Changes for version 0.002 - 2016-01-16
- Jean-Damien Durand <jeandamiendurand@free.fr>
- Impl/_top.pm: few things for xtest
- dev+: removed "secure"
- dev+: simplication
- ResourceIdentifier/BNF.pm: commented a useful trace statement -;
- Role/ldap.pm: restore previous percent_decode() and build_percent_encoding_normalizer() arounds
- ResourceIdentifier/BNF.pm: cosmetic typo
- uri_compat removal
- ldap: dev+
- Role/ldap.pm: dev+
- ResourceIdentifier/BNF.pm: fix warning on undef variable
- Accept ambiguous parse if no parse tree value, install any remaining method in read-only by default
- LDAPString decoding
- ldaps
- Role/ldap.pm: default port
- ResourceIdentifier/BNF.pm: use ref() to detect if we push or not
- ldap
- Possibility to overwrite charset by locally setting $MarpaX::Role::Parameterized::ResourceIdentifier::BNF::pct_encoded_default_charset
- https common role
- around on the builder instead of the method
- query_form and query_keywords
- overload fix, encoding charset dependency on spec
- centralized arg2arg handling uri compat mode
- Role/http.pm
- Fixed comment in percent_encode
- generic helpers
- issues seen when testing ftp scheme
- ftp: [0] issue
- add can_scheme
- ftp: dev+
- host(): fix for ipv6
- Role/ftp.pm: removed
- ResourceIdentifier/BNF.pm: server methods dev+
- prepare for server role
- ResourceIdentifier/BNF.pm: path_segments fix
- Impl/Segment.pm: full URI compliance
- path-segments: dev+
- escape/unescape
- perl op optimization
- percent_decode optimization when there is nothing excluded
- escape/unescape: dev+
- Removed and add unescaped structure
- Fix inlined stubs in uri_compat mode
- ResourceIdentifier/BNF.pm: add notion of separator_location
- ResourceIdentifier/BNF.pm: return _escaped_struct members in uri_compat mode
- Move percent_encode in parameterized role, create unescape method
- Move percent_decode in the parameterized role and add escape support
- Impl/_segment.pm add
- ResourceIdentifier/Setup.pm: use || instead of // in methods returning a string
- ResourceIdentifier/Setup.pm: use || instead of // in default_query_form_delimiter()
- ResourceIdentifier/Setup.pm: default_query_form_delimiter()
- dev+: Remove $__PACKAGE__ variable
- path_query: dev+
- secure(): dev+
- rel(): finally a version that looks ok
- rel: dev+
- ResourceIdentifier/BNF.pm; work on rel()
- dev+: rel()
- dev+: rel_normalized
- dev+: $setup->abs_normalized base(), not $self->abs_normalized_base()
- dev+: optionnaly use normalized base in abs()
- use InstanceOf($top) instead of does(__PACKAGE__)
- dev+: abs() and rel()
- clone revisited
- fix fields list in case of bnf extension
- Make plugins independant
- Role/ftp.pm: the ftp role inherits from the _generic role
- ResourceIdentifier/BNF.pm: support of extension, provide role parameter clone
- Impl/_top.pm: protect against missing argument in new()
- dev+: Role/_generic is with Role/_common (in particular because of build_character_normalizer
- dev+: removed after input modifier in BUILD, fix $is_common internal variable
- dev+: common and generic types become _common and _generic, respectively
- dev+: let the implementation deal with case in can_scheme()
- Externalized starting rule
- dev+: Role/ftp.pm
- Role/ftp.pm: ftp role
- Impl/_top.pm: make sure scheme is normalized before trying the subclass, fix subclass usage
- ResourceIdentifier/BNF.pm: convenience proxies normalized_scheme(), normalized_opaque() and normalized_fragment
- ResourceIdentifier/BNF.pm: parse() method is public
- Move _parse in the parameterized role, follow RFC3987 revisited architecture
- dev+: Externalize current_location and parent_location, and move remove_dot_segments in the parameterized role
- ResourceIdentifier/BNF.pm; leading dots cont
- ResourceIdentifier/Setup.pm: abs_remote_leading_dots
- Impl/_top.pm: better eventual scheme argument processing
- dev+: suspend support of remote leading dots in abs(), not obligatory
- dev+: remove_dot_segments regexp simplification
- dev+: Tracking of remote leading dots in excess
- dev+: In abs(), do not croak if ref or abs are not of generic type
- dev+: remote_leading_dots preparation
- dev+: fix some writing style, _structs access, abs() optimization
- dev+: eq method installed in the top package
- dev+: as_string
- dev+: renamed internal variable *_with_accessors to *_call_lazy_builder to make clear why this exist
- dev+: back to hashes for performance reason - this is a boost -;
- dev+: oups almost a revert of previous commit
- ResourceIdentifier/BNF.pm: removed an non-needed new()
- dev+: Cosmetic rearrange of use/with
- dev+: put instance methods in the parameterized code
- dev+: $URI::ABS_ALLOW_RELATIVE_SCHEME
- dev+: canonical(), access to some internals but not all
- t+: use is_absolute on base object in abs()
- t+: uri_compat at runtime and is_absolute support
- dev+: revisited engine
- dev+: Fix BUILDARGS
- dev+: when calling the builders do nothing else but that
- dev+: object construction optimization
- dev+: make sure the lazy builders are called after object construction
- dev+: add reg_name_is_domain_name in slurpy check, removed lazy on _normalizer_wrapper and _converter_wrapper
- dev+: Perl op optimization
- dev+: Algorithmic optimization for normalizers
- dev+: clone of a MooX::Struct is faster than a new
- dev+: cache symbol_display_form
- dev+: profiling bis
- dev+: profiling
- dev+: fix bug in remove_dot_segments
- dev+: perl op optim
- dev+: abs() using an object instead of a string for absolute reference
- dev+: abs()
- dev+: Generic.path is never undef
- dev+: Removed a logging statement
- dev+: inlined version of percent_encoding_normalizer
- dev+: inlined sub
- dev+: pretty Data::Printer'ing
- dev+: MooX::Struct stringification
- dev+: store explicitely final value after all normalizations
- dev+: back to after instead of around for input
- dev+: Types simplification
- t+: BUILDARGS centralization
- dev+: remove some unused variables
- dev+: incorporate BUILDARGS in BNF
- dev+: has_recognized scheme is ro
- dev+: overload
- dev+: prefer around input instead of after input
- dev+: externalized unescape method, fo not use trigger but explicit parsing at BUILD or after input is set
- dev+: build_path_segment_normalizer
- dev+: moved some specs from _common to _generic
- dev+: variables renaming
- dev+: build_percent_encoding_normalizer
- dev+
- dev+: normalizers fix
- dev+: default_port
- dev+: indeed, _indice_description label fix
- dev+: Make _indice_description an attribute
- dev+: cosmetic change in callback names array
- dev+: normalization and convertion ladders optimization, do it once also before parsing
- dev+: normalization before parsing
- dev+: use Undef instead of defined
- dev+: removed unused files
- dev+: use centralzed BUILDARGS
- dev+: fixes
- t+
- dev+: factorization
- BNF role refactorization
- dev+: without Data::Dumper -;
- dev+: use Type::Params
- dev+: new_abs() and removal of all path derivations
- t+: new_abs
- dev+: removed unused use_utf8 constraint
- dev+: fix typo warnff
- dev+: is_absolute method
- Parameterized/ResourceIdentifier.pm: more generic, support of URI_ENCODED and IRI_ENCODED callbacks
- ResourceIdentifier/BUILDARGS.pm: a little bit simplified
- ResourceIdentifier/_top.pm: uri_compat
- dev+: BUILD
- dev+: Removal of all __DATA__ section
- dev+: Install default (empty) normalizers in _common
- dev+: use _common and _generic instead of Common and Generic
- dev+: Fixed structure members accessors
- dev+: moved all Generic normalizers in the role
- dev+: Added escaped section
- dev+: path segment normalization
- dev+: add decoder
- dev+: get once
- dev+: add external field name when calling normalization
- dev+: keep octets and al. information
- dev+: Fix BUILARGS and always do a copy of %args before jumping to subclasses
- dev+: Fix _structs initialization
- dev+: deal with inheritance
- dev+: use Data::Dumper to see Character Normalization effect in trace mode
- dev+: Removal of unused "normalize" method and require "is_character_normalized", "character_normalization_strategy" in the generic class
- dev+: The return of uri_compat
- dev+: More generic constructor
- dev+: put normalizer in the generic instance, not in the grammar and use $self as argument to Marpa::R2s value()
- dev+: removed non-needed ";"s
- dev+: removed unused escape method
- dev+: descriptio of indices factorization
- dev+: fixed comments
- dev+: changed internal indices
- dev+: normalization
- dev+: normalization fix
- dev+: normalizer
- dev+: isdn, nfc
- dev+: eq and normalization
- dev+: revmoed fragments
- dev+: use setters
- dev+: explicits rw (for the moment)
- dev+: segments defaults to "" as per URI
- dev+: escape/unescape modes
- dev+: fix mode without logging
- dev+: engine fix
- More attributes in BNF
- use ${^MATCH} instead of $1 in percent_encode()
- Renamed Singleton to Grammars
- A lor of changes + wrapping Marpa log correctly (finally!)
- Use of to have logging or not
- Factorization of different new methods
- TRUE and FALSE constants
- Revisited scheme parameter check
- removed some catch{} used for debugging
- use Type::Tiny for parameter type checking
- Removed a confusing file
- dev+: _uri_compat
- dev+: less hacks
- dev+: a bit better trace -;
- Structures generic accessors and Logging category fix
- Add URI, URI_reference, absolute_URI and relative_ref
- Make public method required
- soft revert Parameterized/ResourceIdentifier.pm: Note saying caller is required to use MooX::ClassAttribute
- Parameterized/ResourceIdentifier.pm: Note saying caller is required to use MooX::ClassAttribute
Modules
Resource Identifier BNF role
Resource Identifier: segment implementation
Resource Identifier: top level implementation
Marpa Trace Wrapper
Resource Identifier: Common syntax semantics
Resource Identifier: Generic syntax semantics role
Resource Identifier setup
Type tools for Resource Identifiers as per RFC3986 and RFC3987