- [Select] Support escaped characters in CSS selector names, idents, and strings
- [Select] Support Elixir-style unicode code points in CSS selector names, idents, and strings
- [Select] Add better errors when parsing CSS selectors
- [Parse] Deprecate parsing tuple trees with
parse/1
- [Parse] Add
:tuple_treetype toparse/2
- [Parse] Update to
meeseeks_html5ever v0.11.1, which returns a better error when provided with non-UTF-8 input - [Parse] Return parser errors if parsing an invalid tuple tree
- No longer support Elixir 1.3 (minimum tested compatibility is now Elixir 1.4 and Erlang/OTP 19.3)
- Support Elixir 1.8
- [Parse] Update to
meeseeks_html5ever v0.11.0, which is faster and more memory efficient on Erlang/OTP 21
- [Meta] Test more Elixir+OTP combinations with Travis CI
- [Parse] Update to
meeseeks_html5ever v0.10.0, which supports OTP 21
- [Select] Remove optimization in
Select.handle_matchthat could indirectly cause matches stored in the context for filtering to be prematurely cleared
- [Select] Fix error in how context was updated in
Select.filter_nodes - [Select] Fix error in how context was updated in
XPath.Expr.Step.eval - [Select] Fix error in how nodes were filtered in
XPath.Expr.Step.eval - [Select] Include filters when transpiling absolute XPaths to root selectors
- [Parse] Update to
meeseeks_html5ever v0.9.0, which resolves a Dialyzer error
- [Select] The
cssandxpathmacros now accept vars
- [Select] Fix inconsistency in
Document.get_nodes/1 - [Select] Fix bug in
Document.get_nodes/2, courtesy of @asonge - [Select] Fix various typespecs, courtesy of @asonge
- [Errors] Returned and raised errors throughout the project have been updated to use
Meeseeks.Errorinstead of whatever assorted formats they were using before
- [Errors] Add
Meeseeks.Error, a generic error struct implementingException - [Select] Add
Meeseeks.fetch_allandMeeseeks.fetch_one
- [Extract] Fix bug in
Meeseeks.htmlwhen encoding element attribute values that contain double quotes
- [Select] Most
Documentfunctions now raise if an unknownnode_idis provided, when before they might have raised or might have handle the situation gracefully - [Select] Add
get_root_ids/1,get_node_ids/1, andfetch_node/2toDocument - [Select] Add
Document.delete_note/2, courtesy of @willbarrett - [Readability] Remove Credo
- [Readability] Add .formatters.exs and
mix formatproject
- [All] Fix various typespecs
- [Parse] Update to
meeseeks_html5ever v0.8.1, which supports OTP 20.2
- [Parse] Update to
meeseeks_html5ever v0.8.0, which removes panics related to callingmark_script_already_startedandget_template_contents, and removes synchronous parsing, which did not correctly handle panics and broke the <1ms contract on first call - [Select] Update select functions to propagate parse errors
- [Parse] Update to
meeseeks_html5ever v0.7.0, which fixes an erroneous panic related to callingremove_from_parenton a node with no parent
- [Extract] Update extractors to propagate nil input
- [Select] Fix CSS tokenization bug related to ab formulas
- [Select] Fix Elixir 1.5 related warnings
- [Extract] Add
Document.html/1andDocument.tree/1extractors - [Extract] Update
Meeseeks.html/1andMeeseeks.tree/1to accept aDocument - [Extract] Update the extraction functions to return a better error when provided with invalid input
- [Parse] Fix doctype parsing
- [Parse] Update to
meeseeks_html5ever v0.6.1, which supports OTP 20
- [Parse] Update to
meeseeks_html5ever v0.6.0, which supports parsing XML - [Parse] Add
Meeseeks.parse/2which takes either:htmlor:xmlas the second argument to specify how the source gets parsed - [Extract] Update
Meeseeks.data/1to handle CDATA when parsing HTML
- [Select] Rename
Context.new/1toContext.prepare_for_selection/1 - [Select] Rename
Context.with_accumulator/2toContext.add_accumulator/2
- [Parse] Update to
meeseeks_html5ever v0.5.0 - [Parse] Parse
Document.ProcessingInstructionnodes from tuple-trees - [Select] Support
processing-instructionfunctionality inMeeseeks.XPath(when possible) - [Select] Add a
Document.ProcessingInstructionnode type - [Select] Add
Select.select/3andMeeseeks.select/3 - [Select] Add
Context.ensure_accumulator!/1
- [Select] Change the
Selector.match?/3callback toSelector.match/4, which now takes a context and can return a{boolean, context}tuple in addition to returning a boolean.
- [Select] Add XPath selector support (see
Meeseeks.XPath) - [Select] Add
Selector.filters/1callback to theSelectorbehaviour and update selection to allow for filtering matches before proceeding - [Select] Add
Meeseeks.Contextto allow selectors and the selection process to store state - [Select] Add
Meeseeks.Accumulatorbehaviour and updateAccumulator.{All, One}to use it - [Select] Add
NodeandRootselectors - [Select] Add
Ancestors,AncestorsOrSelf,Children,Descendants,DescendantsOrSelf,NextSiblings,Parent,PreviousSiblings, andSelfselector combinators - [Select] Add
parent,ancestors, andprevious_siblingsqueries toDocument
- [Parse] Update to
meeseeks_html5ever v0.4.6, which correctly parses namespaced elements and doesn't try to bring inhtml5ever 0.16.0 - [Extract] Fix
htmlextractor to add namespaces to elements - [Usability] Improve
DocumentandResultopaque inspected values
- [Meta] Add CI via Travis CI
- [Parse] Update to
meeseeks_html5ever v0.4.4, which permits Elixir 1.3 - [Select] Fix CSS tokenization bug involving wildcard or pseudo-class descendants
- [Parse] Replace
html5ever_elixirwithmeeseeks_html5ever - [Select] Allow CSS selector
:not()to accept multiple selectors
- [Parse] Move
Document.new/1toParser.parse_tuple_tree/1 - [Select] Fix ordering in
Document.get_nodes/1andAccumulator.return/1(forAccumulator.All)
- [Extract] Add new
datasetextractor that mimics the HTMLElement.dataset API - [Usability] Raise a better error when trying to select with a string instead of selectors