VERSION-TODO.md

flexmark-java

 

Version History

• Release 0.60.0
  • API Refactoring
• Next 0.61.xx
• 0.62.2
• 0.62.0
• 0.61.34
• 0.61.32
• 0.61.30
• 0.61.28
• 0.61.26
• 0.61.24
• 0.61.22
• 0.61.20
• 0.61.18
• 0.61.16
• 0.61.14
• 0.61.12
• 0.61.10
• 0.61.8
• 0.61.6
• 0.61.4
• 0.61.2
• 0.61.0
• 0.60.2
• 0.60.0
• 0.59.124
• 0.59.122
• 0.59.120
• 0.59.118
• 0.59.116
• 0.59.114
• 0.59.112
• 0.59.110
• 0.59.108
• 0.59.106
• 0.59.104
• 0.59.102
• 0.59.100
• 0.59.98
• 0.59.96
• 0.59.94
• 0.59.92
• 0.59.90
• 0.59.88
• 0.59.86
• 0.59.84
• 0.59.82
• 0.59.80
• 0.59.78
• 0.59.76
• 0.59.74
• 0.59.72
• 0.59.70
• 0.59.68
• 0.59.66
• 0.59.64
• 0.59.62
• 0.59.60
• 0.59.58
• 0.59.56
• 0.59.54
• 0.59.52
• 0.59.50
• 0.59.48
• 0.59.46
• 0.59.44
• 0.59.42
• 0.59.40
• 0.59.38
• 0.59.36
• 0.59.34
• 0.59.32
• 0.59.30
• 0.59.28
• 0.59.27
• 0.59.25
• 0.59.23
• 0.59.21
• 0.59.19
• 0.59.17
• 0.59.15
• 0.59.13
• 0.59.11
• 0.59.9
• 0.59.7
• 0.59.5
• 0.59.1
• 0.50.48
• 0.50.46
• 0.50.44
• 0.50.42
• 0.50.40
• 0.50.38
• 0.50.36
• 0.50.34
• 0.50.32
• 0.50.30
• 0.50.28
• 0.50.26
• 0.50.24
• 0.50.22
• 0.50.20
• 0.50.18
• 0.50.16
• 0.42.14
• 0.42.12
• 0.42.10
• 0.42.8
• 0.42.6
• 0.42.4
• 0.42.2
• 0.42.0
• 0.40.34
• 0.40.32
• 0.40.30
• 0.40.28
• 0.40.26
• 0.40.24
• 0.40.22
• 0.40.20
• 0.40.18
• 0.40.16
• 0.40.14
• 0.40.12
• 0.40.10
• 0.40.8
• 0.40.6
• 0.40.4
• 0.40.2
• 0.40.0

 

Release 0.60.0

API Refactoring

⚠️ Upcoming release of 0.60.0 will have some breaking changes due to re-organization, renaming and clean up of some implementation classes.

Please give feedback on the upcoming changes if you have concerns about breaking your code:

• Break: split out generic AST utilities from flexmark-util module into separate smaller modules. IntelliJ IDEA migration to help with migration from 0.50.40 will be provided where needed if the package or class is changed. com.vladsch.flexmark.util will no longer contain any files but will contain the separate utilities modules with flexmark-utils module being an aggregate of all utilities modules, similar to flexmark-all

  • ast/ classes to flexmark-util-ast
  • builder/ classes to flexmark-util-builder
  • collection/ classes to flexmark-util-collection
  • data/ classes to flexmark-util-data
  • dependency/ classes to flexmark-util-dependency
  • format/ classes to flexmark-util-format
  • html/ classes to flexmark-util-html
  • mappers/ classes to flexmark-util-sequence
  • options/ classes to flexmark-util-options
  • sequence/ classes to flexmark-util-sequence
  • visitor/ classes to flexmark-util-visitor

• Convert anonymous classes to lambda where possible.

• refactor flexmark-util to eliminate dependency cycles between classes in different subdirectories.

• Break: delete deprecated properties, methods and classes

• Add: org.jetbrains:annotations:15.0 dependency to have @Nullable/@NotNull annotations added for all parameters. I use IntelliJ IDEA for development and it helps to have these annotations for analysis of potential problems and use with Kotlin.

• Break: refactor and cleanup tests to eliminate duplicated code and allow easier reuse of test cases with spec example data.

• Break: move formatter tests to flexmark-core-test module to allow sharing of formatter base classes in extensions without causing dependency cycles in formatter module.

• Break: move formatter module into flexmark core. this module is almost always included anyway because most extension have a dependency on formatter for their custom formatting implementations. Having it as part of the core allows relying on its functionality in all modules.

• Break: move com.vladsch.flexmark.spec and com.vladsch.flexmark.util in flexmark-test-util to com.vladsch.flexmark.test.spec and com.vladsch.flexmark.test.util respectively to respect the naming convention between modules and their packages.

• Break: NodeVisitor implementation details have changed. If you were overriding NodeVisitor.visit(Node) in the previous version it is now final to ensure compile time error is generated. You will need to change your implementation. See comment in the class for instructions.

  ℹ️ com.vladsch.flexmark.util.ast.Visitor is only needed for implementation of NodeVisitor and VisitHandler. If you convert all anonymous implementations of VisitHandler to lambdas you can remove all imports for Visitor.

  • Fix: remove old visitor like adapters and implement ones based on generic classes not linked to flexmark AST node.
  • Deprecate old base classes:
    • com.vladsch.flexmark.util.ast.NodeAdaptedVisitor see javadoc for class
    • com.vladsch.flexmark.util.ast.NodeAdaptingVisitHandler
    • com.vladsch.flexmark.util.ast.NodeAdaptingVisitor

Next 0.61.xx

• Fix: replace PrefixedSubSequence and SegmentedSequence.create() with non builder arguments to use SequenceBuilder throughout library code, except where PrefixedSubSequence is used as is without immediately wrapping in SegmentedSequence. This happens anyway in SegmentedSequence.create for non-builder arg, should make it explicit.
• Fix: rewrite LineAppendableImpl to be compatible with SequenceBuilder
  • optimize by not processing one char at a time. Split the sequence into regions of interest and process the regions as one piece which the SequenceBuilder can optimize to base sequence quicker.
  • do not construct temporary StringBuilder but only keep the last line under construction temporary, with all previously constructed lines used for prior content.
  • use separate instance of sequence builder for the line under construction. SegmentedSequence construction is fast enough to not need a parallel string builder for context tests.
• Add: dependency resolution for Node Renderer Factories to make them independent of extension add order

• Fix: Change spec example to variable number of sections
• Add: yaml front matter configurator for modules. See: Yaml Front Matter Configuration
• Fix: table formatting caret position tracking affects alignment even when not inserting, just formatting. Need to keep track of whether format after typing or just format. Then caret position should only track but not force spaces behind it to be preserved. See failed tests in com.vladsch.flexmark.ext.tables.MarkdownTableTest
• Add: spec example language per section options and rendering in HTML.
  • Add: Tests for section name options
• Fix: Document docx form controls in wiki.
• Add: table recovery utility to restore table after it was wrapped as plain text paragraph. Needed when formatting is applied without a table extension. If not noticed will result in a table which is tedious to recover manually. Handle only tables with leading and trailing | for each line that was wrapped. Otherwise, it is impossible to tell where each line ends and another begins.
• Fix: Html converter to not add spaces between end of inline marker and next punctuation: .,:;

0.62.2

• Fix: merge Parsing optimization by using static patterns
• Fix: add a few more Parsing optimization by using static patterns for flag based dynamic values.
• Fix: add cache for dynamic Parsing patterns based on dependent options.

0.62.0

• Fix: HTML to Md converter to convert empty a tags to empty link text links. ie. <a href="#abc"></a> to [](#abc), except when contained in heading elements.
• Break: rename Attributes to MutableAttributes and create new immutable Attributes class.
• Fix: do not remove title attribute provided by LinkResolver when link has no title. Allow link resolver titles to be used.

0.61.34

• Fix: add DocxRenderer.PAGE_SIZE and DocxRenderer.PAGE_LANDSCAPE properties to allow simple changes to page setup for rendered document which is different from one in the template.

0.61.32

• Fix: optimized TOC levels from erroneous #,# to #-# for single level TOC.

0.61.30

• Fix: #402, When using PegdownOptionsAdapter indented code blocks does not work for lists

0.61.28

• Fix: Toc and SimToc levels parsing to properly handle levels as per documentation.

  • single numeric >2 is 2-level, ie. levels=4 should be interpreted as levels=2-4 but was levels=4,4.
  • levels=x- to be same as levels=x-6, was handled as levels=x,x

  levels=levelList where level list is a comma separated list of levels or ranges. Default is to include heading levels 2 and 3. If a range is specified using - as separator between numeric values then the included range is from first numeric to last numeric inclusive. If first numeric is missing then 1 will be used in its place. If last numeric is missing then 6 will be used in its place.

  Examples:

  • levels=1 include level 1 only, same as levels=1,1
  • levels=2 include level 2 only, same as levels=2,2
  • levels=3 include levels 2 and 3
  • levels=4 include levels 2,3 and 4
  • levels=2-4 include levels 2,3 and 4. same as levels=4
  • levels=2-4,5 include levels 2,3,4 and 5
  • levels=1,3 include levels 1 and 3
  • levels=-3 include levels 1 to 3
  • levels=3- include levels 3 to 6

0.61.26

• Fix: SequenceBuilder.toString() not to add space between sequence parts.

• Break: TextCollectingVisitor needs TextContainer.F_ADD_SPACES_BETWEEN_NODES to ensure there is at least one space between texts of different nodes in collected text. Previously this was added automatically by sequence builder. Use one of the getAndCollect functions taking options flags and pass TextContainer.F_ADD_SPACES_BETWEEN_NODES.

• Fix: MarkdownParagraph wrapping with preserving tracked offsets with spaces to preserve spaces right before last non-blank character.

• Fix: unify handling of link generating inline element handling in link text elements.

  • Add: LinkRendered interface to identify link rendering elements derived from link/image ref syntax identified by implementation of LinkRefDerived interface
  • Fix: WikiNode to implement LinkRefDerived
  • Fix: WikiLink to implement LinkRendered
  • Fix: Footnote to implement LinkRendered

  ℹ️ nested link ref derived elements in link ref text have priority and cannot be used to embed into link ref text using [link ref text][ref id] syntax.

0.61.24

• Fix: link refs in link text should be collapsed.

  • Add: Parser.LINK_TEXT_PRIORITY_OVER_LINK_REF, default false, commonmark spec does not specify how to parse:

    [[moon]](/uri)
    
    [moon]: moon.jpg
    

    with default implementation priority given to contained ref, renders as:

    <p>[<a href="moon.jpg">moon</a>](/uri)</p>
    

    when set to true, priority given to outer link and will render as:

    <p><a href="/uri">[][moon]</a></p>
    

    ℹ️ Undefined reference links are always treated as text when in a link text element.

0.61.22

• Fix: #407, Link text inline content fails to parse image references, link and image refs in link text would be collapsed even when they are not tentative but defined.

0.61.20

• Fix: JekyllTagExtension handling of inline include tag. If included file contains a single paragraph then the paragraph is unwrapped so only the text is included. Otherwise, the included file AST is included as is, which may result in invalid HTML if included file has block elements since these will be included in the p tag of including element.

  Break: ℹ️ when include tag embedding is enabled, the included file elements will be added as child nodes of the JekyllTag node, not JekyllTagBlock node as was done in previous versions.

0.61.18

• Break: remove from TocOptions confusing and error prone rawTitle and rawTitleLevel. Now mutable toc options have to set title and titleLevel then call TocOptions.AsMutable.normalizeTitle()
• Fix: add more debug trace information to markdown wrap tracked offset mismatch.
• Add: UriContentResolver extension for unified content handling in parser and renderers by extensions.
  • Add: JekyllTagExtension.CONTENT_RESOLVER_FACTORIES, default Collections.emptyList(). If this option is an empty list then a singleton list of FileUriContentResolver.Factory instance will be used.
  • Fix: IncludeNodePostProcessor to use content resolvers for getting file content instead of getting file directly.
  • Add: DocxRenderer.DEFAULT_CONTENT_RESOLVER, default true, set to false not to add FileUriContentResolver to content resolvers factories.
  • Add: DocxRendererContext.resolvedContent(ResolvedLink) to allow generic URI to content resolution.
  • Fix: CoreNodeDocxRenderer to use content resolvers for resolving image links, only resolving http:// and https:// links if content is not resolved.

0.61.16

• Fix: add range limit to Escaping.replaceAll(Pattern, BasedSequence, List<Range>, Replacer, ReplacedTextMapper) when setting matcher range.
• Add: DocxRenderer.CUSTOM_PROPERTIES, default Collections.emptyMap(), if present then custom properties of the document will be set.

0.61.14

• Fix: #398, Autolinks get cut off if they contain `&` (escaped query params) for multiple html entities per link.

0.61.12

• Fix: update docx4j to 8.1.6
• Fix: docx rendering tests for full spec example
• Add: JekyllTagExtension.EMBED_INCLUDED_CONTENT, default false. Set to true to embed included markdown content.
• Fix: tracked paragraph wrapping when inserting space before EOL on a line when the word before caret will be wrapped to the next line. ie. the EOL will be removed.
• Deprecate: JekyllTagExtension.ENABLE_RENDERING, not used nor needed.
• Fix: #398, Autolinks get cut off if they contain `&` (escaped query params)

0.61.10

• Add: HtmlRenderer.FENCED_CODE_LANGUAGE_DELIMITERS, default " \t", to configure which chars terminate the language part of info string
• Add: HtmlRenderer.FENCED_CODE_LANGUAGE_CLASS_MAP, default new HashMap(), to provide individual language to class mapping string. If language string is not in the map then HtmlRenderer.FENCED_CODE_LANGUAGE_CLASS_PREFIX + language will be used. Intended to allow creating alias mapping and per-language class configuration.

0.61.8

• Fix: wrong assert conditions in LineAppendableImpl.setPrefixLength(int, int)
• Fix: assert condition for MarkdownParagraph.resolveTrackedOffsetsEdit, space and non-break space should be considered a match.
• Fix: MarkdownParagraph.resolveTrackedOffsetsEdit, space and non-break space should be considered a match.
• Add: details to assert for SegmentTree.findSegmentPos(int, int[], int, int) to allow diagnostics in reports

0.61.6

• Add: HtmlRenderer.HEADER_ID_REF_TEXT_TRIM_LEADING_SPACES default true and HtmlRenderer.HEADER_ID_REF_TEXT_TRIM_TRAILING_SPACES default true. When false then corresponding reference link text in heading text used for generating heading id is not trimmed.
• Add: HtmlRenderer.HEADER_ID_ADD_EMOJI_SHORTCUT default false When true then emjoi nodes add emoji shortcut to collected text for heading id.
• Fix: ParserEmulationProfile.GITHUB
  • Set HtmlRenderer.HEADER_ID_REF_TEXT_TRIM_TRAILING_SPACES to false because GitHub does not trim trailing spaces for generating heading ids.
  • Emoji node to generate text for heading id consisting of the emoji shortcut, for GitHub compatibility.
  • Fix: ImageRef not to add text to heading text for heading id generation
• Fix: add more information to MarkdownParagraph.resolveTrackedOffsetsEdit assert failures to allow better diagnostics in reported stack traces.
• Fix: disable wrapping of multiline URL image links
• Fix: Paragraph content node to preserve leading spaces on lines. Otherwise, multiline URL image content would loose indents since these were stripped from paragraph content during parsing.

  • Fix: TablesExtension to accept non-indenting leading spaces on all table rows. This results in previously non-table text as valid table texts:

    Abc|Def
      |---|---

• Fix: LinkRef and ImageRef changing spaces to   for wrap formatting when not part of a Paragraph, ie. in heading text.

0.61.4

• Fix: merge #397, PR: Add base64 image support with docx rendering thanks to @Xaelis
  • Add: test for base64 encoded docx conversion
• Break: WikiLinkLinkResolver to take WikiLinkExtension.ALLOW_ANCHORS and WikiLinkExtension.ALLOW_ANCHOR_ESCAPE into account when extracting page ref from link.
  • Fix: # or \ included in the URL of the resolved link are now URL encoded.
  • Fix: default WikiLinkLinkResolver handles its own text unescaping. When resolving wiki
    link with default resolver do not unescape. If replacing default resolver ensure you unescape the text. This is needed to properly handle anchor escaping.
  • Fix: PegdownOptionsAdapter to set WikiLinkExtension.ALLOW_ANCHORS to true for pegdown compatibility
• Fix: deprecate CoreNodeFormatter.ATTRIBUTE_UNIQUIFICATION_ID_MAP and move declaration to Formatter.ATTRIBUTE_UNIQUIFICATION_ID_MAP
• Fix: deprecate CoreNodeFormatter.UNIQUIFICATION_ID_MAP and move declaration to Formatter.UNIQUIFICATION_ID_MAP
• Fix: wiki links should not be wrapped during formatting.
  • Add: WikiLink formatter extension and tests
  • Add: WikiLink translating formatter functionality and tests

0.61.2

• Fix: ScopedDataSet.getAll() and ScopedDataSet.getKeys() would not return keys from parent scope.
• Fix: #396, DocumentParser stops reading too early resulting in the document being cut off

0.61.0

• Fix: enable original spec tests

• Fix: merge #391, PR: Fix: CRLF line separator in fenced code blocks produce redundant CR.

• Fix: merge #387, JUnit is in the compile scope

• Add: ability to insert anchor ref targets to HtmlIdGenerator when scanning a document.

• Add: add LinkResolverBasicContext as base interface of LinkResolverContext to allow simple context to pass to LinkResolver when resolving links without needing to implement all unused methods of LinkResolverContext.

• Break: change argument to LinkResolver and LinkResolverFactory to LinkResolverBasicContext. Except for JekyllTagExtension, the argument is still LinkResolverContext and passed value can be cast to this if required.

• Add: support for JekyllTagBlock with include directive to DocxRenderer

  Use {% include includeFile %} to include the file into the source document. includeFile is resolved relative to DocxRenderer.DOC_RELATIVE_URL or DocxRenderer.DOC_ROOT_URL. The latter is used if includeFile starts with /

  Will only include files if JekyllTagExtension.LINK_RESOLVER_FACTORIES is not empty, in which case the link resolvers will be used to resolve includeFile to full file path. Use singleton list of DocxLinkResolver.Factory instance to resolve using doc relative and root url paths.

• Fix: MergeLinkResolver and DocxLinkResolver

0.60.2

• Fix: change formatter sub-context rendering node to nullable

0.60.0

• Fix: rename BaseSequence.extendToAny to BaseSequence.extendByAnyNot which better reflects the operation of the method.
• Deprecate: BaseSequence.extendToAny
• Add: FlexmarkHtmlConverter div to table options to allow converting div based table code to markdown tables.
• Fix: move BaseSequenceManager and related to flexmark-util-experimental since these are not used.

0.59.124

• Add: migrate flexmark-java 0_50_x to 0_60_0.xml IntelliJ Migration
• Fix: rename abbreviation FormatOptions AbbreviationFormatOptions
• Fix: rename definition FormatOptions DefinitionFormatOptions
• Fix: rename task list FormatOptions TaskListFormatOptions
• Fix: formatter core test for invalid indentation
• Fix: formatter EqualizeTrailingMarker.EQUALIZE adding marker when there is none.
• Fix: missing TextContainer implementation by TypographicSmarts
• Fix: restore IRenderBase.NullRenderer and IRenderBase.TextRenderer static classes

0.59.122

• Fix: cleanup and simplify dependency handler use. Too convoluted in the current implementation.

0.59.120

• Fix: Formatter not inserting blank line after paragraph when followed by list sub-item in tighten mode.

0.59.118

• Fix: split flexmark-util module sub-directories into separate modules
  • ast/ classes to flexmark-util-ast
  • builder/ classes to flexmark-util-builder
  • collection/ classes to flexmark-util-collection
  • data/ classes to flexmark-util-data
  • dependency/ classes to flexmark-util-dependency
  • format/ classes to flexmark-util-format
  • html/ classes to flexmark-util-html
  • mappers/ classes to flexmark-util-sequence
  • options/ classes to flexmark-util-options
  • sequence/ classes to flexmark-util-sequence
  • visitor/ classes to flexmark-util-visitor

0.59.116

• Fix: add missing semi in convert to visual test functions HTML entities.
• Fix: remove static char predicate fields from SequenceUtils
• Deprecate: static char predicate fields in SequenceUtils

0.59.114

• Fix: change docx converter to not include <> around auto and mail links in generated code.
• Fix: #384, Markdown parser produces invalid HTML
• Fix: render ref link and image as text if link resolved to empty URL
• Fix: formatter to not output tail blank line after fenced code if it is last child of list item and current list spacing is not LOOSE.
• Fix: data key toString() to use Class.getSimpleName()

0.59.112

• Fix: MarkdownTable not prefixing caption line with requested line prefix.

0.59.110

• Add: IRichSequence.padding to return the required padding sequence.
• Fix: IRichSequence.padStart was not padding the sequence but only returned the required padding.

0.59.108

• Fix: Change the meaning of ListSpacing.LOOSEN and ListSpacing.TIGHTEN for formatter with Parser.BLANK_LINES_IN_AST set to true:

  • ListSpacing.LOOSEN adds blank lines only if list is loose because it has items in the list followed by blank lines.

    ⚠️ This will work as expected only if Parser.LISTS_AUTO_LOOSE is true. Otherwise, a blank line will only be inserted if there is already a blank line after the item, same as selecting ListSpacing.AS_IS.

  • ListSpacing.TIGHTEN removes blank lines which are not significant in determining whether the list is loose so as to make the list as tight as possible.

• Add: ParagraphItemContainer.isParagraphInTightListItem(Paragraph) to allow testing for paragraph items not wrapped without having list options.
• Fix: remove debug exception in MarkdownTable when an offset is not found in tracked offsets.
• Fix: clean up output to stdout in tests for MarkdownParagraph

0.59.106

• Fix: LineAppendableImpl.getLines() when requested with maxTrailingBlankLines of -1 then last line should not have and EOL.
• Fix: SimToc formatting:
  • allow empty title string in parser
  • blank title should not add TOC heading
  • empty title string should be converted to empty string and not removed
  • change no title TocOptions to be empty title and if an empty title string is specified then title string will be blank (" ").
• Break: TocUtils.renderMarkdownToc(MarkdownWriter, List<Integer>, List<String>, TocOptions) and TocUtils.renderTocContent(MarkdownWriter, TocOptions, TocOptions, List<Heading>, List<String>) to not have Heading arguments but only relevant heading information to allow reuse.

0.59.104

• Fix: #382, Is there an option for number of whitespaces needed to create sub-lists?
  • FIXED_INDENT list parser did not convert list item looking text with >= 4 spaces to lazy continuation.

0.59.102

• Break: Add validation to parser that passed BasedSequence is contiguous with no out of base parts and throw descriptive IllegalArgumentException if it is not. Otherwise, it is possible to provide sequences which will cause exceptions during parser.
• Fix: re-implement markdown paragraph tracked offset resolution by scanning the unwrapped and wrapped text segments and mapping them to original sequence offsets while analyzing text context and keeping track of: start of line, first non-blank, last non-blank, end of line.
  • Fix: implement generic code for resolving index in wrapped sequence based on tracking offset in base sequence without inexplicable fudge factor adjustments.
• Fix: validate that MarkdownParagraph can properly compute spaces before/after at offset from altSeq.getBaseSequence() and remove these fields from TrackedOffset.create.
• Fix: LineAppendableImpl.appendTo would trim spaces on blank lines. Lines should be appended as is since formatting is done during accumulation not when output to sequence or appendable.
• Break: invert LineAppendable.F_ALLOW_LEADING_EOL to LineAppendable.F_TRIM_LEADING_EOL to ensure that flags of 0 do no trimming.
  • Fix: Formatter.FORMAT_FLAGS to add LineAppendable.F_TRIM_LEADING_EOL to defaults used for formatting if none specified.
• Fix: LineAppendableImpl empty forward iterator had next instead of none.
• Add: TrackedOffset spaces before/after the offset. This is impossible to determine from segmented sequence in edge cases. Best to have it provided from original sequence where it is easy to determine and combined with what is in the final sequence since some spaces could have been removed when parent prefixes are removed.
• Fix: removal of all prefixes messes up format tracked offset resolution
• Fix: SegmentTree.getSegmentRange(int, int, int, int, BasedSequence, Segment) was not computing startOffset correctly if first segment was text and had no previous ANCHOR segment.
  • Fix: SegmentedSequenceTree.subSequence() was not computing correct start offset due to SegmentTree bug.
• Add: Formatter.DOCUMENT_FIRST_PREFIX and Formatter.DOCUMENT_PREFIX to allow setting prefixes for the whole document.
• Add: SequenceBuilder treating of any based sequence whose base is equal in characters to its own as equivalent since offsets will match.
• Fix: LineAppendable.isEmpty() return true only if no lines or pending lines
• Fix: SubSequence constructor to only use the base CharSequence hash code if it is String
• Add: Formatter.TRACKED_SEQUENCE option to allow formatting of document based on SegmentedSequence which cannot be used for parsing without converting to contiguous sequence. Used to construct offset tracker for tracked offset resolution.
  • Fix: Formatter paragraph wrapping to use trackedSequence.
  • Fix: tracked offset resolution to use trackedSequence.

0.59.100

• Fix: make all reference nodes compareTo() not case sensitive
• Fix: change back to jetbrains.annotations:15.0
• Fix: formatter to not change HTML block indentations by
• Break: rename HtmlFormattingAppendable to HtmlAppendable
• Break: rename HtmlFormattingAppendableBase to HtmlAppendableBase
• Add: LineInfo to have the line CharSequence
• Add: to LineAppendable
  • extend Iterable<LineInfo>
  • Add: LineAppendable.append(LineAppendable, int, int, boolean) to append another line appendable without prefixes
  • Add: LineAppendable.copyAppendable(LineAppendable, int, int, boolean) to make a copy and append selected content with or without prefixes
  • Add: LineAppendable.insertLine(int, CharSequence, CharSequence) to insert lines

0.59.98

• Fix: create TestSuite for each util module subdirectory
• Fix: change UtilTestSuite to run other module test suites.
• Fix: refactor flexmark-util to allow separate modules:
  • visitor has no dependencies
  • misc: classes contained in util directory to misc, has no dependencies
  • collection depends on misc only
  • dependency depends on misc, collection
  • data depends on misc only
  • builder depends on misc, data
  • sequence depends on misc, collection, data
  • options depends on misc, sequence
  • html depends on misc, sequence
  • ast depends on misc, collection, data, sequence
  • format depends on misc, collection, data, sequence, html, ast
• Fix: refactor flexmark-util to eliminate cyclic dependencies between future util modules
  • Break: delete deprecated classes contained in mappers
    • LowerCaseMapper
    • UpperCaseMapper
  • Break: move classes contained in mappers directory to sequence/mappers subdirectory
  • Break: move classes contained in util directory:
    • SharedDataKeys to data
    • the rest to misc
  • Break: move from html to sequence to eliminate dependency cycle
    • Escaping
    • Html5Entities
    • LineAppendable
    • LineAppendableImpl
  • Break: move from builder to misc to eliminate dependency cycle
    • Extension
  • Break: move from sequencetomisc` to eliminate dependency cycle
    • CharPredicate
  • Break: move from Utils to BasedUtils to eliminate dependency cycle
    • asBased
  • Break: move from Utils to SharedDataKeys to eliminate dependency cycle
    • runningTests
  • Break: move from Utils to SequenceUtils to eliminate dependency cycle
    • parseUnsignedIntOrNull
    • parseUnsignedIntOrNull
    • parseIntOrNull
    • parseIntOrNull
    • parseLongOrNull
    • parseLongOrNull
    • parseUnsignedIntOrDefault
    • parseUnsignedIntOrDefault
    • parseIntOrDefault
    • parseIntOrDefault
    • parseNumberOrNull
    • parseNumberPrefixOrNull
• Fix: MarkdownParagraph tracked offset correction when deleting first non-blank char surrounded by spaces, index would be 0 instead of at first non-blank of wrapped with space inserted after index.

0.59.96

• Fix: disable wrapping of link and image text if wrapping enabled.
• Break: remove unused ContinuationIndent
• Break: remove Formatter.BLOCK_QUOTE_CONTINUATION_MARKERS and related. It was never properly implemented and even if it would be it can mess up the document for compound prefixes depending on selected settings.
• Fix: AttributesNodeFormatter adding quotes around implicit attribute values: # and . prefixed ones.
• Fix: FormatterUtils using node.getDocument() instead of context.getDocument()
• Add: TrackedOffsetUtils for common tracked offset resolving code
• Add: tracked offset option to HtmlRenderer, NOT TESTED, Experimental
• Break: remove kludged Formatter.render builder arguments and only use Appendable, pass builder since builder interface is extending appendable.
• Fix: LineAppendable
  • Break: replace ISequenceBuilder arguments with Appendable, if LineAppendable or ISequenceBuilder then will use its builder, otherwise will use StringSequenceBuilder
  • Add: LineAppendable.setPrefixLength(int, int) to change prefix/line content split for lines
  • Add: LineAppendable.setLine(int, CharSequence, CharSequence) to change prefix/line content split for lines
  • Add: appendTo without prefixes to allow extracting content only.
  • Add: LineAppendable.setLine(int, CharSequence, CharSequence) to set prefix and text of line.
  • Add: LineAppendable.setPrefixLength(int, int) to change the prefix/text index for the line without changing actual line content.
  • Break: LineAppendable.append(LineAppendable, int, int) does not add a line() call before appending, which means if there is an unterminated line, the first line of appended appendable will be added to it. This allows adding text before the first line of appendable.
  • Break: LineAppendable.line() and variants no longer allow adding a blank line at the top of the text if allow leading EOL is enabled. Use blankLine() for this.
  • Break: LineAppendable.blankLine() and variants no longer allow adding a blank line at the top of the text if allow leading EOL is not enabled.
• Break: clean up BitFieldSet:
  • method names, ones adding/removing enum fields are now add/remove, ones taking bit mask are orMask/andNotMask.
  • add getUnsigned() versions for int, short, byte getUInt(), getUShort() and getUByte() to get unsigned values of fields when they are shorter than the storage type. setUnsignedField() versions for all field types, signed ones are a PITA to deal with when encoding bit mask values. For signed values use the setBitField and getLong(), getInt(), getShort(), getByte().

0.59.94

• Fix: task bullet marker would not be changed if complete task changed to non-task during formatting.
• Fix: Spec test file formatting
  • not incrementing example count
  • Fix: TestUtils.addSpecExample() to add EOL after ast if not already terminated. This was causing spec example formatting to mess up.
  • Fix: add extra blank line after spec example
  • Fix: parser to not remove trailing EOLs in spec example sections
  • Fix: formatter to preserve all whitespace in spec examples
  • Fix: format options by spacing commas and removing any repeated commas
• Fix: Formatter tracked offset at EOL resolving for any offsets not handled by elements.

0.59.92

• Break: rename SpecReader.TYPE_TEST_BREAK and SpecReader.TYPE_BREAK to SpecReader.SECTION_TEST_BREAK and SpecReader.SECTION_BREAK respectively.
• Fix: LineAppendableImpl.appendTo and all context rendering methods flushTo() now take a separate arguments for maxBlankLines which applies to body of text. maxTrailingBlankLines which applies to trailing blank lines in appendable.
• Fix: LineAppendableImpl re-implement and clean up using a more consistent model for accumulating and manipulating lines of output.
  • Add: separate arguments for maxBlankLines which applies to body of text and maxTrailingBlankLines which applies to trailing blank lines in appendable.
  • Change: accumulation now done as full lines with information on prefix, and eol
• Add: TrackedOffsetList for efficient tracked offset location
• Fix: add formatter tracking when formatting first block element child of an empty list item to allow having block element on same line as list item prefix.
• Add: AsideExtension custom formatter
• Fix: formatter tracking offset adjustment when tracking offset is not in table or paragraph, ie. not resolved after formatting then resolve in Formatter.render when using builder.
• Add: SpecExampleExtension formatting for spec test files.
• Add: NodeFormatterContext.delegateRender() method to delegate to format handler registered after the current one. If none then will delegate to Node renderer, with CoreNodeFormatter.render(Node) being the last in line.
• Add: separate TextContainer.F_LINK_NODE_TEXT collection type to allow link element text to be collected separate from TextContainer.F_NODE_TEXT flag.
• Add: SpaceInsertingSequenceBuilder to track intervening spaces between nodes and insert on non-space.
• Fix: TextCollectingVisitor to add intervening spaces to match old implementation.
• Add: Formatter.TRACKED_OFFSETS default empty list, to provide which offsets into the base sequence being formatted should be tracked so their index can be resolved in the formatted result.
  • Add: tracked offset use to Paragraph formatting wrapping code
  • Add: tracked offset use to TableBlock formatting code

0.59.90

• Add: MarkdownTable.sorted(ColumnSort[], int, Predicate<String>) suffix testing predicate to allow sorting of columns with numeric prefix by numeric then by suffix if numeric part is equal.
• Add: MarkdownTable constructor param for character sequence of the table. Can be any sequence. For use by applications if they need access to original table character sequence.

0.59.88

• Fix: change RefNode text collection to resolve reference if link url collection is requested.

0.59.86

• Fix: replace old TextCollectingVisitor implementation with new one independent of specific nodes.
• Add: TextContainer interface to allow text collection without knowing implementation node class.
• Add: TextCollectingVisitor to utils which does not require specific node classes. Takes option flags:
  • F_LINK_TEXT - link text or image alt for link/image nodes
  • F_LINK_PAGE_REF - link page ref for link/image nodes
  • F_LINK_ANCHOR - link anchor for link/image nodes
  • F_LINK_URL - link url for link/image nodes
  • F_NODE_TEXT - collect node text without unescaping or percent url decoding
• Add: MarkdownTable.transposed(int) to get a transposed version of the table
• Add: MarkdownTable.sorted(ColumnSort[], int flags)) to get a sorted version of a table by given columns/sort type. Columns sorted by text extracted by TextCollectingVisitor using given collection flags
• Add: TableFormatOptions.FORMAT_TABLE_FILL_MISSING_MIN_COLUMN, default null, to control min column where filling of missing columns begins when formatting table.

0.59.84

• Fix: convert MarkdownTable to use common TrackedOffset for offset tracking API
• Fix: convert MarkdownParagraph to use common TrackedOffset for offset tracking API
• Fix: deprecate old addTrackedOffset methods.
• Fix: rename formatter HEADER options to HEADING
• Fix: deprecate old header named options
• Fix: delete unused formatter option KEEP_TRAILING_SPACES
• Fix: delete unused formatter option CODE_KEEP_TRAILING_SPACES

0.59.82

• Add: missing formatter options:
  • HeadingStyle
  • BLOCK_QUOTE_CONTINUATION_MARKERS
  • LIST_RESET_FIRST_ITEM_NUMBER applies if renumbering items
  • LIST_ALIGN_NUMERIC
  • flexmark-ext-attributes formatting of individual attributes instead of dumping the attributes node text.
    • FORMAT_ATTRIBUTES_COMBINE_CONSECUTIVE
    • FORMAT_SORT_ATTRIBUTES
    • FORMAT_ID_ATTRIBUTE, AttributeImplicitName enum, as is, implicit preferred, explicit preferred
    • FORMAT_CLASS_ATTRIBUTE, AttributeImplicitName enum, as is, implicit preferred, explicit preferred
    • FORMAT_ATTRIBUTES_SPACES
    • FORMAT_ATTRIBUTE_EQUAL_SPACE
    • FORMAT_ATTRIBUTE_VALUE_QUOTES
  • prioritized task items: flexmark-ext-gfm-tasklist
    • FORMAT_ORDERED_TASK_ITEM_PRIORITY: Integer, priority to use for parent item. Ordered task items do not not sort on priority since they are already ordered by number.
    • FORMAT_ORDERED_TASK_ITEM_PRIORITY: Integer, priority to use for parent item. Ordered task items do not not sort on priority since they are already ordered by number.
    • FORMAT_TASK_ITEM_PRIORITIES: int[], specifies integer priority for item marker characters corresponding to LISTS_ITEM_PREFIX_CHARS used for list items. Higher number is higher priority. A parent item's priority is the max priority of itself and any of its descendant task items. Non-task items do not have their own priority, only priority derived from their descendant items.
  • ElementPlacementSort
    • SORT_DELETE_UNUSED
    • DELETE_UNUSED
  • LINK_MARKER_COMMENT_PATTERN
  • CONTINUATION_ALIGNMENT
  • Add: Formatter control for controlling non-formatting regions.
    • <!-- @formatter:on --> and <!-- @formatter:off --> tags
    • FORMAT_CONTROL_ON: String
    • FORMAT_CONTROL_OFF: String
    • FORMAT_CONTROL_REGEX: Boolean to treat control on/off strings as regex pattern.
  • implement SimTocExtension.FORMAT_UPDATE_ON_FORMAT option
• Break: rename CharWidthProvider methods and change zeroWidth argument to CharPredicate
• Fix: TestUtils.getOptions(SpecExample, String, Function<String, DataHolder>) to return immutable data holder to prevent leak of options between spec examples
• Fix: escape of plain . at start of continuation.
• Fix: ListBlockParser escape/unEscape handler not checking if returned index == -1 and not handling empty char sequence.
• Fix: MarkdownParagraph
  • not checking if passed sequence is empty before invoking special lead-in escape/unEscape handlers, which is not part of the API contract.
  • backspace after typing a char should not remove spaces which surrounded the char
• Fix: make Paired extend Map.Entry
• Fix: #381, StackOverflowError with long base64 image and LINKS_ALLOW_MATCHED_PARENTHESES disabled, USE_HARDCODED_LINK_ADDRESS_PARSER was not used if matched parentheses were disabled causing stack overflow in regex evaluation. szxvalsdfj lasdf jffdddd

0.59.80

• Fix: erroneous not-null annotation on nullable return in InlineParser.parseCustom(BasedSequence, Node, BitSet, Map<Character, CharacterNodeFactory>)

0.59.78

• Fix: MarkdownParagraph wrap tracked offset adjustment:
  • on typing space in middle of word with second word becoming first non blank on a continuation line.
  • on typing space at start of continuation line with LS ahead of position.
  • on typing space after non-space at start of continuation line with LS ahead of position.
  • on typing space after end of paragraph after the right margin.

0.59.76

• Fix: MarkdownParagraph
  • Add: tracked offsets handling to prepare and resolve tracked offsets in wrapped text
    • handle inserting a space when tracked offset is followed by a space to allow two spaces at tracked offset
    • handle tracked offset followed by LS to resolve offset to end of previous line
    • handle backspace on first character of continuation line to splice word to last word of previous line.
  • Fix: To take CharSequence for indent properties instead of int for number of spaces to allow arbitrary indentation characters.
• Fix: SegmentOffsetTree aggregated length replacement to contain startOffset of next segment instead of endOffset of segment at position, to mimic equivalent aggregated length of SegmentTree.
• Fix: BasedOffsetTracker to handle 0 span and 1 span offset ranges and return info for the requested offset.

0.59.74

• Fix: MarkdownParagraph line break getter/setter names.
• Fix: BasedOffsetTracker
  • if segment end offset equals offset and no segment containing offset is found, then use segment for index calculation.
  • assertion conditions when search returns null, offset can be >= end offset

0.59.72

• Add: Formatter Paragraph wrapping options and code.
  • Add: Formatter.RIGHT_MARGIN, default 0, if >0 then text will be wrapped to given margin
  • Add: Formatter.APPLY_SPECIAL_LEAD_IN_HANDLERS, default true, when true will escape special lead-in characters which wrap to beginning of line and un-escape any which wrap from beginning of line. Used to prevent special characters inside paragraph body from starting a new element when wrapped to beginning of line.
• Add: CustomBlockParserFactory.getLeadInHandler(DataHolder) and handle escaping/unescaping special lead in characters for the block elements. Add escape/unescape to:
  • BlockQuoteParser
  • ListBlockParser
  • HeadingParser
  • AdmonitionBlockParser
  • AsideBlockParser
  • DefinitionItemBlockParser
• Add: Parser.Builder.specialLeadInHandler(SpecialLeadInHandler) to allow parser extensions to register special lead in escaper/unescaper handlers for parser extension block elements if they do not have any custom block parser factories but do create extensions.
• Add: special lead-in escape/unescape handlers during text wrapping
• Add: SegmentOffsetTree for binary search on offset in base sequence for conversion to index in result sequence.
• Fix: BitFieldSet.toString() to output field values when fields are more than one bit long.
• Add: BasedOffsetTracker to take BasedSequence result (could be segmented, offset in based sequence, PositionAnchor and return index of this offset in the resulting based sequence.

0.59.70

• Fix: convert BitEnumSet to BitFieldSet to allow bit fields of more than 1 bit.
  • Add: Optional BitField interface, if implemented by Enum for use by BitFieldSet then bit fields can be 1 to 64 bits in size, with manipulation of bit packed values and iteration over non-zero valued elements.

0.59.68

• Fix: rename LineFormattingAppendable to LineAppendable
• Fix: remove unused overloads and methods from LineFormattingAppendable
• Add: new methods to LineFormattingAppendable for prefix changes and combined prefix/line manipulation.
• Break: rename LineAppendable.SUPPRESS_TRAILING_WHITESPACE to LineAppendable.F_TRIM_TRAILING_WHITESPACE
• Break: rename and invert LineAppendable.ALLOW_LEADING_WHITESPACE to LineAppendable.F_TRIM_LEADING_WHITESPACE
• Deprecate: convenience copies of LineAppendable flags in Formatter and HtmlRenderer. Use LineAppendable directly.
• Break: Formatter.FORMAT_FLAGS default changed from 0 to LineAppendable.F_TRIM_LEADING_WHITESPACE to reflect inversion of option flag and preserve behavior
• Break: HtmlRenderer.FORMAT_FLAGS default changed from 0 to LineAppendable.F_TRIM_LEADING_WHITESPACE to reflect inversion of option flag and preserve behavior
• Fix: remove my prefix from fields in flexmark-utils to keep consistent naming convention in module.
• Add: BitEnumSet to expose the bit mask of elements to use as efficient and convenient bit masked options
• Add: ArrayUtils.toArray(BitSet) to return an int[] of all bit numbers that are set.
• Break: make constructor Range(int, int) protected. Use Range.of to create instances.

0.59.66

• Fix: move experimental concept classes out of the library. Too much clutter of unused old experimental stuff.
  • Fix: move PositionList related classes to flexmark-util-experimental
  • Fix: move unused classes from utils to experimental
• Fix: rename BasedSequenceBuilder to SequenceBuilder
• Fix: MarkdownParagraph
  • use SegmentBuilder for accumulating wrapped text.
  • Fix: remove any position tracking code from paragraph formatter

0.59.64

• Fix: SoftBreakNode in link ref would be trimmed and loose contained EOL as a result.
• Fix: SegmentedSequence to take SegmentBuilder parts for sequence generation instead of list of based sequences. The parts are already resolved by builder, based sequences duplicates useless work on both ends.
  • Fix: optimize storage by implementing binary search with segments serialized to byte array.
    • Add: build SegmentBuilder from SegmentTree
    • Add: code to extract startOffset/endOffset and treeData range for subSequence of SegmentedSequenceTree
    • Add: SegmentTree binary search tree for segmented sequence segments
  • Add: SegmentedSequenceTree binary tree based segmented sequences with option to use old SegmentedSequenceFull if desired.
  • Add: build SegmentTree from SegmentBuilder segments
• Add: segment builder text part stats for first256 and repeatedText, former if all chars in part < 256, latter if all chars are the same. In preparation for storage optimized segmented sequence impl.
• Fix: optimized SegmentBuilder
  • Change TEXT segments from String content to start/end being the -ve of start+1/end+1 of the subsequence for this TEXT in StringBuilder
  • Change Seg to remove myText field and require CharSequence argument for to be passed to all methods which returned myText in previous implementation. Returned value should be CharSequence which is the subSequence representing the TEXT content.
  • TEXT segments no longer have a start/end offset of prev/next BASE since their start/end offsets represent offset in StringBuilder for their character sequence. Their base start offset is endOffset of previous BASE or startOffset of SegmentBuilder if first segment. Their endOffset is startOffset of next BASE or endOffset of SegmentBuilder if it is a dangling TEXT segment.
  • TEXT segments to be accumulated in StringBuilder so a single char sequence for out of base characters is available.
  • SegmentBuilder tracks last immutableOffset in its StringBuilder to reflect the accumulated, immutable out of base TEXT segments in the segment builder.
  • immutableOffset and StringBuilder.length() represents the start/end offsets of dangling TEXT segment.
  • dangling TEXT segment is mutable until a BASE segment is added after it which results in SegmentBuilder endOffset change and after optimizeText method has been invoked. If TEXT segment becomes immutable then immutableOffset is changed to the endOffset of the TEXT segment's subSequence end.
  • adding an ANCHOR after dangling TEXT with an offset equal to StringBuilder endOffset will invoke optimizeText to allow extending previous BASE range and reducing dangling TEXT content. It will not change the dangling TEXT to immutable TEXT to allow appending more text after the call.
  • appending TEXT to SegmentBuilder appends the text to StringBuilder and updates out of base stats for appended TEXT content. Out of base stats always reflect the current content of SegmentBuilder
  • methods:
    • handleOverlap is invoked when an overlapping, BASE segment is added. It may result in calling optimizeText if the resolution of overlap causes a BASE segment to be appended after a dangling TEXT segment.
      • An exception is thrown if:
        • an overlap is found in returned segments
        • text length of returned segments does not equal the text length of original sequence.
      • Arguments are Object[3]: last BASE range, optional dangling TEXT and new BASE range passed as Object[] with Range and CharSequence representing BASE and TEXT segments.
      • StringBuilder content is truncated to immutableOffset and out of base statistics are updated to remove the TEXT content character information before method invocation because they will be updated for changed segments returned by the method.
      • It must return an Object[] of segments that resolve the overlap. No optimization needs to be done at this point because optimizeText will be called if needed as the resulting segments are appended to the SegmentBuilder
      • SegmentBuilder will replace last range and append the rest of the returned segments and update all internal structures, eliminating the concern for these in handleOverlap.
    • optimizeText called when a BASE or ANCHOR segment is added after dangling TEXT segment.
      • An exception is thrown if:
        • an overlap is found in returned segments
        • text length of returned segments does not equal the text length of original arguments.
      • Arguments are Object[3]: last BASE range or Range.NULL, dangling TEXT and new BASE/ANCHOR range or Range.NULL are passed as Object[] with Range and String representing BASE/ANCHOR and TEXT segments respectively.
      • StringBuilder content is truncated to immutableOffset and out of base statistics are updated to remove the TEXT content character information before method invocation because they will be updated for changed segments returned by the method.
      • Method to return an Object[] of optimized segments, with Range representing BASE/ANCHOR segments and CharSequence representing TEXT segments.
      • SegmentBuilder will replace last range and append the rest of the returned segments and update all internal structures, eliminating the concern for these in optimizeText.

0.59.62

• Fix: remove space to   conversion of spec example first line. Was done to allow GitHub to display spec examples as fenced code. No longer needed since GitHub switched to CommonMark parser.
• Fix: update all spec files to have no non-break spaces in example opener.
• Add: TestUtils.customIntOption(String, String, Function<Integer, DataHolder>) and TestUtils.customStringOption(String, String, Function<String, DataHolder>) to ease creating CUSTOM_OPTION option types.

0.59.60

• Add: SettableInstance and SettableExtractedInstance classes to facilitate implementation of settings which can be changed through spec example options.
• Break: move SpecReader.getSpecInputStream(ResourceLocation) and SpecReader.getReadSpec(ResourceLocation) to ResourceLocation.getResourceInputStream(ResourceLocation) and ResourceLocation.getResourceText(ResourceLocation)
• Add: ResourceLocation.getResourceInputStream() and ResourceLocation.getResourceText()
• Add: static BuilderBase.removeExtensions(DataHolder, Collection<Class<? extends Extension>>) for removing extension classes from SharedDataKeys.EXTENSIONS.

0.59.58

• Add: SequenceUtils.equals(CharSequence, Object) to consolidate char sequence equals
• Fix: SegmentedBuilder if anchor position is less than current endOffset then ignore anchor.
• Deprecate: TestUtils.showTabs(String) and TestUtils.unShowTabs(String) in favour of TestUtils.toVisibleSpecText(String) and TestUtils.fromVisibleSpecText(String) because these convert more than tabs.

0.59.56

• Fix: bug causing duplicated text in segmented sequences

• Fix: optimize SegmentedSequence, non base characters are now stored in base offsets as -(char+1), this eliminates nonBasedChars completely.

• Add: segmented sequence stats collection

• Add: options to based sequences which are not used unless a sequence passed to BasedSequence.of is first wrapped in BasedOptionsSequence. Then all sequences created from this base will have int bit options testable with BasedSequence.isOption(int) and get arbitrary options with BasedSequence.getOption(DataKeyBase<T>) or just get the whole DataHolder options via BasedSequence.getOptions() for more sophisticated needs. If the DataHolder passed to BasedOptionsSequence is mutable then

  Makes adding customized behavior and features easy and transparent to implementation with no overhead since SubSequence delegates to its baseSeq if it is a BasedOptionsHolder for these methods but otherwise does no checking. The overhead is one level of delegation of CharSequence methods for sequences with options with no impact on the rest.

0.59.54

• Fix: rework PositionList and Position

  • Add: concept of scope framing to invalidate all position created in the frame and not taken out via Position.unframed(), will also take out listeners and preview listeners added in the frame. This optimizes removal of positions from notifications. Otherwise the update list grows fast and with every iteration, until gc is run, which may be a while. Problem becomes worse if the position list is long lived and accumulates positions which were not detached over its lifetime, between gc runs. Every iteration should be scoped. Iterating 10x over a 10 item list without scoping has max 110 listener updates, scoping inner loop only 10 max listeners, scoping inner loop 1.

    Usage is: Object frame = openFrame(); try { ... } finally { closeFrame(frame); } with all positions created between open/close frame being detached. Frames can be nested and are validated for double closing, unclosed nested, and other common errors. Positions will throw an exception if used after frame is closed.

  • Add: IPreviewPositionListener to PositionList so changes can be previewed and changed. Useful for SegmentBuilder to keep invariants and tracked counts updated when these are changed by optimizers so optimizers do not need to concern themselves about these.

• Add: SegmentBuilder option flags for tracking unique non-based characters and including anchors (0 span ranges) in the segment list.

0.59.52

• Add: BasedSequence.addSegments(BasedSegmentBuilder) so each sequence adds its own optimized segment list without requiring by character scanning ofg of BasedSequence.getIndexOffset(int) and building the segments the hard way.
  • Fix: BasedSequenceBuilder to use BasedSequence.addSegments
    • ReplacedBasedSequence it has no use other than to test if sequence replaces base
  • Fix: Change segment parts to EditOp which extends range and holds an nullable string. This will allow keeping track of replaced parts of base sequence.
    • if string is null then it is a range from base
    • range is empty then string is not null and is an insert at position in base
    • range not empty and string not empty then it is a replacement of base segment.
• Fix: rename BasedCharsRecoverySegmentOptimizer to CharRecoveryOptimizer
• Fix: SegmentBuilder change append...() to plain append()
• Fix: SegmentBuilder to keep track of start and end range in the list based on first/last range added.
  • Test: need to test this more thoroughly
  • Fix: simplify overlap by passing index to last range in list and the range overlapping it. No position.
  • Fix: simplify optimization by passing the full list to optimizer. No position list.
  • Fix: eliminate multiple methods for the same functionality, append: EditOp, start/end offsets or String. The rest are history.
  • Test: cached start/end/length/last range index.
• Add: BasedSegmentBuilder to convert overlap in added ranges to out of base text
• Fix: change BasedSequenceBuilder to use BasedSegmentBuilder for segment accumulation instead of its own implementation.
  • Add: construction with optional BasedSegmentOptimizer list and apply optimizers to segment list before generating sequence or string.
  • Fix: if there is an overlap in appended sequence with previous segments, add the overlap as out of base text instead of throwing an exception.
  • Fix: optimize and simplify for speed, too many methods. Keep only add/append of CharSequence and add/append for start/end offsets. The rest are gone.
• Add: Common char sets:
  • DECIMAL_DIGITS
  • HEXADECIMAL_DIGITS
  • OCTAL_DIGITS
• Fix: infinite recursion in deprecated SubSequence.of
• Break: delete BasedSequence.getIndexRange(int, int) does not do the intended job and is complicated and useless at the same time for its purpose.
• Deprecate: to be delete BasedSequence.getIndexOffset(int) as soon as SegmentedSequence is migrated to segment builder sequence part list.

0.59.50

• Fix: move spec resource files to flexmark-test-specs resources. Any modules needing the spec have to add a test dependency on this module.
• Break: try to reduce footprint of IRichSequence and IRichSequenceBase by making getting rid off some overloads with default params.
  • Add: factor out bulk of IRichSequenceBase to SequenceUtils so functionality could be used on any CharSequence.
  • Fix: change all character inclusion based functions to use CharPredicate
• Add: CharPredicate to consolidate all character inclusion testing via predicates.
• Fix: make BaseSequence.of use plain based sequences and not managed for now. Issues when mutable sequences are passed as parameters make it unstable. Have to use or immutable char sequence or char backed sequence for managed and that will create too much overhead if used for all cases.
• Add: PositionList and Position classes to represent list and position in that list which allow modification, deletion, insertions while iterating over the list. Multiple position instances to different indexed elements can be created and all will be updated to reflect changes in their element's position.
  • Fix: cleanup naming conventions in new classes.
  • Fix: rewrite implementation and optimizer using PositionList and Position instead of hand rolled code.
  • Fix: make Position handle insert/delete index change invalidation instead of PositionList
  • Add: add anchor for position to allow next/prev anchoring
  • Add: positions are now have iterable from position to end/start of list, or from next to end/previous to start.
  • Fix: replace flags and position anchor field with byte flags.

0.59.48

• Fix: SegmentBuilder to not add empty text.

• Add: SegmentOptimizer for SegmentBuilder segment optimization

• Add: CharMatchingSegmentOptimizer to change replaced text to range in sequence when the replacement text matches sequence characters. Eliminates out of base unnecessary characters.

• Fix: add quick fail/success test for equality of base object. base strings through delete range/insert string operations.

• Fix: make all based sequence test for same base use getBase() instead of getBaseSequence(). It is the underlying base object that is important not the based sequence used to wrap it.

• Add: BasedSequenceManager to allow re-use of equivalent base sequences to eliminate the problem of having different base sequences for the same text content. Only implemented for SubSequence not CharSubSequence

  Uses weak refs and weak hash map so as long as there is a reference to previously created sequence base, requests to BaseSequenceManager.INSTANCE.getBaseSequence(T, Function<T, BasedSequence>) will return an existing based sequence for the passed object.

  Uses fast fail on length() and hashCode() differences with fallback on weak hash map for getting previously computed equality comparisons to minimize expensive duplicate content comparisons for previously determined equal/unequal sequence bases.

  Weak refs and maps ensure cache is cleared when last used reference to base or its subsequence is released and gc() has run.

• Fix: change SubSequence to create a String if passed sequence implements Appendable to eliminate use of mutable sequences as bases the same way CharSubSequence did for StringBuilder. It is important not to use mutable CharSequences for BasedSequence.of() argument. If not sure toString() on the argument is the best option.

• Fix: change all possible uses of CharSubSequence to SubSequence

• Add: SegmentBuilder to track offset ranges in original or base sequence and inserted out of

• Fix: BasedSequenceBuilder.add(CharSequence) now does not complain when adding a BasedSequence which is from a different base than the builder. It simply treats it as it does any other non-based sequence and adds it as out of base chars.

• Fix: SequenceBuilder.addAll(Collection<? extends CharSequence>) now takes list of anything extending CharSequence

• Add: javadoc to BasedSequence.baseSubSequence(int, int) that it should return sequence equal to base sequence without any modifications or additions. This should really be implemented only in real base sequences like CharSubSequence and SubSequence the rest should rely on BasedSequenceImpl.

0.59.46

• Fix: merge formatter always adding <> around autolinks even when original did not have them.
• Fix: docx conversion renders mail link text with mailto: prefix
• Fix: add '\0' to '\uFFFD' conversion in RichSequence
• Fix: #376, convert markdown to html, delimiters are not allowed to span table cells.
• Add: Text addText(String value, boolean noProofRPr, boolean createR)
• Add: Text addTextCreateR(String value, boolean noProofRPr)
• Add: Text addTextCreateR(String value)
• Add: Text addText(String value, boolean noProofRPr)
• Add: Text addText(String value)
• Add: Text addInstrText(String value, boolean noProofRPr, boolean createR)
• Add: Text addInstrTextCreateR(String value, boolean noProofRPr)
• Add: Text addInstrTextCreateR(String value)
• Add: Text addInstrText(String value, boolean noProofRPr)
• Add: Text addInstrText(String value)
• Add: FldChar addFldChar(STFldCharType charType, boolean createR)
• Add: FldChar addFldCharCreateR(STFldCharType charType)
• Add: FldChar addFldChar(STFldCharType charType)
• Deprecate: DocxContext.addWrappedText() use DocxContext.addText(String)
• Deprecate: DocxContext.text(String) use DocxContext.addTextCreateR(String)
• Fix: IRichSequenceBase.split and IRichSequenceBase.splitList some combinations would pass parameters out of place causing wrong operation.
• Fix: MarkdownParagraph wrapping to use simplified left aligned wrapping with no trailing spaces

0.59.44

• Add: docx renderer form fields via [name]{.class attributes} where name is given by DocxRenderer.FORM_CONTROLS and if not empty or blank then an unresolved reference link with that name will be converted to a word form input field with type and options given in teh attributes.
  • drop down list, class name .dropdown attributes:
    • name - name for the form field
    • default - default selection first item if no default provided. Can be option text (case sensitive tried first, if no match then case insensitive is tried), if text match fails then attempted to parse as an integer index 1..number of options
    • help - text to show in status bar and on hitting F1
    • options - list of options separated by |, individual options will be trimmed but otherwise left as is
  • checkbox, class name .checkbox attributes:
    • name - name for the form field
    • checked - default will be checked, otherwise unchecked
    • help - text to show in status bar and on hitting F1
  • text, class name .text attributes:
    • name - name for the form field
    • default - default value
    • help - text to show in status bar and on hitting F1
    • max-length - number or leave out for unlimited
    • type - gives the type of text field
      • regular - default if type not provided, regular text
        • format - UPPERCASE, LOWERCASE, FIRST CAPITAL, TITLE CASE, not case sensitive, space between words can be eliminated or replaced by -
      • date - date
        • format - word date format
      • number - number
        • format - word number format
      • current-date - current date
        • format - word date format
      • current-time - current time
        • format - word time format

0.59.42

• Break: BasedSequence.append(CharSequence[]) now constructs as segmented sequence so proper ordering of segments is enforced. This means IllegalArgumentException will be thrown when this condition is not met. To get old functionality you need to create an appended string of all sequences and BasedSequence.of() on that string.
• Break: rename MappedSequence to MappedBasedSequence to allow for MappedRichSequence
• Break: rename RichCharSequence to IRichSequence to separate base interfaces for rich and based sequences and allow for RichSequence to be the base interface for rich sequences similar to BasedSequence for based sequences
• Break: rename RichCharSequenceBase to IRichSequenceBase
• Break: rename RichCharSequenceImpl to RichSequenceImpl
• Break: rename RepeatedCharSequence to RepeatedSequence
• Add: RichSequence.of for creating RichSequence instances
• Add: BasedSequence.of for creating BasedSequence instances
• Deprecate: RichSequenceImpl.of, use RichSequence.of instead
• Deprecate: BasedSequenceImpl.of, use BasedSequence.of instead
• Deprecate: SubSequence.of, use BasedSequence.of instead
• Deprecate: UpperCaseMapper and LowerCaseMapper use ChangeCase.toUpperCase and ChangeCase.toLowerCase
• Deprecate: all static factory method named of to more specific so that errors are less likely. repeatOf, prefixOf otherwise it is possible to have the wrong factory method called and not see it.
• Break: move
  • com.vladsch.flexmark.tree.iteration.ArrayIterable to com.vladsch.flexmark.util.collection.iteration.ArrayIterable
  • com.vladsch.flexmark.util.html.CellAlignment to com.vladsch.flexmark.util.collection.iteration.ArrayIterable
  • com.vladsch.flexmark.ast.DelimitedNode to com.vladsch.flexmark.util.ast.DelimitedNode
• Fix: IRichSequence and by extension RichSequence and BasedSequence
  • Add: IRichSequence.trimToEndOfLine(CharSequence eolChars, boolean includeEol, int index) - trim end to end of line at index
  • Add: IRichSequence.trimToStartOfLine(CharSequence, boolean, int) - trimmed end after next EOL
  • Add: IRichSequence.leadingBlankLinesRange(CharSequence eolChars, int fromIndex, int endIndex) - find next range of blank lines
  • Add: IRichSequence.trailingBlankLinesRange(CharSequence eolChars, int startIndex, int fromIndex) - find previous range of blank lines
  • Add: IRichSequence.blankLinesRemovedRanges(CharSequence eolChars, int fromIndex, int endIndex) - list of ranges in sequence between fromIndex and endIndex excluding all blank lines
  • Add: range based methods:
  • IRichSequence.subSequence(Range)
  • IRichSequence.subSequenceAfter(Range)
  • IRichSequence.subSequenceBefore(Range)
  • Add: editing methods:
  • IRichSequence.insert(CharSequence, int) insert another char sequence at index
  • IRichSequence.delete(int, int) delete range in this sequence
  • IRichSequence.replace(CharSequence, CharSequence) one char sequence in this sequence with another
  • IRichSequence.replace(int, int, CharSequence) range in this sequence with char sequence
• Add: to BasedSequence
  • Add: BasedSequence.extendToEndOfLine(CharSequence eolChars, boolean includeEol) - extend end to end of line in basedSequence
  • Add: BasedSequence.extendToStartOfLine(CharSequence eolChars, boolean includeEol) - extend start to start of line in basedSequence
  • Convenience methods returning coordinates from the base sequence
    • @NotNull Pair<Integer, Integer> baseLineColumnAtIndex(int index);
    • @NotNull Range baseLineRangeAtIndex(int index);
    • int baseEndOfLine(int index);
    • int baseEndOfLineAnyEOL(int index);
    • int baseStartOfLine(int index);
    • int baseStartOfLineAnyEOL(int index);
    • int baseColumnAtIndex(int index);
    • @NotNull Pair<Integer, Integer> baseLineColumnAtStart();
    • @NotNull Pair<Integer, Integer> baseLineColumnAtEnd();
    • int baseEndOfLine();
    • int baseEndOfLineAnyEOL();
    • int baseStartOfLine();
    • int baseStartOfLineAnyEOL();
    • @NotNull Range baseLineRangeAtStart();
    • @NotNull Range baseLineRangeAtEnd();
    • int baseColumnAtEnd();
    • int baseColumnAtStart();
• Add: to Node convenience methods delegated to getChars()
  • int getStartOffset() to getStartOffset()
  • int getEndOffset() to getEndOffset()
  • int getTextLength() to length()
  • BasedSequence getBaseSequence() to getBaseSequence()
  • Range getSourceRange() to etSourceRange()
  • BasedSequence baseSubSequence(int startIndex, int endIndex) to baseSubSequence(startIndex, endIndex)
  • BasedSequence baseSubSequence(int startIndex) to baseSubSequence(startIndex)
  • BasedSequence getEmptyPrefix() to getEmptyPrefix()
  • BasedSequence getEmptySuffix() to getEmptySuffix()
  • int getStartOfLine() to baseStartOfLine()
  • int getEndOfLine() to baseEndOfLine()
  • int startOfLine(int index) to baseStartOfLine(index)
  • int endOfLine(int index) to baseEndOfLine(index)
  • Pair<Integer, Integer> lineColumnAtIndex(int index) to baseLineColumnAtIndex(index)
  • Pair<Integer, Integer> lineColumnAtStart() to baseLineColumnAtStart()
  • Pair<Integer, Integer> getLineColumnAtEnd() to baseLineColumnAtEnd()
• Add: ArrayUtils for searching arrays of <T> on value or predicate
  • indexOf() variations for forward search
  • lastIndexOf() variations for backward search
  • firstOf() variations for forward search returning matched element or null
  • lastOf() variations for backward search returning matched element or null
• Break: RichCharSequence all trim, trimStart, trimEnd keepLength argument is now keep with different meaning. Instead of min length of string to keep, it is min number of trimmable characters to keep, which is more useful but does break compatibility. Now it can be used to leave whitespace padding up to a maximum of keep characters.
• Fix: reformat tests for compound sections
• Deprecate: Range.subSequence(CharSequence) use Range.basedSubSequence(CharSequence), Range.richSubSequence(CharSequence) or for plain sequences Range.charSubSequence(CharSequence)
• Fix: rename Range.subSequence() to Range.basedSubSequence()
• Fix: rename Range.safeSubSequence() to Range.safeBasedSubSequence() which limit the subsequence within 0 to length
• Add: Range.richSubSequence()
• Add: Range.safeRichBasedSubSequence() which limit the subsequence within 0 to length
• Add: Range.charSubSequence()
• Add: Range.safeCharBasedSubSequence() which limit the subsequence within 0 to length

0.59.40

• Break: remove example option parsing related TestUtils methods.
• Add: spec example language per section options and rendering in HTML.
• Add: ExampleOption to parse and provide information about
• Add: compound spec-example sections, to combine previous lower level headings via splice(" - ") for better organization of tests and to allow hierarchical structure view for spec files in Markdown Navigator
• Fix: remove all file url prefix for tests. Either provide the file url for the spec or default resolution will be used.

0.59.38

• Fix: remove unnecessary type param from MutableDataHolder.remove(DataKeyBase<?>)
• Fix: Parser.HEADING_NO_EMPTY_HEADING_WITHOUT_SPACE had wrong regex. Did not allow non-empty headings without space either.

0.59.36

• Fix: remove synchronization. The issue was with using HashMap.computeIfAbsent() and passing the factory, which could and would access and create other keys. This caused concurrent modification, not threading. Now computing value with factory then adding it to the dataSet.

0.59.34

• Add: synchronized around HashMap modification to ensure thread safety. Elimination of nullable keys would allow using ConcurrentHashMap in DataSet but access and modification to options is done at start of parse as extensions and core load their options, afterwards the data set is mostly not modified or accessed.

0.59.32

• Fix: add nullable annotations to DataSet and DataKey classes
• Break: DataKey now represents not null data values, use NullableDataKey if using nullable values, for processing of either type of data key, use DataKeyBase super class which does not specify nullability for the data value.
• Break: DataValueFactory now has non-nullable dataHolder argument and nullable result. Use:
  • DataNotNullValueFactory for nullable result and non nullable dataHolder
  • DataValueNullableFactory for nullable result and nullable dataHolder
  • DataNotNullValueNullableFactory for non-nullable result and non-nullable dataHolder
• Add: DataKey 3 argument constructor with pre-computed non null default value to be used for null dataHolder defaults. This way the factory is DataNotNullValueFactory which takes a non-null data holder and returns a computed value.
• Add: NullableDataKey 3 argument constructor with pre-computed nullable default value to be used for null dataHolder defaults. This way the factory is DataValueFactory which takes a non-null data holder and returns a computed value.
• Fix: add DataKey.get(DataHolder) to replace DataKey.getFrom(DataHolder), shorter and compatible with Kotlin array access syntax, also handles nullability of data.
• Fix: deprecate DataKey.getFrom()
• Fix: replace all DataHolder.get() by DataKey.get()
• Fix: add nullability annotations to a boat load of classes.
• Add: MutableDataHolder.set(@NotNull DataKey<T>, @NotNull T) and MutableDataHolder.set(@NotNull NullableDataKey<T>, @Nullable T) to respect nullability of key's data value.
• Add: DataKey.set(@NotNull MutableDataHolder, @NotNull T) and NullableDataKey.set(@NotNull MutableDataHolder, @Nullable T) to allow setting data values via keys, if needed.

0.59.30

• Fix: change all test case construction param from Map<String, DataHolder> to Map<String, ? extends DataHolder> to allow for any type implementing the right map.

0.59.28

• Fix: pass through opening line of spec example to DumpSpecReader so it can construct an accurate full spec version of expected text.
• Fix: add ResourceResolverManager to handle ResourceUrlResolver registration.

0.59.27

• Add: parameterized test case options, ones that have option name followed by [ and terminated by ]. Used for custom handling variations of option. These should set TestUtils.CUSTOM_OPTION to a BiFunction<String option, String text, DataHolder> where option is the option name (type in example), text is the text between [].

  optionsMap.put("type", new MutableDataSet().set(CUSTOM_OPTION, ActionSpecTestCase::typeOption));
  

    public static DataHolder typeOption(String option, String params) {
      // allow escape
      String text = params
              .replace("\\\\", "\\")
              .replace("\\]", "]")
              .replace("\\t", "\t")
              .replace("\\n", "\n")
              .replace("\\r", "\r")
              .replace("\\b", "\b");
  
      return new MutableDataSet().set(ACTION_NAME, TYPE_ACTION).set(TYPE_ACTION_TEXT, text);
  }
  

• Add: now if test option DataHolder with TestUtils.CUSTOM_OPTION set then it will be invoked even when no [] present to allow handling of empty params, which will be null in this case.

• Add: before checking for parameterized option the full option with params and [] is tried. If present then its results are used, but also checked if TestUtils.CUSTOM_OPTION is present.

• Add: ResourceUrlResolver, registered via ResourceLocation.registerUrlResolver(Function<String, String>) to allow resource URL conversion to URL used in example messages to be adjusted by project. Must register resolvers before trying to get an instance of ResourceLocation since resolution of file url is done in constructor.

• Add: example's URL in exceptions when processing undefined example option.

• Add: HTML comment handling to SpecReader, start/end comment tags must be the first non-blank on the line and outside of spec example.

0.59.25

• Fix: remove moved test class from core to utils module.

0.59.23

• Fix: change all tests to use static const ResourceLocation instead of creating a new instance.
• Fix: make ComboSpecTestCase.getSpecResourceLocation() final and remove it from all subclasses.

0.59.21

• Fix: move flexmark modules test classes out of test.util package which is for flexmark-test-util module's test classes
• Add: DataSet.registerDataKeyAggregator(DataKeyAggregator) to allow for aggregation action keys and aggregation for combining keys which require this. Only applied in
  • DataSet.aggregate()
  • DataSet.aggregate(DataHolder, DataHolder)
  • DataSet.aggregateActions(DataHolder, DataHolder)

0.59.19

• Fix: convert all resource info to ResourceLocation
• Add: class location resolution via file under test resources named for the root package of the module and containing directory path(s) from parent of resources directory to get source root for test files. for maven builds it is java/. This allows resolving absolute file path for class files in tests for generating the location file:///
  • Fix: package for some modules did not follow module/package naming convention
    • Fix: move flexmark-ext-superscript to proper package by naming convention com.vladsch.flexmark.ext.superscript
    • Fix: move flexmark-profile-pegdown to proper package by naming convention com.vladsch.flexmark.profile.pegdown
    • Fix: move flexmark-test-util spec package under com.vladsch.flexmark.test.util package
  • Add: test class location helpers to all modules:
    • flexmark-core-test/src/test/resources/com.vladsch.flexmark.core.test.txt
    • flexmark-docx-converter/src/test/resources/com.vladsch.flexmark.docx.converter.txt
    • flexmark-ext-abbreviation/src/test/resources/com.vladsch.flexmark.ext.abbreviation.txt
    • flexmark-ext-admonition/src/test/resources/com.vladsch.flexmark.ext.admonition.txt
    • flexmark-ext-anchorlink/src/test/resources/com.vladsch.flexmark.ext.anchorlink.txt
    • flexmark-ext-aside/src/test/resources/com.vladsch.flexmark.ext.aside.txt
    • flexmark-ext-attributes/src/test/resources/com.vladsch.flexmark.ext.attributes.txt
    • flexmark-ext-autolink/src/test/resources/com.vladsch.flexmark.ext.autolink.txt
    • flexmark-ext-definition/src/test/resources/com.vladsch.flexmark.ext.definition.txt
    • flexmark-ext-emoji/src/test/resources/com.vladsch.flexmark.ext.emoji.txt
    • flexmark-ext-enumerated-reference/src/test/resources/com.vladsch.flexmark.ext.enumerated.reference.txt
    • flexmark-ext-escaped-character/src/test/resources/com.vladsch.flexmark.ext.escaped.character.txt
    • flexmark-ext-footnotes/src/test/resources/com.vladsch.flexmark.ext.footnotes.txt
    • flexmark-ext-gfm-issues/src/test/resources/com.vladsch.flexmark.ext.gfm.issues.txt
    • flexmark-ext-gfm-strikethrough/src/test/resources/com.vladsch.flexmark.ext.gfm.strikethrough.txt
    • flexmark-ext-gfm-tasklist/src/test/resources/com.vladsch.flexmark.ext.gfm.tasklist.txt
    • flexmark-ext-gfm-users/src/test/resources/com.vladsch.flexmark.ext.gfm.users.txt
    • flexmark-ext-gitlab/src/test/resources/com.vladsch.flexmark.ext.gitlab.txt
    • flexmark-ext-ins/src/test/resources/com.vladsch.flexmark.ext.ins.txt
    • flexmark-ext-jekyll-front-matter/src/test/resources/com.vladsch.flexmark.ext.jekyll.front.matter.txt
    • flexmark-ext-jekyll-tag/src/test/resources/com.vladsch.flexmark.ext.jekyll.tag.txt
    • flexmark-ext-macros/src/test/resources/com.vladsch.flexmark.ext.macros.txt
    • flexmark-ext-media-tags/src/test/resources/com.vladsch.flexmark.ext.media.tags.txt
    • flexmark-ext-resizable-image/src/test/resources/com.vladsch.flexmark.ext.resizable.image.txt
    • flexmark-ext-spec-example/src/test/resources/com.vladsch.flexmark.ext.spec.example.txt
    • flexmark-ext-superscript/src/test/resources/com.vladsch.flexmark.ext.superscript.txt
    • flexmark-ext-tables/src/test/resources/com.vladsch.flexmark.ext.tables.txt
    • flexmark-ext-toc/src/test/resources/com.vladsch.flexmark.ext.toc.txt
    • flexmark-ext-typographic/src/test/resources/com.vladsch.flexmark.ext.typographic.txt
    • flexmark-ext-wikilink/src/test/resources/com.vladsch.flexmark.ext.wikilink.txt
    • flexmark-ext-xwiki-macros/src/test/resources/com.vladsch.flexmark.ext.xwiki.macros.txt
    • flexmark-ext-yaml-front-matter/src/test/resources/com.vladsch.flexmark.ext.yaml.front.matter.txt
    • flexmark-ext-youtube-embedded/src/test/resources/com.vladsch.flexmark.ext.youtube.embedded.txt
    • flexmark-ext-zzzzzz/src/test/resources/com.vladsch.flexmark.ext.zzzzzz.txt
    • flexmark-formatter-test-suite/src/test/resources/com.vladsch.flexmark.formatter.test.suite.txt
    • flexmark-html2md-converter/src/test/resources/com.vladsch.flexmark.html2md.converter.txt
    • flexmark-integration-test/src/test/resources/com.vladsch.flexmark.integration.test.txt
    • flexmark-java.wiki/src/test/resources/com.vladsch.flexmark-java.wiki.txt
    • flexmark-jira-converter/src/test/resources/com.vladsch.flexmark.jira.converter.txt
    • flexmark-pdf-converter/src/test/resources/com.vladsch.flexmark.pdf.converter.txt
    • flexmark-profile-pegdown/src/test/resources/com.vladsch.flexmark.profile.pegdown.txt
    • flexmark-tree-iteration/src/test/resources/com.vladsch.flexmark.tree.iteration.txt
    • flexmark-util/src/test/resources/com.vladsch.flexmark.util.txt
    • flexmark-youtrack-converter/src/test/resources/com.vladsch.flexmark.youtrack.converter.txt
    • flexmark/src/test/resources/com.vladsch.flexmark.txt
• Add: creation of SpecExample instance based on caller information with correct file/line of the source calling RenderingTestCase.assertRendering(String, String, String) or its variants.
• Fix: convert typographic smarts inline parser to match strings instead of using regex which sometimes takes long to execute for no apparent reason.

0.59.17

• Fix: make SpecExampleRendererBase reusable by moving any references to flexmark parsing rendering specifics to FlexmarkSpecExampleRenderer and moving up not related to flexmark parsing, like caching of generated HTML and AST.
• Fix: consolidate full spec and individual example test case in to one method.

0.59.15

• Fix: delete deprecated methods, classes, fields.
• Fix: delete deprecated flexmark-ext-gfm-tables
• Fix: format code

0.59.13

• Break: move formatter tests and html renderer to flexmark-core-test module to allow sharing of test base classes in extensions without causing dependency cycles in formatter module.

• Break: move formatter module into flexmark core. this module is almost always included anyway because most extension have a dependency on formatter for their custom formatting implementations. Having it as part of the core allows relying on its functionality in all modules.

• Break: move com.vladsch.flexmark.spec and com.vladsch.flexmark.util in flexmark-test-util to com.vladsch.flexmark.test.spec and com.vladsch.flexmark.test.util respectively to respect the naming convention between modules and their packages.

• Break: remove flexmark-html-parser, flexmark-html2md-converter is a replacement

• Fix: revert back to MutableDataSet.set(DataKey<T>, T) to enforce compile time type checking for typed class values

• Fix: clean up all tests to eliminate duplication and unnecessary interface methods.

  Metric	Before	After
  Files	1264	1297
  Lines	124,192	118,755
  Source Lines	96,669	92,145
  Comment Lines	9,878	9,597
  Blank Lines	17,646	17,013

0.59.11

• Fix: remove need to load/unload extensions. Properly handle removal from EXTENSIONS property for those tests that need it by implementing LOAD_EXTENSIONS property and UNLOAD_EXTENSIONS property handling in RenderingTestCase.combineOptions(DataHolder, DataHolder).
• Fix: remove loading/unloading and other test support code from builder, no longer needed, including withOptions() and builder construction from another builder.
• Fix: change EXTENSIONS from Iterable to Collection, which should not affect most code and easier to work with.
• Fix: translation mode formatting would loose indented code blocks in items if these blocks were indented more than the minimum indent for the item.
• Fix: paragraphs in containers that have non-blank termination marker would be erroneously marked as isTrailingBlankLine when they were terminated because the container was closed not because of a blank line.
• Fix: First paragraph of AdmonitionBlock would always format with leading blank line. Now it formats as it was in original markdown to allow preserving format as is when it is not significant or confusing.

0.59.9

• Fix: NO_EOL option for tests was applied inconsistently.
• Fix: add SpecExample argument to getSpecExampleRenderer
• Fix: add SpecExampleParse argument to addSpecExample
• Fix: regression bug #372 #372, [Regression?] Attributes extension not applied to `code` tag of code blocks
  • Add: AttributesExtension.FENCED_CODE_ADD_ATTRIBUTES, default FencedCodeAddType.ADD_TO_PRE_CODE for backward compatibility with 0.42, but if this is option is not set and AttributesExtension.FENCED_CODE_INFO_ATTRIBUTES is set to true, default will change to FencedCodeAddType.ADD_TO_PRE since attributes after info are used to add to the code tag.

0.59.7

• Add: SpecExampleRenderer.finalizeRender() to allow tests and others to clean up after each rendering
• Add: ability to combine data sets with DataKey<Consumer<?>> keys in a custom way. Needed if some option values do not contain a single value but multiple values that are set with a consumer changing some values in a structure. In such cases overwriting of these values may not be correct and the consumers need to invoked so the results of second consumer get to apply changes to options.
• Fix: add dual argument constructor to MutableDataSet

0.59.5

• Fix: change AttributesExtension.FENCED_CODE_INFO_ATTRIBUTES to default false to match previous versions.

0.59.1

• Fix: MarkdownTable index out of bounds in some functions when table has no rows
• Fix: DataSet copy constructors to accept null
• Fix: change DataSet.keySet() to DataSet.getKeys()
  • Deprecate: DataSet.keySet()
• Add: BasedSequence.extendByAny(CharSequence charSet, int maxCount) to extend the based sequence to include any following contiguous characters from the underlying based sequence that are in charSet. Variations include BasedSequence.extendByAny(CharSequence) for unlimited count and BasedSequence.extendByOneOfAny(CharSequence) for max count of 1.
• Add: BasedSequence.extendToAny(CharSequence charSet, int maxCount) to extend the based sequence to include first of any following characters from the underlying based sequence that are in charSet. Variations include BasedSequence.extendToAny(CharSequence) for unlimited count.
• Add: BasedSequence.prefixWithIndent(int maxColumns) to extend the based sequence to include leading indent from the underlying based sequence up to a maximum of given columns. Tabs are taken into account as set to 4 space columns. Variations include BasedSequence.prefixWithIndent() for unlimited indent.
• Break: make NodeAdaptedVisitor.myCustomHandlersMap private. Use NodeAdaptedVisitor.getHandler(Node), NodeAdaptedVisitor.getHandler(Class<?>), and NodeAdaptedVisitor.getNodeClasses() to get access to contained data.

0.50.48

• Fix: #382, Is there an option for number of whitespaces needed to create sub-lists?
  • FIXED_INDENT list parser did not convert list item looking text with >= 4 spaces to lazy continuation.

0.50.46

• Fix: #381, StackOverflowError with long base64 image and LINKS_ALLOW_MATCHED_PARENTHESES disabled, USE_HARDCODED_LINK_ADDRESS_PARSER was not used if matched parentheses were disabled causing stack overflow in regex evaluation.
• Fix: docx conversion renders mail link text with mailto: prefix

0.50.44

• Fix: #376, convert markdown to html, delimiters are not allowed to span table cells.

0.50.42

• Fix: regression bug #372, [Regression?] Attributes extension not applied to `code` tag of code blocks
  • Add: AttributesExtension.FENCED_CODE_ADD_ATTRIBUTES, default FencedCodeAddType.ADD_TO_PRE_CODE for backward compatibility with 0.42, but if this is option is not set and AttributesExtension.FENCED_CODE_INFO_ATTRIBUTES is set to true, default will change to FencedCodeAddType.ADD_TO_PRE since attributes after info are used to add to the code tag.
• Fix: data set copy constructors to accept null
• Fix: change DataSet.keySet() to DataSet.getKeys()
  • Deprecate: DataSet.keySet()
• Add: BasedSequence.extendByAny(CharSequence charSet, int maxCount) to extend the based sequence to include any following contiguous characters from the underlying based sequence that are in charSet. Variations include BasedSequence.extendByAny(CharSequence) for unlimited count and BasedSequence.extendByOneOfAny(CharSequence) for max count of 1.
• Add: BasedSequence.extendToAny(CharSequence charSet, int maxCount) to extend the based sequence to include first of any following characters from the underlying based sequence that are in charSet. Variations include BasedSequence.extendToAny(CharSequence) for unlimited count.
• Add: BasedSequence.prefixWithIndent(int maxColumns) to extend the based sequence to include leading indent from the underlying based sequence up to a maximum of given columns. Tabs are taken into account as set to 4 space columns. Variations include BasedSequence.prefixWithIndent() for unlimited indent.
• Break: make NodeAdaptedVisitor.myCustomHandlersMap private. Use NodeAdaptedVisitor.getHandler(Node), NodeAdaptedVisitor.getHandler(Class<?>), and NodeAdaptedVisitor.getNodeClasses() to get access to contained data.

0.50.40

• Add: BasedSequence.emptyPrefix() and BasedSequence.emptySuffix() to return empty sequence at start or end of the current sequence. Useful for PrefixedSubSequence.of(String, BasedSequence, int, int) variants without needing to extract the subsequence manually.
• Fix: factor out link URL parser to separate class with own tests
• Fix: link URL parser not handling char escaping, nested parenthesis tracking and other edge cases in jekyll URL macro parsing.

0.50.38

• Fix: real fix for Stack Overflow when parsing long URLs, by hand rolling parsing of link destination.
  • Add: Parser.USE_HARDCODED_LINK_ADDRESS_PARSER, default true, setting to false will use regex parsing for link destination address. May cause StackOverflowError exception on long input but does allow customizing the regex if needed.
• Add: BasedSequence.safeCharAt(int) will return \0 if index out of range for the sequence.
• Add: BasedSequence.safeBaseCharAt(int) will return character from the base sequence if index out of range for sequence or \0 if out of base sequence range.

0.50.36

• Fix: SegmentedSequence would return sequence with start > end if ending with PrefixedSubSequence whose position was > end of real sequences before.
• Add: BitIntegerSet.toArray(int[], int) for easy int[] creation from bit set.
• Add: BitIntegerSet.addAll(int[], int, int) for setting from int[]

0.50.34

• Add: source Node to com.vladsch.flexmark.util.format.TableCell
• Fix: Stack overflow when parsing large embedded images with space in URL enabled. Now link urls starting with data:image/png;base64, do not allow spaces in the link.

0.50.32

• Add: DocxRenderer image max-width tests
• Add: DocxRenderer image max-height attribute processing and tests
• Fix: refactor DocxRenderer tests to extract common code to base class
• Fix: heading rendering to work with styles using numbering

0.50.30

• Fix: DocxRenderer
  • Add: page break via empty paragraph with only {.pagebreak} attributes
  • Add: tab via {.tab} attributes
  • Add: inline image alignment with {align=}:
    • left - left align, wrap text to right
    • right - right align, wrap text to left
    • center - center align, wrap text to left and right
    • else no wrapping around image
  • Add: handling of font-size attribute, expects float of font size in pt, rounds to nearest 1/2 pt
  • Add: image size attribute handling for in and cm for inch and cm dimensions. Can be given in:
    • % for percent of page width
    • pt points
    • cm cm
    • in inches
    • if no units given then pixels are assumed.
  • Fix: if image size only given for one dimension, compute the other preserving aspect ratio
  • Fix: change default emoji size to 0.9 of line height
  • Fix: remove <> wrapper from URL target
  • Fix: lists to set style id to DocxRenderer.PARAGRAPH_BULLET_LIST_STYLE for unordered lists and DocxRenderer.PARAGRAPH_NUMBERED_LIST_STYLE for ordered lists, while still using DocxRenderer.BULLET_LIST_STYLE and DocxRenderer.NUMBERED_LIST_STYLE for the numbering style. Tight/loose is now uses paragraph settings from DocxRenderer.TIGHT_PARAGRAPH_STYLE and DocxRenderer.LOOSE_PARAGRAPH_STYLE respectively as mods on list style.
  • Fix: start image ids at 100000 to avoid id conflicts with images already in the document being added. For example in footers/headers.
  • Fix: for R tags, only add space="preserve" if contained text starts or ends with a space. Does not affect result but reduces xml noise.
• Fix: Formatter translation and merge rendering

0.50.28

• Fix: #362, ArrayIndexOutOfBoundsException in BasedSequence.indexOfAll
• Add: DocxRenderer.BULLET_LIST_STYLE default "BulletList", numbering style to use for bullet list item paragraph.
• Add: DocxRenderer.NUMBERED_LIST_STYLE default "NumberedList", numbering style to use for numbered list item paragraph.
• Add: Simple Attributes processing to docx conversion:
  • .className on paragraph elements will set the docx styleId to className if the style id is found.
  • Use {style=""} to set attributes on text or block elements. Only the following are processed:
    • color - text color
    • background-color - shade fill color, pattern always solid.
    • font-family - not implemented
    • font-size - not implemented
    • font-weight - set/clear bold (if using numeric weights then >= 550 sets bold, less clears it)
    • font-style - set/clear italic
• Add: merge multiple markdown documents into single document functionality. Includes making conflicting references between documents unique by adding an integer numeric suffix.
  • Conflicting headings will get a unique explicit id attribute by the flexmark-ext-attributes extension.
  • Added to Formatter module using modified translation API functionality.
    • Formatter.mergeRender(Document[], Appendable, HtmlIdGeneratorFactory)
    • Formatter.mergeRender(Document[], Appendable, int, HtmlIdGeneratorFactory)
    • Formatter.mergeRender(Document[], int, HtmlIdGeneratorFactory)
  • Add: merge tests and functionality to:
    • flexmark-formatter core rendering of anchors and headings.
    • flexmark-ext-abbreviation, by default making references unique is disabled. Enable by setting AbbreviationExtension.MAKE_MERGED_ABBREVIATIONS_UNIQUE to true for the formatter options used to merge documents.
    • flexmark-ext-attributes
    • flexmark-ext-enumerated-reference
    • flexmark-ext-footnotes
    • flexmark-ext-macros
  • Add: link mapping based on document's Formatter.DOC_RELATIVE_URL and Formatter.DOC_ROOT_URL properties.
• Fix: Formatter attributes would eliminate separator space before or after following text possibly changing the applicable node for the attribute.

0.50.26

• Fix: docx converter to use BulletList and NumberedList numbering list styles for list conversions to allow easy list item styling for multi-level lists
  • Add: DocxConverterEmpty to samples for generating flexmark-empty-template.docx from empty.md and empty.xml

0.50.24

• Fix: update docx4j to version 8.1.2
• Fix: update jsoup to version 1.11.3

0.50.22

• Fix: Attributes with spaces after { generated wrong previous text offsets

0.50.20

• Fix: #357, HTML to markdown and removed nested list
• Remove: FlexmarkHtmlConverter.EXT_TABLES unused conversion option.
• Add: allow attributes after fenced code info string as last non-blank text after the info string.
  • Add: AttributesExtension.FENCED_CODE_INFO_ATTRIBUTES default true, if false will not parse attributes after the fenced code info string

0.50.18

• Add: flexmark-html2md-converter module which implements HTML to Markdown conversion with an extension API to allow customizing the conversion process. Sample: HtmlToMarkdownCustomizedSample.java
  • Fix: #313, Ability to override tags processing in FlexmarkHtmlParser
• Fix: deprecate the old flexmark-html-parser classes

0.50.16

• Fix: GitLab block quotes should have >>> as termination, not <<<, macros are still terminated by >>> to allow nested block quotes in macros. Affects:
  • HTML converter
  • Formatter
• Fix: #351, Is there any special format requirement for processing html data to markdown
• Fix: #349, Translation Helper bugs, mix-up between anchors and other non-translating elements after translation.
• Fix: #349, Translation Helper bugs
• Fix: #348, WRAP_AUTO_LINKS defaults to false, Markdown loses a potential useful link
• Break: change FlexmarkHtmlParser.WRAP_AUTO_LINKS default to true
• Fix: add SpecReader.readExamples(String, String) which takes resource name and file URL to allow for non-standard resource file locations.
• Fix: html to markdown string not add EOL by default.
• Break: make Java 8 minimum version and use JDK 8 for compilation
  • Fix: IntelliJ Migration contained in migrate flexmark-java 0_42_x to 0_50_0.xml, to use:
    • copy to IntelliJ application settings to migration subdirectory
    • if you have the project which you want to migrate open, then close it
    • open the project in IntelliJ Ultimate or Community
    • update the flexmark-java dependency version to 0.42.0 (or later) and make sure the new library is downloaded/updated in the project.
    • use menu Refactor > Migrate... or search everywhere (double shift activation) and search for migrate action.
    • select migrate flexmark-java 0.42.x to 0.50.0
    • press Run
    • in the refactoring preview tool window that opens hit Do Refactor
  • Remove classes from utils which are implemented in Java 8
  • replace RunnableValue with Supplier, requires changing run() with get()
  • replace ValueRunnable with Consumer, requires changing run() with accept()
  • replace Factory with Supplier, requires changing create() with get()
  • replace Computable with Function, requires changing compute() with apply() and reversing template parameters.
  • replace ComputableFactory with Function, requires changing create() with apply() and reversing template parameters.
  • replace ComputeFactory with Function, requires changing create() with apply() and reversing template parameters.
  • replace all factory interfaces to be compatible with Function interface and change create() with apply()
  • replace com.vladsch.flexmark.util.collection.Consumer with java.util.function.Consumer
  • clean up DataKey and DataHolder related classes
    • move com.vladsch.flexmark.util.collection.DataValueFactory to com.vladsch.flexmark.util.DataValueFactory
    • replace com.vladsch.flexmark.util.collection.DynamicDefaultKey with com.vladsch.flexmark.util.options.DataKey
    • change MutableDataHolder.getOrCompute(DataKey<T>, DataValueFactory<T>) to MutableDataHolder.getOrCompute(DataKey<T>) with the data value factory taken from they key.
  • remove unused ItemIndexSetMap, use IndexedItemSetMap instead
  • move flexmark-util Node related classes and interfaces to com.vladsch.flexmark.util.ast
    • move com.vladsch.flexmark.util.NodeTracker to com.vladsch.flexmark.util.ast.NodeTracker
    • move com.vladsch.flexmark.util.BlockTracker to com.vladsch.flexmark.util.ast.BlockTracker
    • move com.vladsch.flexmark.util.KeepType to com.vladsch.flexmark.util.ast.KeepType
    • move com.vladsch.flexmark.util.IParse to com.vladsch.flexmark.util.ast.IParse
    • move com.vladsch.flexmark.util.IRender to com.vladsch.flexmark.util.ast.IRender
    • move com.vladsch.flexmark.util.collection.NodeClassifierVisitor to com.vladsch.flexmark.util.ast.NodeClassifierVisitor
    • move com.vladsch.flexmark.util.collection.NodeCollectingVisitor to com.vladsch.flexmark.util.ast.NodeCollectingVisitor
    • move com.vladsch.flexmark.util.collection.ClassifyingNodeTracker to com.vladsch.flexmark.util.ast.ClassifyingNodeTracker
    • move com.vladsch.flexmark.util.mappers.NodeClassifier to com.vladsch.flexmark.util.ast.NodeClassifier
  • move data key/set/holder classes to flexmark.util.data package from flexmark.util.options package
    • move com.vladsch.flexmark.util.options.DataHolder to com.vladsch.flexmark.util.data.DataHolder
    • move com.vladsch.flexmark.util.options.MutableDataSet to com.vladsch.flexmark.util.data.MutableDataSet
    • move com.vladsch.flexmark.util.options.MutableScopedDataSet to com.vladsch.flexmark.util.data.MutableScopedDataSet
    • move com.vladsch.flexmark.util.options.ScopedDataSet to com.vladsch.flexmark.util.data.ScopedDataSet
    • move com.vladsch.flexmark.util.options.DataValueFactory to com.vladsch.flexmark.util.data.DataValueFactory
    • move com.vladsch.flexmark.util.options.MutableDataHolder to com.vladsch.flexmark.util.data.MutableDataHolder
    • move com.vladsch.flexmark.util.options.DataSet to com.vladsch.flexmark.util.data.DataSet
    • move com.vladsch.flexmark.util.options.MutableDataSetter to com.vladsch.flexmark.util.data.MutableDataSetter
    • move com.vladsch.flexmark.util.options.DataKey to com.vladsch.flexmark.util.data.DataKey
• Add: LineFormattingAppendable and LineFormattingAppendableImpl
  • Fix: deprecate FormattingAppendable to be replaced by LineFormattingAppendable
  • Fix: deprecate FormattingAppendableImpl to be replaced by LineFormattingAppendableImpl
  • Fix: replace all uses of FormattingAppendable by LineFormattingAppendable
  • Fix: replace all uses of FormattingAppendableImpl by LineFormattingAppendableImpl
  • Delete: FormattingAppendable and FormattingAppendableImpl
• Fix: Factor out BasedSequenceImpl functionality that does not depend on BasedSequence and can be applied to any CharSequence into its own RichCharSequence interface with default abstract implementation in RichCharSequenceBase and implementation in RichCharSequenceImpl.

0.42.14

• Fix: #351, Is there any special format requirement for processing html data to markdown
• Fix: HTML parser converts a tags in preformatted text to links, should convert to URL only

0.42.12

• Fix: #349, Translation Helper bugs, mix-up between anchors and other non-translating elements after translation.

0.42.10

• Fix: #349, Translation Helper bugs
• Fix: #348, WRAP_AUTO_LINKS defaults to false, Markdown loses a potential useful link

0.42.8

• Add: PDF converter landscape sample PdfLandscapeConverter.java
• Fix: revert to OpenHtmlToPDF version 0.0.1-RC15 which is the last Java 7 byte code version
• Fix: fill missing columns StringIndexOutOfBoundsException when prev cell consists of consecutive pipes without text
• Fix: add explicit Locale.US to String.format() when using %d for integers to prevent conversion to arabic on Java 11.

0.42.6

• Fix: #338, getLineNumber incorrect with Windows end of line separators
  • merge: #339, PR: Fix to line number when using Windows EOL characters. thanks to

0.42.4

• Fix: Trailing URI prefix only auto-links do not get parsed

0.42.2

• Fix: #334, CR line separators don't produce line break nodes,
  • merge: #335, PR: Fix CR-only line separator handling thanks to Kijimuna

0.42.0

• Fix: #332, withOptions forgets about old link resolvers
  • Break: move com.vladsch.flexmark.Extension to com.vladsch.flexmark.util.builder.Extension
  • Fix: IntelliJ Migration contained in migrate flexmark-java 0_40_x to 0_42_0, to use:
    • copy to IntelliJ application settings to migration subdirectory
    • if you have the project which you want to migrate open, then close it
    • open the project in IntelliJ Ultimate or Community
    • update the flexmark-java dependency version to 0.42.0 (or later) and make sure the new library is downloaded/updated in the project.
    • use menu Refactor > Migrate...
    • select migrate flexmark-java 0.42.x to 0.50.0
    • press Run
    • in the refactoring preview tool window that opens hit Do Refactor
  • Add: common builder base to handle unloading extensions
  • Add: BuilderBase.RELOAD_EXTENSIONS default true, if true then will unload loaded extension from the builder when withOptions() is used on: Parser, HtmlRenderer, Formatter or DocxRenderer. If false then will not reload extensions which are already loaded. The latter is slightly faster because loaded extensions are not re-loaded but will not change extension configuration for loaded extensions based on new options.
  • Fix: HtmlParser.Builder constructor with options does not preserve all already loaded extensions and custom api factories
  • Fix: Parser.Builder constructor with options does not preserve all already loaded extensions and custom api factories
  • Fix: Formatter.Builder constructor with options does not preserve all already loaded extensions and custom api factories
  • Fix: DocxRenderer.Builder constructor with options does not preserve all already loaded extensions and custom api factories
  • Add: BuilderBase.UNLOAD_EXTENSIONS, default Extension.EMPTY_LIST, all extensions in the list will be removed. Used only when using withOptions() and providing new options with this key set. Used during testing to remove default extensions.

0.40.34

• Fix: #328, Html2mark - missing newline when paragraph followed by div
• Fix: #331, Ability to replace empty <p> with <br> during html2mark conversion
• Fix: NPE in TableParagraphPreProcessor
• Fix: AutolinkNodePostProcessor processing links out of order causing sequence end/start reversal.

0.40.32

• Fix: compound enumerated references in attributes ids and outside of headings to output with last enumerated reference ordinal.

0.40.30

• Fix: limit EnumeratedReferenceBlock to single line of text without processing other block elements.
• Fix: change Abbreviations to custom block parser from paragraph pre-processor. Allows abbreviation definitions without preceding blank line.

0.40.28

• Add: enumerated reference text in heading to be used with only a format reference, the id is taken from the heading attributes.
• Add: compound enumerated references for creating legal numbering for enumerated references.

0.40.26

• Fix: upgrade dependencies
  • OpenHtmlToPdf -> 0.0.1-RC19
  • docx4j -> 6.1.2
• Add: parse int or default to Utils.java

0.40.24

• Break: test case related classes changed to allow providing URL string for the file resource used in the text instead of relying on heuristic conversion of resource URL to file path.

• Fix: #326, flexmark-html-parser - multiple <code> inside <pre> bug

• Fix: EscapedCharacterExtension disabled for DoNotLinkDecorate nodes if another extension specified its NodePostProcessorFactory as dependent on Text with DoNotDecorate and DoNotLinkDecorate and the extension was added after EscapedCharacterExtension

• Add: Parser.WWW_AUTO_LINK_ELEMENT, default false. If true then strings of the form <www.someurl> will be parsed as auto-links

• Fix: copy Extensions from flexmark-profile-pegdown module to PegdownExtensions in core parser. Old class still available and extends the new named class.

• Add: ParserEmulationProfile.getOptions(DataHolder) to ParserEmulationProfile to allow using current option values in deciding what the settings should be for the profile. Used by ParserEmulationProfile.PEGDOWN

• Fix: #323, TOC generation improvement, profile setting HeaderId generation to false even though PegdownExtensions.ANCHORLINKS is not used

• Add: css option to PDF export and use as default css from #323, TOC generation improvement, thanks to @jvdvegt

  • Add: PdfConverterExtension.DEFAULT_CSS data key with default value of embedded CSS.
  • Add: PdfConverterExtension.embedCss(String html, String css) will embed the css in html by inserting it between <head> and </head> wrapped in <style> and </style>. Does its best to generate valid HTML, use it if your don't want to bother creating your own HTML document with embedded CSS.

  ℹ️ Default CSS is only added if PdfConverterExtension.exportToPdf(OutputStream, String, String, DataHolder) is used to provide options. For all other calls you need to embed the default css into your HTML string before exporting to PDF.

• Fix: HTML parser link/image conversion to ref link and reference when cannot create a valid reference fallback to generating explicit link instead.

• Fix: HTML parser ignoring NONE for FlexmarkHtmlParser.EXT_INLINE_LINK and FlexmarkHtmlParser.EXT_INLINE_IMAGE options

• Fix: HTML parser to add blank line before block quote and aside if they are not the first child of their parent.

• Fix: HTML parser, divs always generated line before and after, instead of before if not first child element and after if not last child element.

• Add: IParse.transferReferences(Document, Document, Boolean), to allow overriding repository KeepType selected copying behaviour for references already defined in the destination document.

• Deprecate: IParse.transferReferences(Document, Document), in favour of IParse.transferReferences(Document, Document, Boolean)

0.40.22

• Fix: merge util tests from @James-Adam and fix bugs
• Fix: change to MutableDataSet.set(DataKey<? extends T>, T)

0.40.20

• Fix: #316, Github user extension incorrectly formats some text
  • Add: test to make sure previous character to @ is not isUnicodeIdentifierPart(), - nor .
• Add: FlexmarkHtmlParser options:
  • Fix: #318, Ability to disable table caption in FlexmarkHtmlParser,
    • Add: TABLE_CAPTION option as a convenience alias for Formatter.FORMAT_TABLE_CAPTION. Add documentation to Extensions: Html To Markdown
  • Fix: #314, Ability to override character replacements map in FlexmarkHtmlParser,
    • Add: TYPOGRAPHIC_REPLACEMENT_MAP option key taking a Map<String,String>, if not empty will be used instead of the bundled map. Any typographic characters or HTML entities missing from the map will be output without conversion. If you want to suppress a typographic so it is not output add "" to its mapped value.
  • Fix: #317, FlexmarkHtmlParser outputs extra newline when converting nested <ol>, <ul> lists
  • Fix: when rendering raw HTML for inline elements, have to escape contained special markdown characters.
  • Add: FOR_DOCUMENT option to use as document into which the generated Markdown will be inserted. This is to allow re-use matching references generated from HTML instead of new ones.
  • EXT_INLINE_LINK and EXT_INLINE_IMAGE option default LinkConversion.MARKDOWN_EXPLICIT, specifies type of link and image conversion to apply: NONE, MARKDOWN_EXPLICIT, MARKDOWN_REFERENCE, TEXT, HTML
• Add: Remove reliance on YouTrack: IDEA-207453 and instead change resource file URL to path in SpecReader and add a message to all example tests with the file URL with :xxx where xxx is the line number of the spec example in the file.
• This will print the source location of the test in the console. Clicking the link goes right to the spec source file and line of the failed test if Awesome Console plugin is installed.

0.40.18

• Add: EmbeddedAttributeProvider documentation and add it to the provider's list by default unless HtmlRenderer.EMBEDDED_NODE_PROVIDER is set to false. Add attributes to nodes in the AST by inserting a EmbeddedAttributeProvider.EmbeddedNodeAttributes node with the desired attributes. See: NodeInsertingPostProcessorSample.java for example.

• Add: document that DocxRenderer emoji with GitHub preferred is not able to download images, compiling ImageUtils library with Java 8 eliminates the problem.

• Add: resource file URL to SpecReader and add a message to all example tests with the file URL with :xxx where xxx is the line number of the spec example in the file.

• This will print the source location of the test in the console. Clicking the link goes right to the spec source file and line of the failed test. Unfortunately this URL will only work in IntelliJ when JetBrains add this to console to handle such file:// URLs: YouTrack: IDEA-207453

• Fix: #310, PR: Change URL of GitHub CDN thanks to @benelog

• Fix: AsideExtension option keys to be dynamic data keys dependent on corresponding Parser block quote options for their defaults.

• ⚠️ This can potentially break code relying on versions of the extension before 0.40.18 because parsing rules can change depending on which block quote options are changed from their default values.

• To ensure independent options for aside blocks and block quotes, set aside options explicitly. The following will set all aside options to default values, independent from block quote options:

  .set(EXTEND_TO_BLANK_LINE, false)
  .set(IGNORE_BLANK_LINE, false)
  .set(ALLOW_LEADING_SPACE, true)
  .set(INTERRUPTS_PARAGRAPH, true)
  .set(INTERRUPTS_ITEM_PARAGRAPH, true)
  .set(WITH_LEAD_SPACES_INTERRUPTS_ITEM_PARAGRAPH, true)
  

• Deprecate: Parser.BLOCK_QUOTE_TO_BLANK_LINE, use more mnemonic Parser.BLOCK_QUOTE_EXTEND_TO_BLANK_LINE

• Deprecate: CustomNode and CustomBlock. Block and Node should be used directly. The library aims to make no distinction between core and extension implementations, these classes add no useful information.

• Deprecate: BaseSequence had old named functions which were misleading and duplicated under proper names:

  • countChars() -> countLeading()
  • countCharsNot() -> countLeadingNot()
  • countCharsReversed() -> countTrailing()
  • countNotCharsReversed() -> countTrailingNot()

• First of all they only counted leading characters which the name did not imply. Second they were duplicated.

• Add character counting functions:

  • countOfAny()
  • countOfAnyNot()

• Fix: DefinitionExtension does not correctly set the child parse column, causing list items to be expecting 1 extra space for child item recognition.

• Add: ExtensionConversion.NONE to suppress any output from corresponding element

• Add: FlexmarkHtmlParser.EXT_MATH, default ExtensionConversion.HTML, for selecting <math> tag processing. For now only MARKDOWN does nothing useful. Later it will be used to convert math ml to GitLab math inline element.

0.40.16

• Fix: EnumeratedReferenceExtension would process GitHub issue link text as enumerated reference. Now enumerated reference must start with a non-digit character to be interpreted as enumerated reference.
• Fix: Reference Text, Reference Id, Link Text and Image Alt Text child text was not trimmed in the AST causing the parent text range to be smaller than the contained children's text range.
• Add: PdfConverterExtension.PROTECTION_POLICY, default null. Set the protection policy for the generated PDF document.
  • Merge: #306, PR: Add password protection support, send appreciations to niksw7
• Add: youtu.be link handling to YouTubeLinkExtension
  • Merge: #305, PR: add new youtube link style support to flexmark-ext-youtube-embedded for youtu.be/xyz(?t=123), send appreciations to jjYBdx4IL

0.40.14

• Fix: AutolinkExtension removing leading Typographic nodes when the first link occurs in text following the typographic node.
• Add: PDF converter sample with non-latin character set rendering information.
• Fix: missing simple_smile emoji cheat sheet shortcut

0.40.12

• Fix: #300, Typography extension breaks some auto links
• Add: TypographicText interface to mark nodes which hold text which is replaced with typographic for rendering but treated as text for decoration processing. For now AutolinkExtension is the only one making use of it to prevent typographic smarts from breaking up a link and causing part of it to be left out of the URL.

0.40.10

• Fix: GitLab inline math parser to allow multi-line inline math elements
• Fix: #295, CoreNodeFormatter does not descend into children on Link nodes , more like kludge, Formatter.OPTIMIZED_INLINE_RENDERING when false and not translating to always render children of link text.

0.40.8

• Fix: change AttributeProviderFactory extensions to eliminate duplicate registrations of factories.
• Fix: AttributesExtension to assign attributes to explicit/refs links/images
• Fix: #299, FlexmarkHtmlParser produces extra empty list item for eclosing </p> element

0.40.6

• Fix: MarkdownTable.appendTable(FormattingAppendable) to set FormattingAppendable.ALLOW_LEADING_WHITESPACE so indentation prefix is not eliminated.
• Fix: FormattingAppendableImpl to not skip pending spaces if FormattingAppendable.ALLOW_LEADING_WHITESPACE is selected, these will be prefixed before indent.
• Add: FlexmarkHtmlParser.SKIP_LINKS, default false. When true links are converted to text part of the link.
• Add: DoNotAttributeDecorate interface to be implemented by all text nodes which do not result in span-like rendering of their text, such as TypographicSmarts, TypographicQuotes and SoftLineBreak but will erroneously create a break in the text and mess up attribute assignment to text spans.
• Add: AttributesExtension.WRAP_NON_ATTRIBUTE_TEXT, default true. When false does not wrap nodes marked as DoNotAttributeDecorate in TextBase. When true, will allow intuitive attribute processing for text containing DoNotAttributeDecorate as single text span instead of one interrupted by the typographic smarts nodes.
• Fix: AttributesExtension incorrectly parsed invalid empty sequences and absorbed the text instead of skipping it. ie. {}, { }, {#} and {.}
• Add: AttributesExtension.USE_EMPTY_IMPLICIT_AS_SPAN_DELIMITER, default false. When set to true will treat {#} or {.}, without embedded spaced, as start attribute span delimiter to mark start of attribute assignment to text between {.} or {#} and the matching attributes element.

0.40.4

• Fix: #295, CoreNodeFormatter does not descend into children on Link nodes
  • Link, LinkRef, Image and ImageRef node formatter renderer now descends into child nodes during all formatting, not just translation formatting.
• Add: Formatter.OPTIMIZED_INLINE_RENDERING default false. When set to true will use previous rendering for links and images which appends the node characters without descending into child nodes.

0.40.2

• Fix: #294, BlankLine nodes within BlockQuote change in behaviour
  • remove blank lines from AST for BlockQuote and AsideBlock nodes if Parser.BLANK_LINES_IN_AST is not set.
• Fix: #292, 0.40.0: JUnit is in the compile scope
  • remove unnecessary ast collecting visitor in DocumentParser
  • add test scope to flexmark pom
• Fix: #293, YamlFrontMatterBlock rendered as markdown does not preserve nested lists
  • Add: YamlFrontMatterValue node containing yaml value(s) inserted as children of YamlFrontMatterNode
  • Fix: change stored key as BasedSequence instead of string, can be retrieved as YamlFrontMatterNode.getKeySequence()
  • Add: code to return List<String> from child nodes of YamlFrontMatterNode
  • Fix: resolve offsets in YamlFrontMatterNode and YamlFrontMatterValue nodes.

0.40.0

• Fix: Java9+ Compatibility, IntelliJ Migration contained in migrate 0_35_x to 0_40_0.xml, to use:
  • copy to IntelliJ application settings to migrations subdirectory
  • if you have the project which you want to migrate open, then close it
  • open the project in IntelliJ Ultimate or Community
  • update the flexmark-java dependency version to 0.40.0 (or later) and make sure the new library is downloaded/updated in the project.
  • use menu Refactor > Migrate...
  • select migrate flexmark-java 0.35.x to 0.40.0
  • press Run
  • in the refactoring preview tool window that opens hit Do Refactor
• ⚠️ In my projects, the IDE can miss some class migrations, especially on files which are not open. Re-running the migration tends to apply the migration. In cases where migration is not applied manual editing will be needed. In my projects Document, Node and Formatter were the problem classes:
  • com.vladsch.flexmark.ast.Document to com.vladsch.flexmark.util.ast.Document
  • com.vladsch.flexmark.ast.Node to com.vladsch.flexmark.util.ast.Node
  • com.vladsch.flexmark.formatter.internal.Formatter to com.vladsch.flexmark.formatter.Formatter
• ⚠️ flexmark-docx-converter and flexmark-pdf-converter cannot be used with Java 9+ modules because the underlying libraries used have package load conflicts that will be resolved in a later release.

• Fix: overlapping packages in different modules, change:

• com.vladsch.flexmark.ast.AllNodesVisitor to com.vladsch.flexmark.util.ast.AllNodesVisitor

• com.vladsch.flexmark.ast.BlankLine to com.vladsch.flexmark.util.ast.BlankLine

• com.vladsch.flexmark.ast.BlankLineContainer to com.vladsch.flexmark.util.ast.BlankLineContainer

• com.vladsch.flexmark.ast.Block to com.vladsch.flexmark.util.ast.Block

• com.vladsch.flexmark.ast.BlockContent to com.vladsch.flexmark.util.ast.BlockContent

• com.vladsch.flexmark.ast.BlockNodeVisitor to com.vladsch.flexmark.util.ast.BlockNodeVisitor

• com.vladsch.flexmark.ast.Content to com.vladsch.flexmark.util.ast.Content

• com.vladsch.flexmark.ast.ContentNode to com.vladsch.flexmark.util.ast.ContentNode

• com.vladsch.flexmark.ast.CustomBlock to com.vladsch.flexmark.util.ast.CustomBlock

• com.vladsch.flexmark.ast.CustomNode to com.vladsch.flexmark.util.ast.CustomNode

• com.vladsch.flexmark.ast.DescendantNodeIterable to com.vladsch.flexmark.util.ast.DescendantNodeIterable

• com.vladsch.flexmark.ast.DescendantNodeIterator to com.vladsch.flexmark.util.ast.DescendantNodeIterator

• com.vladsch.flexmark.ast.Node to com.vladsch.flexmark.util.ast.Node

• com.vladsch.flexmark.ast.Document to com.vladsch.flexmark.util.ast.Document

• com.vladsch.flexmark.ast.DoNotCollectText to com.vladsch.flexmark.util.ast.DoNotCollectText

• com.vladsch.flexmark.ast.DoNotDecorate to com.vladsch.flexmark.util.ast.DoNotDecorate

• com.vladsch.flexmark.ast.DoNotLinkDecorate to com.vladsch.flexmark.util.ast.DoNotLinkDecorate

• com.vladsch.flexmark.ast.KeepTrailingBlankLineContainer to com.vladsch.flexmark.util.ast.KeepTrailingBlankLineContainer

• com.vladsch.flexmark.ast.NodeAdaptedVisitor to com.vladsch.flexmark.util.ast.NodeAdaptedVisitor

• com.vladsch.flexmark.ast.NodeAdaptingVisitHandler to com.vladsch.flexmark.util.ast.NodeAdaptingVisitHandler

• com.vladsch.flexmark.ast.NodeAdaptingVisitor to com.vladsch.flexmark.util.ast.NodeAdaptingVisitor

• com.vladsch.flexmark.ast.NodeIterable to com.vladsch.flexmark.util.ast.NodeIterable

• com.vladsch.flexmark.ast.NodeIterator to com.vladsch.flexmark.util.ast.NodeIterator

• com.vladsch.flexmark.ast.NodeRepository to com.vladsch.flexmark.util.ast.NodeRepository

• com.vladsch.flexmark.ast.NodeVisitor to com.vladsch.flexmark.util.ast.NodeVisitor

• com.vladsch.flexmark.ast.NodeVisitorBase to com.vladsch.flexmark.util.ast.NodeVisitorBase

• com.vladsch.flexmark.ast.NonRenderingInline to com.vladsch.flexmark.util.ast.NonRenderingInline

• com.vladsch.flexmark.ast.Package to com.vladsch.flexmark.util.ast.Package

• com.vladsch.flexmark.ast.ReferenceNode to com.vladsch.flexmark.util.ast.ReferenceNode

• com.vladsch.flexmark.ast.ReferencingNode to com.vladsch.flexmark.util.ast.ReferencingNode

• com.vladsch.flexmark.ast.VisitHandler to com.vladsch.flexmark.util.ast.VisitHandler

• com.vladsch.flexmark.ast.Visitor to com.vladsch.flexmark.util.ast.Visitor

• com.vladsch.flexmark.IParse to com.vladsch.flexmark.util.IParse

• com.vladsch.flexmark.IRender to com.vladsch.flexmark.util.IRender

• Fix: API classes under internal package:

• flexmark:

• com.vladsch.flexmark.internal.inline.AsteriskDelimiterProcessor to com.vladsch.flexmark.parser.core.delimiter.AsteriskDelimiterProcessor

• com.vladsch.flexmark.internal.inline.EmphasisDelimiterProcessor to com.vladsch.flexmark.parser.core.delimiter.EmphasisDelimiterProcessor

• com.vladsch.flexmark.internal.Bracket to com.vladsch.flexmark.parser.core.delimiter.Bracket

• com.vladsch.flexmark.internal.Delimiter to com.vladsch.flexmark.parser.core.delimiter.Delimiter

• com.vladsch.flexmark.internal.ThematicBreakOptions to com.vladsch.flexmark.internal.ThematicBreakParser.ThematicBreakOptions

• com.vladsch.flexmark.internal.HeadingOptions to com.vladsch.flexmark.internal.HeadingParser.HeadingOptions

• com.vladsch.flexmark.internal.BlockParserTracker to com.vladsch.flexmark.parser.block.BlockParserTracker

• com.vladsch.flexmark.internal.BlockQuoteParser to com.vladsch.flexmark.parser.core.BlockQuoteParser

• com.vladsch.flexmark.internal.DocumentBlockParser to com.vladsch.flexmark.parser.core.DocumentBlockParser

• com.vladsch.flexmark.internal.FencedCodeBlockParser to com.vladsch.flexmark.parser.core.FencedCodeBlockParser

• com.vladsch.flexmark.internal.HeadingParser to com.vladsch.flexmark.parser.core.HeadingParser

• com.vladsch.flexmark.internal.IndentedCodeBlockParser to com.vladsch.flexmark.parser.core.IndentedCodeBlockParser

• com.vladsch.flexmark.internal.ListBlockParser to com.vladsch.flexmark.parser.core.ListBlockParser

• com.vladsch.flexmark.internal.ListItemParser to com.vladsch.flexmark.parser.core.ListItemParser

• com.vladsch.flexmark.internal.ParagraphParser to com.vladsch.flexmark.parser.core.ParagraphParser

• com.vladsch.flexmark.internal.ThematicBreakParser to com.vladsch.flexmark.parser.core.ThematicBreakParser

• com.vladsch.flexmark.internal.HtmlBlockParser to com.vladsch.flexmark.parser.core.HtmlBlockParser

• com.vladsch.flexmark.internal.ReferencePreProcessorFactory to com.vladsch.flexmark.parser.core.ReferencePreProcessorFactory

• com.vladsch.flexmark.internal.InlineParserImpl to com.vladsch.flexmark.parser.internal.InlineParserImpl

• com.vladsch.flexmark.internal.BlockStartImpl to com.vladsch.flexmark.parser.internal.BlockStartImpl

• com.vladsch.flexmark.internal.BlockContinueImpl to com.vladsch.flexmark.parser.internal.BlockContinueImpl

• com.vladsch.flexmark.internal.LinkRefProcessorData to com.vladsch.flexmark.parser.internal.LinkRefProcessorData

• com.vladsch.flexmark.internal.CommonmarkInlineParser to com.vladsch.flexmark.parser.internal.CommonmarkInlineParser

• com.vladsch.flexmark.internal.MatchedBlockParserImpl to com.vladsch.flexmark.parser.internal.MatchedBlockParserImpl

• com.vladsch.flexmark.internal.PostProcessorManager to com.vladsch.flexmark.parser.internal.PostProcessorManager

• com.vladsch.flexmark.internal.DocumentParser to com.vladsch.flexmark.parser.internal.DocumentParser

• com.vladsch.flexmark.internal.HtmlDeepParser to com.vladsch.flexmark.parser.internal.HtmlDeepParser

• resource /com/vladsch/flexmark/internal/util/entities.properties to /com/vladsch/flexmark/util/html/entities.properties in flexmark-util module

• flexmark-docx-converter

• com.vladsch.flexmark.docx.converter.internal.DocxRenderer to com.vladsch.flexmark.docx.converter.DocxRenderer

• com.vladsch.flexmark.docx.converter.internal.DocxRendererPhase to com.vladsch.flexmark.docx.converter.DocxRendererPhase

• com.vladsch.flexmark.docx.converter.internal.NodeDocxRendererHandler to com.vladsch.flexmark.docx.converter.NodeDocxRendererHandler

• com.vladsch.flexmark.docx.converter.internal.DocxRendererOptions to com.vladsch.flexmark.docx.converter.DocxRendererOptions

• flexmark-formatter

• com.vladsch.flexmark.formatter.internal.Formatter to com.vladsch.flexmark.formatter.Formatter

• com.vladsch.flexmark.formatter.internal.FormattingPhase to com.vladsch.flexmark.formatter.FormattingPhase

• com.vladsch.flexmark.formatter.internal.MarkdownWriter to com.vladsch.flexmark.formatter.MarkdownWriter

• com.vladsch.flexmark.formatter.internal.NodeFormatter to com.vladsch.flexmark.formatter.NodeFormatter

• com.vladsch.flexmark.formatter.internal.NodeFormatterContext to com.vladsch.flexmark.formatter.NodeFormatterContext

• com.vladsch.flexmark.formatter.internal.NodeFormatterFactory to com.vladsch.flexmark.formatter.NodeFormatterFactory

• com.vladsch.flexmark.formatter.internal.NodeFormattingHandler to com.vladsch.flexmark.formatter.NodeFormattingHandler

• com.vladsch.flexmark.formatter.internal.NodeRepositoryFormatter to com.vladsch.flexmark.formatter.NodeRepositoryFormatter

• com.vladsch.flexmark.formatter.internal.PhasedNodeFormatter to com.vladsch.flexmark.formatter.PhasedNodeFormatter

• com.vladsch.flexmark.formatter.internal.NodeFormatterSubContext to com.vladsch.flexmark.formatter.NodeFormatterSubContext