinksProcessor): cleaned up the processor 537f6228 * refact(base/extractLinks): rename and also extract id attributes 3ef7b3af * refact(base/resolveUrl): simplify the code f393d4ac * feat(anchors): detect dangling anchors b9b3ef8c * refact(base/computePathsProcessor): warn if no path or outputPath template/getter is found a5b9f536 * refact(jsdoc/jsdocfileReader): move processing into extractJSDocCommentsProcessor cbcbac23 * feat(jsdoc/extractJSDocComments): add new processor b1362834 * fix(jsdoc/parseTagsProcessor): don't error if doc has no content b3da03c3 * feat(ngdoc/templates/base.template): make 'Improve this Doc' lead to appropriate line 55f5f002 ## v0.10.0 29th September 2014 Finally, only 4 betas and 7 release candidates and here is 0.10.0. This version is current being used in production by the angular.js project. * feat(ngdoc/templates/base.template): prefill 'Improve this Doc' commit msg eafb1d2f ## v0.10.0-rc.7 26th September 2014 * fix(ngdoc/templates): fix the github URLs ce548496 * feat(examples/generateExamplesProcessor): allow css-dependencies in examples 10b8278f ## v0.10.0-rc.6 17th September 2014 * feat(examples/index.template.html): add `ng-csp` conditionally 4e173614 ## v0.10.0-rc.5 15th September 2014 ### Bug fixes - Nunjucks fix(nunjucks/marked filter): use renderMarkdown service c82e6a26 fix(nunjucks/marked custom tag): use renderMarkdown service 1ec39d19 feat(nunjucks/renderMarkdown): add new service to allow custom markdown rendering 47ee5ec1 ### New features - JSDoc File Reader * feat(jsdoc/jsdocFileReader): add AST to the fileInfo 998738a7 * refact(jsdoc/jsdocFileReader): use estraverse rather than hacked walk.js 6f8279ce ## v0.10.0-rc.4 12th September 2014 Bug Fixes: * fix(examples/generateExamplesProcessor): manifest files should be an array of strings caf321ec * fix(ngdoc/memberDocs): strip modifier from member name b0b53ec5 ## v0.10.0-rc.3 11th September 2014 Fixes: * feat(ngdoc/providerDocsProcessor): add missing processor 2ea87c15 * fix(ngdoc/moduleDocsProcessor): add missing var declaration 196c1758 Testing Improvements * test(ngdoc/getLinkInfo): test missing and URL based links 4170e83f * refact(ngdoc/moduleDocsProcessor): improve and test error messages d749563f * test(base/writeFilesProcessor): test corner cases fef7a919 * test(nunjucks/filters): add missing tests 13dc8a77 * test(jsdoc/extractTagsProcessor): test corner cases b8157723 * test(base/readFilesProcessor): add corner case tests 17765d52 ## v0.10.0-rc.2 10th September 2014 Bug fix * fix(examples/generateProtractorTestsProcessor): enable testing 'ng-app-included' examples 46a0c516 ## v0.10.0-rc.1 10th September 2014 Minor refactoring * chore(package.json): fix path to tests in npm scripts 2924a3ac * chore(package.json): remove pre and post install hooks 6dd05b21 * chore(examples): move tests into source folder f128fede * chore(ngdoc): move tests into source folder b5df9a16 * chore(nunjucks): move tests into source folder cce54a00 * chore(jsdoc): move tests into source folder 9b337073 * chore(base): move tests into source folder 43b64bbf ## v0.10.0-beta.4 6th September 2014 ### Refactorings #### ComputeIdsProcessor The major refactoring in this release was to move computation of ids (and aliases, which used to be called partialIds) into a generic computeIdsProcessor in the base package. This is then configured (similar to computePathsProcessor) for each docType that a package introduces. * fix(ngdoc): fix idTemplate for member docTypes dbad1945 * fix(examples): add idTemplates for examples docTypes 16086476 * fix(ngdoc/getDocFromAlias): start from scratch if filtering by area fails f742f5a3 * fix(ngdoc/getLinkInfo): pass through currentDoc when searching for member 039deedf * refact(computeIdsProcessor): rename partialIds to aliases 00607193 * refact(base/aliasMap): remove unused reference to lodash 3321da6b * refact(ngdoc): use computeIdProcessor and aliasMap 569312c9 * refact(getPartialNames): rename to getAliases ee83af87 * feat(ngdoc/getDocFromAlias): add service to find relative docs from alias bef71895 * refact(base/computeIdProcessor): rename partialIdMap to aliasMap e336bbc3 * refact(partialIdMap): rename to aliasMap 46075fb3 * feat(base/computeIdsProcessor): add new processor to compute ids based on configuration templates aeec7e1e * feat(base/partialIdMap): add new service for storing partial ids 1c3f82ab * refact(base/computePathsProcessor): always regenerate the maps on $process 4caa96c7 * refact(ngdoc/getAliases): combine getPartialNames and parseCodeName into getAliases 2d3071f0 * refact(ngdoc/getLinkInfo): use getDocFromAlias 0cc9f490 * refact(ngdoc/computeIdProcessor): remove and use base/computeIdsProcessor e17d808c * test(jsdoc): disable computeIdsProcessor for test e303f3c7 * refact(ngdoc/partialNameMap): remove and use aliasMap from base instead 925c01a1 * docs(ngdoc/getDocFromAlias): document the service 167a982b * fix(base/computeIdsProcessor): improve error messages 88ea6e10 * fix(jsdoc): provide basic getPartialIds function 54dc8118 * refact(jsdoc): update to use computeIdsProcessor 2938a4c5 #### apiDocsProcessor Also the apiDocsProcessor was broken up. Part of it is now handled by the `computeIdsProcessor` but the rest was refactored into `membersDocsProcessor` and `moduleDocsProcessor`. * refact(ngdoc/generateComponentGroupsProcessor): apiDocs is going away 1fbed695 * refact(ngdoc/apiDocsProcessor): remove old processor c1a2fee0 * refact(ngdoc/moduleDocsProcessor): use getDocFromAlias fa0408ca * refact(ngdoc/memberDocsProcessor): use getDocFromAlias 74063d3f * fix(ngdoc/memberDocsProcessor): test and ensure doc.memberof is a full id f6b063dd * fix(ngdoc/moduleDocsProcessor): run after memberDocsProcessor c3fd9398 * fix(ngdoc/moduleDocsProcessor): better resolve ambiguous module references b27d8ff7 * test(ngdoc/moduleDocsProcessor): test the processor 3e42370c * refact(ngdoc): reorganization of processors and services c7397fa1 * feat(ngdoc/moduleDocsProcessor): extract functionality from apiDocsProcessor 8cd51dc4 * feat(ngdoc/memberDocsProcessor): extract functionality from apiDocsProcessor c7302275 #### `dgeni.configureInjector` and tests The a new version of dgeni allowed us access to the injector without having to run `dgeni.generate()`. This meant that we could have much cleaner tests, where we can let the injector create our services rather than having to wire them up manually - which was a pain and error prone. * test(ngdoc/getLinkInfo): use mockPackage for testing 896cad63 * test(ngdoc/getDocFromAlias): test with mockPackage b819e338 * test(ngdoc/mockPackage): add mock Package for testing 659e8d09 * test(ngdoc/filterNgDocsProcessor): use mockPackage and injector 2eea5026 * test(ngdoc/generateComponentGroups): use mockPackage and injector 13da2040 * test(jsdoc): use mockPackage 9d08564b * test(base/computePathsProcessor): get mockLog from injector b154d5a2 * test(base/computeIdsProcessor): use configureInjector and mockPackage 7ef665dc * test(base/aliasMap): use configureInjector and mockPackage e641efd3 * test(base/computePathsProcessor): use configureInjector and mockPackage 2528b445 * test(base/createDocMessage): use configureInjector and mockPackage 5d45118f * test(base/debugDumpProcessor): use configureInjector and mockPackage 4b4eeb9e * test(base/encodeCodeBlock): use configureInjector and mockPackage 54d6d19a * test(base): use mockPackage e6366e83 * test(base/templateFinder): use configureInjector and mockPackage 1b8559e3 * test(base/trimIndentation): use configureInjector and mockPackage 176d6241 * test(mockPackage): add a mock package to be used in testing 293563c3 #### link `node_modules/dgeni-package` -> `..` To make it easier to reference modules within dgeni-packages, preinstall and postinstall hooks are now in place to create a soft link to the root of the projet from within node_modules: * chore(package.json): add link to this module in node_modules 68a08ec3 * chore(package.json): make node_module linking more tolerant bfbe201b #### writeFile Service * feat(base/writeFile): add simple service to write a file to disk 1c9681e9 * feat(base/writeFile): add simple service to write a file to disk f035c854 * refact(base/writeFilesProcessor): user writeFile service aa3be581 * refact(base/debugDumpProcessor): use writeFile service adc64c7c ### Bug Fixes * fix(generateComponentGroup processor): ensure doc has a name 27469bec * fix(jsdoc/codeNameProcessor): don't override codeName if already provided 942a72f6 ## v0.10.0-beta.3 15th August 2014 The major change in this release was the removal of dependency on the ES6-shim and the use of ES6 `Map` for storing maps of information. The shimmed Map was not playing nicely with Jasmine, Nunjucks and `console.log`. Instead these mappings are now using [`StringMap`](https://github.com/olov/stringmap), which is cleaner than using bare objects (made from `Object.create(null)`). Thanks as always to Stéphane for his translations and removal of unused code. * refact(*) : remove unused dependencies 5c893d85 * refact(examples/exampleMap): rename and convert to a StringMap bdbb7c26 * fix(ngdoc/templates): fix github links 2f01d163 * fix(examples): use computePathsProcessor for example file paths c4b02945 * fix(examples/generateExamplesProcessor): ensure correct files are in manifest dc60336e * fix(ngdoc/templates): ensure github links are to valid paths 7676f187 * feat(base/readFilesProcessor): add projectRelativePath to fileInfo 7931f230 * fix(examples): ensure example index files have correct path e5e62554 * refact(jsdoc/moduleMap): use a StringMap to store the modules 0c37674a * refact(TagColletion) use StringMap for tagsByName property d6655cf0 * refact(jsdoc/parseTagsProcessor): use StringMap b61e3e20 * refact(jsdoc/inlineTagProcessor): use StringMap 33f6e5fc * refact(computePathsProcessor): use stringmap and no longer allow default template 93fc95cf * refact(base/readFilesProcessor): use stringmap for fileReaderMap be33702c * fix(ngdoc/macro.html): default param values are in `doc. e3899b67 * fix(partialNameMap): use pure object to prevent property collisions 3750c928 * refact(*): remove dependency on 'es6-shim' and `Map` 15751dfb ## v0.10.0-beta.2 7th August 2014 Bug fixes and refactoring. The big change in this revision is the introduction of the `computePathsProcessor`. * refact(examples): use computePathsProcessor f85487f7 * refact(examples/parseExamplesProcessor): ditch using ES6 Map for examples c7d2b6db * refact(examples/generateExamplesProcessor): ditch using ES6 Map for examples a770bf91 * refact(examples/runnableExample inlineTag): ditch using ES6 Map for examples f1b463dc * feat(examples/generateProtractorTestsProcessor): add new processor from Angular.js c08b9be4 * refact(examples/examples): ditch using ES6 Map 37ac555c * fix(ngdoc): ensure API docs go into 'partials' folder e9aa9f81 * fix(ngdoc/generateComponentGroupsProcessor): use moduleName and moduleDoc rather than module a0f32a03 * fix(ngdoc/module tagDef): don't calc module if docType is "overview" ba819910 * fix(ngdoc/area tagDef): docs now use relativePath not filePath 620a2b07 * fix(createDocMessage): docs now use relativePath not filePath efb7397c * fix(computePaths): use log.debug not console.log 1a8f18ba * refact(examples/templates): move files up a folder 7e7fdea5 * feat(base/createDocMessage): include docType in message 7add32a5 * feat(base/computePathsProcessor): improve error messages 14086ed2 * refact(examples): use computePathProcessor from base package f6891e9b * fix(computeIdProcessor): don't compute if id is already defined bab945be * refact(ngdoc): use computePathProcessor from base package fe187590 * fix(jsdoc): add 'js' docType path template a3d5f5b1 * feat(jsdoc file-reader): give each read doc a `docType` of `"js"` 2f15a7cf * refact(jsdoc): use computePathProcessor from base package 90adbf8b * refact(jsdoc/computePathProcessor): use the computePathProcessor from base package a2e7a6f8 * feat(computePathsProcessor): add configurable processor 15179ce5 ## v0.10.0-beta.1 25th July 2014 **This is a major rearchitecture in line with changing to [dgeni@0.4.0](https://github.com/angular/dgeni/blob/master/CHANGELOG.md#v040-beta1-25th-july-2014)** There are numerous breaking changes surrounding this release and that of dgeni 0.4.0. The most important changes are * Configuration is done directly on the Processors and Services using Configuration Blocks, which are defined on Packages. * Everything is now dependency injected. Dgeni deals with instantiating Processors and Services but if you have properties on these that reference objects that should also be instantiated by the DI system then you can either ask for them to be injected into the config block: ```js myPackage.config(function(processor1, someObj) { processor1.someProp = someObj; }); ``` use the `injector` directly: ```js myPackage.config(function(processor1, injector) { processor1.someProp = injector.invoke(someObjFactory); }); ``` or use the `getInjectables()` helper service: ```js myPackage.config(function(processor1, getInjectables) { processor1.someProp = getInjectables([someObjFactory, someOtherObjFactory]); }); ``` * All real processors have changed their names from dash-case to camelCase. This is because it is easier for their names to be valid JavaScript identifiers. The most significant commits are: * fix(utils/code): encode HTML entities 13b99152 * feat(base/debugDumpProcessor): add new processor 4c126792 * feat(dgeni package): add initial package for documenting dgeni configurations 2bfa92b2 * refact(parseExamplesProcessor): use Map() for example.files d926879a * fix(ngdoc/module tag-def): module is the first segment of the relative path 649f3051 * fix(jsdoc/description tag-def): capture non-tag specific description ed68438d * feat(base/createDocMessage): add new service db11bc44 * fix(*): doc.file is now doc.fileInfo.filePath fb600502 * fix(ngdoc/collectPartialNamesProcessor): compute-id was renamed to computeIdProcessor 0158fb3b * fix(ngdoc/apiDocs): compute-path was renamed to computePathProcessor 9396f8c3 * fix(ngdoc/apiDocsProcessor): compute-id was renamed to computeIdProcessor a83d7fc9 * feat(ngdoc package): add getTypeClass service addebf63 * refact(base/code): rename to encodeCodeBlock 2ae134ff * feat(getTypeClass): add new service 9c49ff9d * fix(ngdocFileReader): must have an explicit name 9c5dd397 * test(nunjucks/templateEngine): templateEngine now relies on templateFinder 0269acf5 * fix(nunjucks/marked custom tag): add service name for DI injection 9069c24f * refact(trimIndentation): convert to DI service 3ab6c9ed * fix(nunjucks/templateEngine): get the templateFolders from the templateFinder 8760aa7f * fix(inlineTagProcessor): inline tag definitions are optional 380dd474 * fix(computePathProcessor): let writeFileProcessor deal with outputFolder 4057deb8 * fix(jsdoc fileReader): file-readers need explicit names cad6d043 * fix(jsdoc package): pseudo processors need to use $runBefore, etc. 5d32020f * fix(renderDocsProcessor): extra and helpers and optional 6bc1f16a * fix(readFilesProcessor): resolve include and exclude paths correctly f7ea5f78 * refact(computePathProcessor): read outputFolder config from writeFilesProcessor 828b48c3 * feat(api-docs): allow packageName to be specified as a tag 83c7e1fa * fix(jsdoc package): add trimWhitespaceTransform to package 75f52df1 * refact(templateFinder): now call getFinder() to get the actual function e8c015d9 * refact(jsdoc transforms): convert to DI services 5e92ff46 * feat(ngdocs/moduleMap): add new service to support apiDocsProcessor et al 19a37a27 * fix(runnableExample inline-tag): examples is now a Map d1857779 * test(tag-defs): split out tests into separate files 9be421f7 * fix(jsdoc package): jsdocFileReader should be loaded as a service 0c92271c * refact(ngdoc/tag-defs): convert to DI injectables d8512597 * refact(ngdoc/link inline tag): convert to DI injectable 77e51df3 * refact(ngdoc/code tag): convert to DI injectable 1d23cd4e * refact(ngdoc/code filter): convert to DI injectable 8c871b3d * feat(getLinkInfo): add new service 32aa8427 * refact(partialNameMap): rename and convert to DI service 49acc949 * feat(getPartialNames): add new service c6326268 * feat(parseCodeName): add new service e02fe91c * refact(base/services): move code and trimIndentation to be DI services 00d17816 * refact(examples service): move to its own service (as part of no-config update) 16d7819e * refact(renderDocsProcessor): templateEngine now has a `getRenderer()` method 4f50737b * refact(jsdoc): convert transforms to services 6dc5417f * refact(extractTagsProcessor): move computation into smaller functions fa1821c1 * refact(tagExtractor): move into the extractTagsProcessor 49e51da4 * refact(tagParser): move into the parseTagsProcessor efba4e9b * feat(read-files): add path exclusion and update to no-config 6853d759 * refact(*): update to use new processor configuration style c54fd8d6 * refact(*): use new dgeni Packages 128c2e61 ## v0.9.3 05/22/2014 * fix(extract-type): cope with missing type 4584a423 ## v0.9.2 05/09/2014 * fix(jsdoc package): trim whitespace from tags afa5c8c6 ## v0.9.1 05/02/2014 * fix(ngdoc/compute-path): use ngdoc specific version of this processor 3e17d31b * fix(code-name): cope with additional code situations 8b456b08 * fix(jsdoc/trim-whitespace transform): only trim strings 5aa2376d * fix(jsdoc/jsdoc file-reader): cope with comments that have no code node 4e3857db * fix(code-name): recognise CallExpression nodes 14c5c103 ## v0.9.0 05/01/2014 This is a major refactoring release which is compatible with Dgeni v0.3.x. There are many breaking changes. ### New Packages The packages have been refactored into smaller more focussed sets of processors. * base - The minimal set of processors to get started with Dgeni * jsdoc - Tag parsing and extracting * nunjucks - The nunjucks template rendering engine. No longer in jsdoc - you must add this explicitly to your config or you will get `Error: No provider for "templateEngine"! (Resolving: templateEngine)` * ngdoc - The angular.js specific tag-defs, processors and templates. This loads the jsdoc and nunjucks packages for you. * examples - Processors to support the runnable examples feature in the angular.js docs site. ### New Processor Exports With Dgeni 0.3.0 processors can declaratively export services to be injected into processors' `process()` method. This release take full advantage of this, refactoring the structure of the dependencies of various processors to simplify and enable more flexibility. ### New Tag Definition Transforms Previously processing of tags was somewhat distributed between the tagParser and the tagExtractor, with various features rather hard-coded, such as `canHaveType` and `canHaveName`. This has all been moved into **tag definition transforms**, which provide a much more flexible and powerful way to define how to transform the simple text "parsed" from the tag into a rich object that can be attached to the document. ### Detailed List of Changes **Features** * feat(jsdoc/tag-defs): add `@type` tag 904aa00b * feat(jsdoc/tag-defs): add `@method` tag 6fc99313 * feat(jsdoc file-reader): add more code metadata 7820317a * feat(jsdoc/name-from-code): extract the name of the doc from the code 1115c431 **Refactorings** * refact(ngdoc/tag-defs): use new tagExtractor syntax b0848557 * refact(jsdoc/tag-defs): use new tagExtractor syntax 0048547e * refact(extract-tags processor): rename and use tagExtractor d2ef2237 * refact(tagExtractor): major reworking to use 'transformFns' a90734ae * refact(parse-tags): simplify using tagParser 02cc093d * refact(tagParser): move into its own processor 1d2e689e * refact(tagDefinitions): move into its own processor d2d916ef * refact(defaultTagTransforms): move into its own processor 354a9489 * refact(tag-def/transforms): convert "tagProcessors" into tag "transforms" d4ca4e94 * refact(nunjucks): move basic filters and tags to nunjucks package 5547409c * refact(dash-case): change name to change-case 39ab9be1 * refact(walk): remove unused code 66f86a50 * refact(marked): remove unused code 4591bd90 * refact(doc-writer): remove unused code 41c0cd12 * refact(dash-case): remove unused code 8f9dd805 * refact(check-property): remove unused code 6e39ba70 * refact(code-name): move to jsdoc package 4061e64f * refact(packages): align with renames and moves of processors a21f804e * refact(doc-extractor): complete rename to read-files 8bc760bf * refact(escaped-comments): rename to unescape-comments d030ca95 * refact(rendering): move nunjucks stuff out 0b36e95a * refact(code-name): rename 6e312504 * refact(doc-extractor): rename to read-files 60ac908c * refact(partial-names processor): remove `init` and provide `exports` 0de2680a * refact(component-groups-generate processor): remove `init` and provide `exports` 61ba9cb1 * refact(api-docs processor): remove `init` and provide `exports` 34d05b1f * refact(jsdoc processors): remove `init` and provide `exports` 3188ff14 * refact(examples-parse): remove `init` and provide `exports` 5cbcbab6 * refact(examples-generate): remove `init` and provide `exports` 0c006bb4 * refact(nunjucks-renderer): remove `init` and provide `exports` f4b42dd6 * refact(doc-extractor): remove `init` and provide `exports` fce833c9 * refact(*): update due to utils move 0245eb22 * refact(utils): moved here from dgeni 1f04843d * refact(jsdoc): moved stuff to base package 5f2ec6be **Bug Fixes** * fix(tagExtractor): invalid injectable parameter name 6c8790f3 * fix(jsdoc): add defaultTagTransforms and tagExtractor processors 12ac7aa0 * fix(jsdoc): tag-extractor processor was renamed 1931f4d4 * fix(tagExtractor): accidental global vars 5de56780 * fix(extract-type transform): ensure tag.description gets updated 0f911d8a * fix(extract-name transform): ensure tag.description gets updated e347c847 * fix(tagDefinitions): throw error if tag definitions are missing from the config 4777ec79 * fix(nunjucks): correctly load up the template engine processor 071e52c7 * fix(base processors): minor fixes to get the tests working efd3e35f * fix(link inline tag): parse newlines in link's title b2ebb415 * fix(ngdoc): don't show first param in filter syntax 4f2ccf52 * fix(walk): hack ancestor to kind of do what I want f19b6940 * fix(compute-path): ensure it runs early enough fe1e0bd7 * fix(jsdoc package): actually append processors to config f2020d47 * fix(marked tag): fix path to trim-indentation module 1206f8fe * fix(nunjucks-renderer): `env` changed to `templateEngine` e4a756e0 * fix(base package): load change-case filter locally a74f2ee4 * fix(compute-paths): ensure it is run before rendering b50baa99 * fix(jsdoc/tag-defs): Allow multiple `@see` tags d73a842f * fix(tag-parser): don't overwrite default tag processors collection 9fc7f58e * fix(code): fix path to utilities 26c26e70 * fix(examples-parse): fix path to utilities 1f9d1488 * fix(jsdoc/tag-defs): missing comma 293ffde2 * fix(jsdoc): fix typo in error message b1e7dd08 ## v0.8.3 04/23/2014 **Bug Fixes** * fix(ngdoc): don't show first param in filter syntax 9c5a7f26 * fix(jsdoc): fix typo in error message b1e7dd08 ## v0.8.2 03/22/2014 **Bug Fixes** * fix(jsdoc/jsdoc-extractor): ensure Windows newlines are respected 7f1e1627 * fix(jsdoc/jsdoc-extractor): fix off-by-one error fedcf6b1 * fix(jsdoc/lib/walk): also walk the keys of properties 4a8ae60d * fix(jsdoc/nunjucks-renderer): improve error message 554c7afd ## v0.8.1 03/18/2014 **Bug Fixes** * fix(examples/index.template): nunjucks uses `not` not `!` e0070e3f ## v0.8.0 03/18/2014 **New Features** * feat(examples/index.template): allow ng-app to be defined in the example 9b760c08 * feat(jsdoc package): use `jsdoc` rather than `js` doc-extractor 72866263 ## v0.7.1 03/11/2014 **Minor Features** * feat(api-docs): add configuration for paths 8ddcf647 * feat(runnableExample template): provide the path to the example 353eef6e ## v0.7.0 03/11/2014 **New Features** * feat(TagCollection): add an array of Tags via the constructor 7c1cca1a * feat(jsdoc/tag-defs): add deprecated tag 0f3a1949 * feat(examples-generate): create manifest.json file for Plunker 1849ec49 ## v0.6.0 03/07/2014 **New Features** * feat(tagParser): only ignore tags that are defined with ignore property 59492bea * feat(jsdoc tags): improve jsdoc tag coverage d8eb2b43 * feat(PartialNames): getLink disambiguates docs by area 6da98dd5 * feat(jsdoc/compute-paths): add new processor 23cc829a * feat(partial-names): add removeDoc method 746a0bc7 * test(link handler): fix test, since the handler now throws on error a15053ab **Bug Fixes** * fix(compute-paths): ensure contentsFolder is applied correctly 301877fc * fix(filter-ngdocs processor): run before tags are extracted a090cf49 **Refactorings** * refact(ngdoc/id tag def): move functionality to its own doc processor 268ac3bd * refact(partial-name processor): move adding docs to own processor 8684226b **BREAKING CHANGE** * If you relied on undefined tags being quietly ignored your processing will now fail. You should add new tag defintions for all tags that you wish to ignore of the form: ``` { name: 'tag-to-ignore', ignore: true } ``` ## v0.5.0 03/07/2014 **New Features** * feat(jsdoc extractor): add next code node to the doc 22a59651 **Bug Fixes** * fix(jsdoc extractor): ignore non-jsdoc comments 50ad83d8 * fix(inline link tag): throw error if link is invalid 07af2f42 ## v0.4.0 03/06/2014 **New Features** * feat(examples): move injected example into a template cc658f31 * feat(jsdoc): add `rendering.nunjucks.config` field to config eb805097 * feat(link inline handler): replace old link processor with new inline handle 723e0e56 * feat(inline-tag processor): add new generic inline tag processor 39083631 * feat(firstParagraph filter): add new filter 4dcabba1 * feat(jsdoc extractor): add esprima powered jsdoc extractor e96da1ea **Bug Fixes** * fix(ngdoc/templates) : improve "View Source" and "Improve Doc" links 049ee59f * fix(write-files): ignore docs that have no output path c666e5e3 * fix(ngdoc templates): show first paragraph not first line 7335fd91 **BREAKING CHANGE** * The `examples` injectable object has changed from being a flat array to a hash indexed on the id of the example. If you only iterated over the examples then things like `forEach` should still just work. But you can no longer access the examples by index, e.g. `examples[0]` will return undefined rather than the first example. ## v0.3.1 03/02/2014 **Bug Fixes** * fix(tagParser): don't break on bad-tags 560eff7b ## v0.3.0 02/28/2014 **New Features** * feat(tagParser): ignore tags inside fenced code blocks 09fb7d64 * feat(trimProcessor): add tag processor to trim off whitespace a81f6231 * feat(nameProcessor): add support for param aliases 5720bfed **Bug Fixes** * fix(typeProcessor): handle escaped braces 786f1ab5 * fix(ngdoc templates): ensure type hints are escaped 4ace02b8 * fix(escaped-comments): re-code HTML escaped comment markers 020fde5c * fix(examples-parse): ensure that code blocks are rendered correctly 10ae5e21 * fix(api-docs processor): don't contaminate the global context 2fa8acf6 * fix(typeProcessor): add better error message 6becbd46 * fix(tag-parser): add better error message 458b26f5 * fix(tagParser): cope with tags that have no following whitespace 04cf4f02 * fix(typeProcessor): attach optional property to tag if type is optional 0188305f ## v0.2.4 02/25/2014 **Bug Fixes** * fix(doctrine-tag-parser): don't rethrow error if tag type is bad 7ac46af6 ## v0.2.3 02/25/2014 **Bug Fixes** * fix(doctrine-tag-parser): support jsdoc3 tags and improve error messages c8ca67a2 ## v0.2.2 02/21/2014 **Bug Fixes** * fix(examples-generate): ensure each index file gets content c4918e05 * fix(ngdoc/members): render member docs correctly c7b98a67 ## v0.2.1 02/20/2014 **Bug Fixes** * fix(example-generation): commonFiles should get scripts from the 'scripts' object 3b41c91a ## v0.2.0 02/20/2014 **New Features** * feat(example-generation): generate examples for multiple deployments 82ba9054 ## v0.1.0 02/20/2014 **Bug Fixes** * fix(doc-extractor): give decent error if projectPath is missing 0e326692