/
home
/
sjslayjy
/
public_html
/
theweavenest
/
vendor
/
league
/
commonmark
/
docs
/
2.0
/
upgrading
/
Upload File
HOME
--- layout: default title: Upgrading from 1.6 to 2.0 (for developers) description: Upgrade guide for those who develop custom extensions and more using this library --- ## Minimum PHP Version The minimum supported PHP version was increased from 7.1 to 7.4. ## `CommonMarkConverter` and `GithubFlavoredMarkdownConverter` constructors The constructor methods for both `CommonMarkConverter` and `GithubFlavoredMarkdownConverter` no longer accept passing in a customized `Environment`. If you want to customize the extensions used in your converter you should switch to using `MarkdownConverter`. See the [Basic Usage](/2.0/basic-usage/) documentation for an example. ```diff -use League\CommonMark\CommonMarkConverter; use League\CommonMark\Environment\Environment; use League\CommonMark\Extension\CommonMark\CommonMarkCoreExtension; use League\CommonMark\Extension\Table\TableExtension; +use League\CommonMark\MarkdownConverter; $config = [ 'html_input' => 'escape', 'allow_unsafe_links' => false, 'max_nesting_level' => 100, ]; -$environment = new Environment(); +$environment = new Environment($config); $environment->addExtension(new CommonMarkCoreExtension()); $environment->addExtension(new TableExtension()); -$converter = new CommonMarkConverter($config, $environment); // or GithubFlavoredMarkdownConverter +$converter = new MarkdownConverter($environment); echo $converter->convertToHtml('Hello World!'); ``` ## `CommonMarkConverter` Return Type In 1.x, calling `convertToHtml()` would return a `string`. In 2.x this changed to return a `RenderedContentInterface`. To get the resulting HTML, either cast the return value to a `string` or call `->getContent()`. (This new interface extends from `Stringable` so you can type hint against that instead, if needed.) ```diff use League\CommonMark\CommonMarkConverter; $converter = new CommonMarkConverter(); -echo $converter->convertToHtml('# Hello World!'); +echo $converter->convertToHtml('# Hello World!')->getContent(); +// or +echo (string) $converter->convertToHtml('# Hello World!'); ``` ## HTML Changes Table of Contents items are no longer wrapped with `<p>` tags: ```diff <ul class="table-of-contents"> <li> - <p><a href="#level-2-heading">Level 2 Heading</a></p> + <a href="#level-2-heading">Level 2 Heading</a> </li> <li> - <p><a href="#level-4-heading">Level 4 Heading</a></p> + <a href="#level-4-heading">Level 4 Heading</a> </li> <li> - <p><a href="#level-3-heading">Level 3 Heading</a></p> + <a href="#level-3-heading">Level 3 Heading</a> </li> </ul> ``` See [#613](https://github.com/thephpleague/commonmark/issues/613) for details. Additionally, the HTML (including URL fragments) for Heading Permalinks have changed: ```diff -<h1><a id="user-content-hello-world" href="#hello-world" name="hello-world" class="heading-permalink" aria-hidden="true" title="Permalink">¶</a>Hello World!</h1> +<h1><a id="content-hello-world" href="#content-hello-world" class="heading-permalink" aria-hidden="true" title="Permalink">¶</a>Hello World!</h1> ``` Note that the `href` now targets the `id` attribute instead of the `name`, which is deprecated in HTML 5. Additionally, the default prefix has changed to `content`. See the [Heading Permalink extension documentation](/2.0/extensions/heading-permalinks/#configuration) for more details on how to configure the prefixes. ## Configuration Option Changes Several configuration options now have new names: | Old Key/Path | New Key/Path | Notes | | ----------------------------------- | ----------------------------------- | --------------------------------------- | | `enable_em` | `commonmark/enable_em` | | | `enable_strong` | `commonmark/enable_strong` | | | `use_asterisk` | `commonmark/use_asterisk` | | | `use_underscore` | `commonmark/use_underscore` | | | `unordered_list_markers` | `commonmark/unordered_list_markers` | Empty arrays no longer allowed | | `heading_permalink/id_prefix` | (unchanged) | Default is now `content` | | `heading_permalink/inner_contents` | `heading_permalink/symbol` | | | `heading_permalink/slug_normalizer` | `slug_normalizer/instance` | | | `max_nesting_level` | (unchanged) | Only integer values are supported | | `mentions/*/symbol` | `mentions/*/prefix` | | | `mentions/*/regex` | `mentions/*/pattern` | Cannot contain start/end `/` delimiters | ## Method Return Types Return types have been added to virtually all class and interface methods. If you implement or extend anything from this library, ensure you also have the proper return types added. ## Configuration Classes Relocated The following classes have been moved to the `league/config` package: | Old Class Namespace/Name (1.x) | Moved To | | ---------------------------------------------------- | ------------------------------------------- | | `League\CommonMark\Util\ConfigurationAwareInterface` | `League\Config\ConfigurationAwareInterface` | | `League\CommonMark\Util\ConfigurationInterface` | `League\Config\ConfigurationInterface` | | `League\CommonMark\Util\Configuration` | `League\Config\Configuration` | ## Classes/Namespaces Renamed Many classes now live in different namespaces, and some have also been renamed. Here's a quick guide showing their new locations: _(Note that the base namespace of `League\CommonMark` has been omitted from this table for brevity.)_ | Old Class Namespace/Name (1.x) | New Class Namespace/Name (2.0) | | ------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------- | | `ConfigurableEnvironmentInterface` | `Environment\ConfigurableEnvironmentInterface` | | `EnvironmentAwareInterface` | `Environment\EnvironmentAwareInterface` | | `Environment` | `Environment\Environment` | | `EnvironmentInterface` | `Environment\EnvironmentInterface` | | `Extension\CommonMarkCoreExtension` | `Extension\CommonMark\CommonMarkCoreExtension` | | `Delimiter\Processor\EmphasisDelimiterProcessor` | `Extension\CommonMark\Delimiter\Processor\EmphasisDelimiterProcessor` | | `Block\Element\BlockQuote` | `Extension\CommonMark\Node\Block\BlockQuote` | | `Block\Element\FencedCode` | `Extension\CommonMark\Node\Block\FencedCode` | | `Block\Element\Heading` | `Extension\CommonMark\Node\Block\Heading` | | `Block\Element\HtmlBlock` | `Extension\CommonMark\Node\Block\HtmlBlock` | | `Block\Element\IndentedCode` | `Extension\CommonMark\Node\Block\IndentedCode` | | `Block\Element\ListBlock` | `Extension\CommonMark\Node\Block\ListBlock` | | `Block\Element\ListData` | `Extension\CommonMark\Node\Block\ListData` | | `Block\Element\ListItem` | `Extension\CommonMark\Node\Block\ListItem` | | `Block\Element\ThematicBreak` | `Extension\CommonMark\Node\Block\ThematicBreak` | | `Inline\Element\AbstractWebResource` | `Extension\CommonMark\Node\Inline\AbstractWebResource` | | `Inline\Element\Code` | `Extension\CommonMark\Node\Inline\Code` | | `Inline\Element\Emphasis` | `Extension\CommonMark\Node\Inline\Emphasis` | | `Inline\Element\HtmlInline` | `Extension\CommonMark\Node\Inline\HtmlInline` | | `Inline\Element\Image` | `Extension\CommonMark\Node\Inline\Image` | | `Inline\Element\Link` | `Extension\CommonMark\Node\Inline\Link` | | `Inline\Element\Strong` | `Extension\CommonMark\Node\Inline\Strong` | | `Block\Parser\BlockQuoteParser` | `Extension\CommonMark\Parser\Block\BlockQuoteParser` | | `Block\Parser\FencedCodeParser` | `Extension\CommonMark\Parser\Block\FencedCodeParser` | | `Block\Parser\ATXHeadingParser` and `Block\Parser\SetExtHeadingParser` | `Extension\CommonMark\Parser\Block\HeadingParser` | | `Block\Parser\HtmlBlockParser` | `Extension\CommonMark\Parser\Block\HtmlBlockParser` | | `Block\Parser\IndentedCodeParser` | `Extension\CommonMark\Parser\Block\IndentedCodeParser` | | `Block\Parser\ListParser` | `Extension\CommonMark\Parser\Block\ListParser` | | `Block\Parser\ThematicBreakParser` | `Extension\CommonMark\Parser\Block\ThematicBreakParser` | | `Inline\Parser\AutolinkParser` | `Extension\CommonMark\Parser\Inline\AutolinkParser` | | `Inline\Parser\BacktickParser` | `Extension\CommonMark\Parser\Inline\BacktickParser` | | `Inline\Parser\BangParser` | `Extension\CommonMark\Parser\Inline\BangParser` | | `Inline\Parser\CloseBracketParser` | `Extension\CommonMark\Parser\Inline\CloseBracketParser` | | `Inline\Parser\EntityParser` | `Extension\CommonMark\Parser\Inline\EntityParser` | | `Inline\Parser\EscapableParser` | `Extension\CommonMark\Parser\Inline\EscapableParser` | | `Inline\Parser\HtmlInlineParser` | `Extension\CommonMark\Parser\Inline\HtmlInlineParser` | | `Inline\Parser\OpenBracketParser` | `Extension\CommonMark\Parser\Inline\OpenBracketParser` | | `Block\Renderer\BlockQuoteRenderer` | `Extension\CommonMark\Renderer\Block\BlockQuoteRenderer` | | `Block\Renderer\FencedCodeRenderer` | `Extension\CommonMark\Renderer\Block\FencedCodeRenderer` | | `Block\Renderer\HeadingRenderer` | `Extension\CommonMark\Renderer\Block\HeadingRenderer` | | `Block\Renderer\HtmlBlockRenderer` | `Extension\CommonMark\Renderer\Block\HtmlBlockRenderer` | | `Block\Renderer\IndentedCodeRenderer` | `Extension\CommonMark\Renderer\Block\IndentedCodeRenderer` | | `Block\Renderer\ListBlockRenderer` | `Extension\CommonMark\Renderer\Block\ListBlockRenderer` | | `Block\Renderer\ListItemRenderer` | `Extension\CommonMark\Renderer\Block\ListItemRenderer` | | `Block\Renderer\ThematicBreakRenderer` | `Extension\CommonMark\Renderer\Block\ThematicBreakRenderer` | | `Inline\Renderer\CodeRenderer` | `Extension\CommonMark\Renderer\Inline\CodeRenderer` | | `Inline\Renderer\EmphasisRenderer` | `Extension\CommonMark\Renderer\Inline\EmphasisRenderer` | | `Inline\Renderer\HtmlInlineRenderer` | `Extension\CommonMark\Renderer\Inline\HtmlInlineRenderer` | | `Inline\Renderer\ImageRenderer` | `Extension\CommonMark\Renderer\Inline\ImageRenderer` | | `Inline\Renderer\LinkRenderer` | `Extension\CommonMark\Renderer\Inline\LinkRenderer` | | `Inline\Renderer\StrongRenderer` | `Extension\CommonMark\Renderer\Inline\StrongRenderer` | | `Extension\SmartPunct\PunctuationParser` | `Extension\SmartPunct\DashParser` and `Extension\SmartPunct\EllipsesParser` | | `Extension\TableOfContents\TableOfContents` | `Extension\TableOfContents\Node\TableOfContents` | | `Block\Element\AbstractBlock` | `Node\Block\AbstractBlock` | | `Block\Element\Document` | `Node\Block\Document` | | `Block\Element\InlineContainerInterface` | `Node\Block\InlineContainerInterface` | | `Block\Element\Paragraph` | `Node\Block\Paragraph` | | `Block\Element\StringContainerInterface` | `Node\StringContainerInterface` | | `Inline\Element\AbstractInline` | `Node\Inline\AbstractInline` | | `Inline\Element\AbstractStringContainer` | `Node\Inline\AbstractStringContainer` | | `Inline\AdjacentTextMerger` | `Node\Inline\AdjacentTextMerger` | | `Inline\Element\Newline` | `Node\Inline\Newline` | | `Inline\Element\Text` | `Node\Inline\Text` | | `Block\Parser\BlockParserInterface` | `Parser\Block\BlockContinueParserInterface` and `Parser\Block\BlockStartParserInterface` | | `Block\Parser\LazyParagraphParser` | `Parser\Block\ParagraphParser` | | `Cursor` | `Parser\Cursor` | | `DocParser` | `Parser\MarkdownParser` | | `DocParserInterface` | `Parser\MarkdownParserInterface` | | `Inline\Parser\InlineParserInterface` | `Parser\Inline\InlineParserInterface` | | `Inline\Parser\NewlineParser` | `Parser\Inline\NewlineParser` | | `InlineParserContext` | `Parser\InlineParserContext` | | `InlineParserEngine` | `Parser\InlineParserEngine` | | `Block\Renderer\DocumentRenderer` | `Renderer\Block\DocumentRenderer` | | `Block\Renderer\ParagraphRenderer` | `Renderer\Block\ParagraphRenderer` | | `ElementRendererInterface` | `Renderer\ChildNodeRendererInterface` | | `HtmlRenderer` | `Renderer\HtmlRenderer` | | `Inline\Renderer\NewlineRenderer` | `Renderer\Inline\NewlineRenderer` | | `Inline\Renderer\TextRenderer` | `Renderer\Inline\TextRenderer` | | `Block\Renderer\BlockRendererInterface` and `Inline\Renderer\InlineRendererInterface` | `Renderer\NodeRendererInterface` | | `HtmlElement` | `Util\HtmlElement` | ## New Block Parsing Approach We've completely changed how block parsing works in 2.0. In a nutshell, 1.x had parsing responsibilities split between the parser and the node. But nodes should be "dumb" and not know anything about how they are parsed - they should only know the bare minimum needed for rendering. As a result, 2.x has delegated the parsing responsibilities to two different interfaces: | Responsibility | Old Method (1.x) | New Method (2.0) | | --------------------------------------------------------- | ------------------------------------------ | -------------------------------------------------------------- | | Identifying the start of a block | `BlockParserInterface::parse()` | `BlockStartParserInterface::tryStart()` | | Instantiating and configuring the new block | `BlockParserInterface::parse()` | `BlockContinueParserInterface::__construct()` | | Determining if the block acts as a container | `AbstractBlock::isContainer()` | `BlockContinueParserInterface::isContainer()` | | Determining if the block can have lazy continuation lines | `AbstractBlock::isCode()` | `BlockContinueParserInterface::canHaveLazyContinuationLines()` | | Determining if the block can contain certain child blocks | `AbstractBlock::canContain()` | `BlockContinueParserInterface::canContain()` | | Determining if the block continues on the next line | `AbstractBlock::matchesNextLine()` | `BlockContinueParserInterface::tryContinue()` | | Adding the next line to the block | `AbstractBlock::handleRemainingContents()` | `BlockContinueParserInterface::addLine()` | | Finalizing the block and its contents | `AbstractBlock::finalize()` | `BlockContinueParserInterface::closeBlock()` | As a result of making this change, the `addBlockParser()` method on `ConfigurableEnvironmentInterface` has changed to `addBlockStartParser()`. See [the block parsing documentation](/2.0/customization/block-parsing/) for more information on this new approach. ## New Inline Parsing Approach The `getCharacters()` method on `InlineParserInterface` has been replaced with a more-robust `getMatchDefinition()` method which allows your parser to match against more than just single characters. All custom inline parsers will need to change to this new approach. Additionally, when the `parse()` method is called, the Cursor is no longer proactively advanced past the matching character/start position for you. You'll need to advance this yourself. However, the `InlineParserContext` now provides the fully-matched text and its length, allowing you to easily `advanceBy()` the cursor without having to do an expensive `$cursor->match()` yourself which is a nice performance optimization. See [the inline parsing documentation](/2.0/customization/inline-parsing/) for more information on this new approach. ## Rendering Changes This library no longer differentiates between block renderers and inline renderers - everything now uses "node renderers" which allow us to have a unified approach to rendering! As a result, the following changes were made, which you may need to change in your custom extensions: | Old Method/Interface (1.x) | New Method/Interface (2.0) | | ------------------------------------------------------- | ------------------------------------------------- | | `BlockRendererInterface` | `NodeRendererInterface` | | `InlineRendererInterface` | `NodeRendererInterface` | | `EnvironmentInterface::getBlockRenderersForClass()` | `EnvironmentInterface::getRenderersForClass()` | | `EnvironmentInterface::getInlineRenderersForClass()` | `EnvironmentInterface::getRenderersForClass()` | | `ConfigurableEnvironmentInterface::addBlockRenderer()` | `ConfigurableEnvironmentInterface::addRenderer()` | | `ConfigurableEnvironmentInterface::addInlineRenderer()` | `ConfigurableEnvironmentInterface::addRenderer()` | | `ElementRendererInterface::renderBlock()` | `ChildNodeRendererInterface::renderNodes()` | | `ElementRendererInterface::renderBlocks()` | `ChildNodeRendererInterface::renderNodes()` | | `ElementRendererInterface::renderInline()` | `ChildNodeRendererInterface::renderNodes()` | | `ElementRendererInterface::renderInlines()` | `ChildNodeRendererInterface::renderNodes()` | | `HtmlRenderer::renderBlock($document)` | `HtmlRenderer::renderDocument()` | Renderers now implement the unified `NodeRendererInterface` which has a similar (but slightly different) signature from the old `BlockRendererInterface` and `InlineRendererInterface` interfaces: ```php /** * @param Node $node * @param ChildNodeRendererInterface $childRenderer * * @return HtmlElement|string|null */ public function render(Node $node, ChildNodeRendererInterface $childRenderer); ``` The element being rendered is still passed in the first argument, and the object that helps you render children is still passed in the second argument. Note that blocks are no longer told whether they're being rendered in a tight list - if you need to know about this, traverse up the `$node` AST yourself and check any `ListBlock` ancestor for tightness. ## AST Node Changes The `AbstractBlock::$data` and `AbstractInline::$data` arrays were replaced with a `Data` array-like object on the base `Node` class. ## Removed Classes The following classes have been removed: | Class name in 1.x | Replacement / Notes | | ------------------------------ | ------------------------------------------------------------------------------------------------------------- | | `AbstractStringContainerBlock` | Use `extends AbstractBlock implements StringContainerInterface` instead. Note the new method names. | | `Context` | Use `MarkdownParserState` instead (has different methods but serves a similar purpose) | | `ContextInterface` | Use `MarkdownParserStateInterface` instead (has different methods but serves a similar purpose) | | `Converter` | Use `MarkdownConverter` instead. | | `ConverterInterface` | Use `MarkdownConverterInterface`. This interface has the same methods so it should be a drop-in replacement. | | `UnmatchedBlockCloser` | No longer needed 2.x | ## Renamed constants The following constants have been moved/renamed: | Old Name/Location (1.x) | New Name/Location (2.0) | | ----------------------------------------- | ------------------------ | | `EnvironmentInterface::HTML_INPUT_ALLOW` | `HtmlFilter::ALLOW` | | `EnvironmentInterface::HTML_INPUT_ESCAPE` | `HtmlFilter::ESCAPE` | | `EnvironmentInterface::HTML_INPUT_STRIP` | `HtmlFilter::STRIP` | | `TableCell::TYPE_HEAD` | `TableCell::TYPE_HEADER` | | `TableCell::TYPE_BODY` | `TableCell::TYPE_DATA` | ## Renamed Methods The following methods have been renamed: | Class | Old Name (1.x) | New Name (2.0) | | -------------------------------------------------- | ------------------ | ----------------------- | | `Environment` / `ConfigurableEnvironmentInterface` | `addBlockParser()` | `addBlockStartParser()` | | `ReferenceMap` / `ReferenceMapInterface` | `addReference()` | `add()` | | `ReferenceMap` / `ReferenceMapInterface` | `getReference()` | `get()` | | `ReferenceMap` / `ReferenceMapInterface` | `listReferences()` | `getIterator()` | ## Visibility Changes The following properties have had their visibility changed: | Property | Was (1.x) | Is Now (2.0) | Notes | | ---------------------- | --------- | ------------ | ----------------------------------------- | | `TableCell::$align` | `public` | `private` | Use `getAlign()` and `setAlign()` instead | | `TableCell::$type` | `public` | `private` | Use `getType()` and `setType()` instead | | `TableSection::$type` | `public` | `private` | Use `getType()` instead | ## Configuration Method Changes Calling `EnvironmentInterface::getConfig()` without any parameters is no longer supported. Calling `ConfigurableEnvironmentInterface::mergeConfig()` without any parameters is no longer supported. The `ConfigurableEnvironmentInterface::setConfig()` method has been removed. Use `getConfig()` instead. ## New approach to the `ReferenceParser` The `ReferenceParser` class in 1.x worked on complete paragraphs of text. This has been changed in 2.x to work in a more-gradual fashion, where parsing is done on-the-fly as new lines are added. Whereas you may have previously called `parse()` on a `Cursor` once on something containing multiple lines, you should now call `parse()` on each line of text and then later call `getReferences()` to check what has been parsed. ## `Html5Entities` utility class removed Use the `Html5EntityDecoder` utility class instead. ## `bin/commonmark` command This command was buggy to test and was relatively unpopular, so it has been removed. If you need this type of functionality, consider writing your own script with a Converter/Environment configured exactly how you want it. ## `CommonMarkConverter::VERSION` constant This previously-deprecated constant was removed in 2.0 Use `\Composer\InstalledVersions` provided by composer-runtime-api instead. ## `HeadingPermalinkRenderer::DEFAULT_INNER_CONTENTS` constant This previously-deprecated constant was removed in 2.0. Use `HeadingPermalinkRenderer::DEFAULT_SYMBOL` instead. ## `ArrayCollection` changes Several methods were removed from this class - here are the methods along with possible alternatives you can switch to: | Removed Method Name | Alternative | | ------------------- | ---------------------------------------------------- | | `add($value)` | `$collection[] = $value` | | `set($key, $value)` | `$collection[$key] = $value` | | `get($key)` | `$collection[$key]` | | `remove($key)` | `unset($collection[$key])` | | `isEmpty()` | `count($collection) === 0` | | `contains($value)` | `in_array($value, $collection->toArray(), true)` | | `indexOf($value)` | `array_search($value, $collection->toArray(), true)` | | `containsKey($key)` | `isset($collection[$key])` | | `replaceWith()` | (none provided) | | `removeGaps()` | (none provided) | This class is also `final` now, so don't extend it. ## `final` classes The following classes are now marked `final` and cannot be extended: - `ArrayCollection` - `Emphasis` - `FencedCode` - `Heading` - `HtmlBlock` - `HtmlElement` - `HtmlInline` - `IndentedCode` - `Newline` - `Strikethrough` - `Strong` - `Text` ## Node setter methods return void All `set*()` methods on all Node types now return `void` (whereas some used to return `$this` in 1.x) for consistency. ## Unused methods The following unused methods have been removed: - `Delimiter::setCanClose()` ## Slug Normalizer Need to generate unique slugs in your extensions? Use the new [Slug Normalizer](/2.0/customization/slug-normalizer/) provided by the `Environment`. ## Text Normalizers The second argument to `TextNormalizerInterface::normalize()` used to allow any arbitrary object. This was changed to an `array` so that multiple things can be passed in at once. ## `title` attribute for `Link` and `Image` nodes The `title` attribute for `Link` and `Image` nodes is now stored using a dedicated property instead of stashing it in `$data`. Use `getTitle()` and `setTitle()` to access the value. ## Node Iteration In 1.x, most custom code used `$node->walker()` to iterate the AST. Although this still exists, consider whether your code could use `$node->iterator()` instead which can be up to twice as fast!