csv-parse

• package: move pad dependency to dev

• relax_column_count: new less and more options
• columns: skip empty records before detecting headers
• errors: rename CSV_INCONSISTENT_RECORD_LENGTH
• errors: rename CSV_RECORD_DONT_MATCH_COLUMNS_LENGTH

New Feature:

• on_record: user function to alter and filter records

Minor improvements:

• test: ensure every sample is valid
• from_line: honours inferred column names
• from_line: new sample
• errors: expose CSV_INVALID_ARGUMENT
• errors: expose CSV_INVALID_COLUMN_DEFINITION
• errors: expose CSV_OPTION_COLUMNS_MISSING_NAME
• errors: expose CSV_INVALID_OPTION_BOM
• errors: expose CSV_INVALID_OPTION_CAST
• errors: expose CSV_INVALID_OPTION_CAST_DATE
• errors: expose CSV_INVALID_OPTION_COLUMNS
• errors: expose CSV_INVALID_OPTION_COMMENT
• errors: expose CSV_INVALID_OPTION_DELIMITER
• error: fix call to supper

Project management:

• package: contributing
• package: code of conduct

• context: column is null when cast force the context creation, fix #260

• errors: don't stringify/parse undefined and null values
• errors: expose CSV_NON_TRIMABLE_CHAR_AFTER_CLOSING_QUOTE
• errors: expose CSV_MAX_RECORD_SIZE

• lint: integrate eslint

• context: null column when columns number inferior to record length

• src: set const in for loop

• skip_lines_with_empty_values: handle non string value
• errors: add context information
• tests: new error assertion framework
• buffer: serialize to json as string
• errors: expose INVALID_OPENING_QUOTE

• errors: start normalizing errors with unique codes and context
• errors: expose CSV_INVALID_CLOSING_QUOTE
• errors: expose CSV_QUOTE_NOT_CLOSED
• errors: expose CSV_INVALID_RECORD_LENGTH_DONT_PREVIOUS_RECORDS
• errors: expose CSV_INVALID_RECORD_LENGTH_DONT_MATCH_COLUMNS
• errors: expose CSV_INVALID_COLUMN_MAPPING

• travis: remove node.js 8 and add 12
• destroy: test inside readable event

• security: remove regexp vulnerable to DOS in cast option, npm report 69742

• ts: add buffer as allowed type for input, fix #248

• package: latest dependencies
• bom: detection when buffer smaller than bom
• package: remove deprecated @types/should dependency
• package: update file path

• package: fix files declaration

• bom: parsing for BOM character #239
• ts: add sync definition
• package: replace npm ignore with file field

Fix:

• columns: allows returning an array of string, undefined, null or false

New features:

• options: new bom option

• columns: enrich error message when provided as literal object
• cast: handle undefined columns
• skip_lines_with_error: new sample

Fix: columns: fix es5 generation

Fix:

• columns: immutable option

Minor enhancements:

• ts: distribute definitions with es5
• ts: unused MatcherFunc type

Project management:

• babel: include .babelrc to git

New features:

• objname: accept a buffer

Minor enhancements:

• to_line: validation refinements
• trim, ltrim, rtrim: validation refinements
• to: validation refinements
• from_line: validation refinements
• objname: validation refinements
• from: validation refinements
• escape: validation refinements
• skip_empty_lines: validation refinements
• skip_lines_with_empty_values: validation refinements
• skip_lines_with_error: validation refinements
• relax_column_count: validation refinements
• relax: validation refinements
• delimiter: validation refinements
• max_record_size: validation refinements

Fix:

• record_delimiter: fix multi bytes with skip_empty_lines and from_line
• rtrim: accept tab

New features:

• options: accept snake case and camel case
• cast: dont call cast for non column-mappable fields

Fix:

• cast: ensure column is a string and not an array
• stream: handle empty input streams
• cast: function may return non-string values
• stream: pass stream options without modification

Fix:

• relax_column_count: handle records with more columns

This is a complete rewrite based with a Buffer implementation. There are no major breaking changes but it introduces multiple minor breaking changes:

• option rowDelimiter is now record_delimiter
• option max_limit_on_data_read is now max_record_size
• drop the record event
• normalise error message as {error type}: {error description}
• state values are now isolated into the info object
• count is now info.records
• lines is now info.lines
• empty_line_count is now info.empty_lines
• skipped_line_count is now info.invalid_field_length
• context.count is cast function is now context.records
• drop support for deprecated options auto_parse and auto_parse_date
• drop emission of the record event
• in raw option, the row property is renamed record
• default value of max_record_size is now 0 (unlimited)
• remove the record event, use the readable event and this.read() instead

New features:

• new options info, from_line and to_line
• trim: respect ltrim and rtrim when defined
• delimiter: may be a Buffer
• delimiter: handle multiple bytes/characters
• callback: export info object as third argument
• cast: catch error in user functions
• ts: mark info as readonly with required properties
• comment_lines: count the number of commented lines with no records
• callback: pass undefined instead of null

API management:

• Multiple tests have been rewritten with easier data sample
• Source code is now written in ES6 instead of CoffeeScript
• package: switch to MIT license

• max_limit_on_data_read: update error msg
• src: simplify detection for more data
• lines: test empty line account for 1 line
• options: extract default options
• package: add a few keywords
• src: precompute escapeIsQuote
• travis: test against Node.js 11

• rowDelimiter: fix overlap with delimiter
• internal: rename rowDelimiterLength to rowDelimiterMaxLength

• readme: fix links to project website

• src: generate code

• package: move to csv.js.org
• samples: new cast sample
• package: upgrade to babel 7
• samples: new mixed api samples
• samples: new column script
• samples: update syntax
• package: improve ignore files

Breaking changes:

• columns: skip empty values when null, false or undefined

Cleanup:

• sync: refactor internal variables
• index: use destructuring assignment for deps

• typescript: make definition header more relevant

• to: ignore future records when to is reached

• trim: after and before quote
• tests: compatibility with Node.js 10
• trim: handle quote followed by escape
• parser: set nextChar to null instead of empty
• travis: run against node 8 and 10

• cast: pass the header property
• auto_parse: deprecated message on tests
• cast: inject lines property

• cast: deprecate auto_parse
• auto_parse: function get context as second argument

• skip_lines_with_error: DRYed implementation
• skip_lines_with_error: Go process the next line on error
• events: register and write not blocking
• test: prefix names by group membership
• events: emit record
• raw: test to ensure it preserve columns
• package: latest dependencies (28 march 2018)
• raw: ensure tests call and success
• package: ignore npm and yarn lock files
• sync: handle errors on last line

• package: move babel to dev dependencies

• package: es5 backward compatibility
• package: ignore yarn lock file

• package: only remove js files in lib
• source: remove unreferenced variables #179
• package: start running tests in preversion
• package: new release workflow

This major version use CoffeeScript 2 which produces a modern JavaScript syntax (ES6, or ES2015 and later) and break the compatibility with versions of Node.js lower than 7.6 as well as the browsers. It is however stable in term of API.

• package: use CoffeeScript 2

• package: revert to CoffeeScript 1

Irrelevant release, forgot to generate the coffee files.

• package: preserve compatibility with Node.js < 7.6

• options: auto_parse as a user function
• options: auto_parse_date as a user function
• test: should require handled by mocha
• package: coffeescript 2 and use semver tilde
• options: ensure objectMode is cloned

• relax_column_count: honors count while preserving skipped_line_count
• api: improve argument validation

• sync: catch err on write

• relax: handle double quote

• src: group state variable initialisation
• package: update repo url
• quote: disabled when null, false or empty
• src: remove try/catch
• src: optimize estimation for row delimiter length
• lines: improve tests
• src: use in instead of multiple is
• src: string optimization

• skip default row delimiters when quoted #58
• auto_parse: cleaner implementation
• src: isolate internal variables

• options: new to and from options

• rowDelimiters: fix all last month issues

• regression with trim and last empty field #123

• rowDelimiter: simplification
• fix regression when trim and skip_empty_lines activated #122
• auto_parse = simplify internal function

• src: trailing whitespace and empty headers #120
• rowDelimiter: adding support for multiple row delimiters #119
• Remove unnecessary argument: Parser.prototype.__write #114