[libppi-perl] annotated tag upstream/1.236 created (now 3038eb0)
Salvatore Bonaccorso
carnil at debian.org
Sun Jun 25 08:49:56 UTC 2017
This is an automated email from the git hooks/post-receive script.
carnil pushed a change to annotated tag upstream/1.236
in repository libppi-perl.
at 3038eb0 (tag)
tagging 2f51bb899806a35c3ad37d467a4f9acf7838e489 (commit)
replaces upstream/1.220
tagged by Salvatore Bonaccorso
on Sun Jun 25 10:24:06 2017 +0200
- Log -----------------------------------------------------------------
Upstream version 1.236
Adam Kennedy (178):
Importing PPI from the SourceForge CVS
Moving t.data to t/data
Compensating for the test data relocation
Updating template
Correcting newlines
Adding a fix to support power operators
Fixing glob cast vs operator
Fixed regressions
Preparing for release
Preparing for an emergency release
Another fix
Adding a skip for strange locales
Noting locale test
Incrementing version to 1.112 for release
Fixing the wording for the unicode-incompatibility message
Removing the use of auto_install
Incrementing some heavily used modules to remove the use of AutoInstall
Incrementing version
Move from IO::Scalar to IO::String.
Incrementing version for release
Correcting test
Minor bug fix
Noting bug fix
Adding test for until () { }
Adding until () { } support
Adding until support
Finally fixing until
Noted fix to until bug
Adding a regression test for bug #19629
Adding support for list statements
Correcting the dump
Correcting the dump
All tests pass
Adding test for ->prune
Adding more tests
Fixing ->prune
Fixing prune
Updating VERSION to 1.115
Correcting bad newlines
Comment tweak
Adding PPI::Document::File
Added readonly document attribute
Fixing tests
Applied #16892: [PATCH] docs and comments
Fixed a range of bugs
Fixed #19629 (End of list mistakenly seen as end of statement)
Incrementing version for release
Fixed a number of bugs
Add the dor operator
Adding support for .123 decimal numbers
Preparing for release of 1.118
Work around the need for a #!/use/bin/perl line
Fixing #19999: Make test fails (undeclared global variable $document) under Perl 5.6.2
Tweaking the test to also cover the leading whitespace bug
Tweaked a test slightly
Noting some changes
Stronger enabling of warnings during tests
Noting Module::Install upgrade
Fixing a bug in PPI::Document::File->new.
Adding literal support for PPI::Token::Quote::Single (i.e. I did the easiest one) :)
Some minor documentation fixes
Some minor documentation fixes
Some minor documentation fixes
Adding tests for invalid nesting
Use the more-correct _INSTANCE test
Renaming a sample
Adding nested statement and nested structure tests, and fixing some uncovered bugs.
Enable tainting for inline tests
Add elements to the tree BEFORE we recurse into them, so that we can do more-complex detections, like for map( { $_ =>
Setting to dev version for release
Added initial timeout support
Tweak
Preparing for next dev release
Preparing for release
Adding new modules from laptop
Removing "All rights reserved"
Incrementing pip version, changing the path to the repository
Make the tests readonly, clean up the quote engine contributed tests
Adding more optimizations, support for \EOF, and some bug fixes
Adding regression test for \EOF
Preparing for release
Correcting newlines, preparing for release
Moving the 99 test to 99_author.t
Tweaks
Adding another regression case
Adding a regression case
Bugs fixed
Updating version to 1.199_05
Newline corrections
Some minor fixes to some tests
Minor tweak
Updating for release
Minor bug fix
Incrementing version for release
Correcting the base call
Incrementing version and setting Changes for release
Removing "use base 'Exporter';"
Preparing back-compatible release
Typo fix
Adding some operators and magic variables for Perl 5.10.0
Converting every use of build_requires: to test_requires: for future compatibility with the FreeBSD test_requires metadata upgrade
Removing a no warnings
Various additions
Adding some missing files
Getting closer to release
PPI changes all work
Preparing for release
Correcting the copyright
Correcting the release date
Updating to the production version
Adding changes for PPI
Adding empty document
Adding missing regression
Bug fixes for release
Updating changes for release
Explicitly copying a capture
Starting on more thorough complete implementation.
Bits and pieces
Adding a ton of missing files before my hard drive dies
Removed a literal match and converted to a resolved set
Various bug fixing
All of the basic forms of for(each) qw{foo} should work now
Fixed 05_lexer.t
More work on compound statements
Correcting $/s --> \z/s
Test suite passes
Preparing for dev release
Updating copyright for release
Various look and feel cleanups.
Renamed PPI::Statement::Switch to PPI::Statement::Given
Major lexer optimisations
Adding inline tests for the lexer
Implementing PPI::Transform::UpdateCopyright
Partially complete Class::XSAccessor integration
Another pre-release
Updating $VERSION for release
Nudging dependencies
Updating $VERSION for release
Preparing for the next release candidate
Changed ::ForLoop to ::For
Fixing the prototype
An attempt at a slight speed up
Layout tweaks
Layout tweak
Incrementing version for release
Updating version
Prep for release
First pass at making PPI incrementally faster via data uncovered by NYTProf 3
Fixed an introduced bug
#50309: literal() wrong result on "qw (a b c)"
Preparing for dev release
Preparing for release
Preparing for production release
Some small performance hacks
Only run the API test during release testing
Small optimisation release
Packaging fixes
Adding experimental tokenizer support.
Added extra charset case
A variety of bug fixes
Fixed #54208: PPI::Token::Quote::Literal::literal is missing due to case-sensitivity error
Adding missing regression files
Added missing method
Fixed RT #48819
Updating version
Updating for release
Adding a custom minimumversion test, ignoring the tests that are smart enough to skip themselves when there are high Perl versions.
Fix for PPI::Token::Pod::merge, plus some housecleaning
Getting a couple of WYANT's bug fixes into a release, and applying one more of my own
Added missing regression test
Prepare for new release
Removed pointless use of List::MoreUtils
Ready for 1.215 release
- Fixed RT #27475: PPI parses regex incorrectly
Bumping version to 1.216_01
Merge pull request #6 from dsteinbrunner/patch-1
Merge pull request #3 from adamkennedy/inaccurate_PDOM_examples
Merge pull request #2 from adamkennedy/xsaccessor_shadowing_fix
Chas. J. Owens IV (1):
make Makefile.PL write README and README.md
Chris Capaci (1):
returning from function if $word isn't defined before using it causes a failure.
Chris Dolan (65):
Doc fixes: 1) svn instead of CVS, 2) 2 typos
RT 19629, part II:
Bugfix: decision about whether location computation had already begun
Bugfix: 'goto' is a PPI::Statement::Break trigger
Update 'Changes' for post-1.115 work
Add a failing regression test for
Oops, revert r865. I got confused about content() vs. serialize()
Add regression test for '1-1', where '-' should be an Operator
Add a base() method to PPI::Token::Number which returns a numeral
Add failing tests for two new regressions:
Fixed RT #21571: symbol() mistakenly returns just sigil
Fixed RT #21575: variables() chokes on my list with whitespace
Add failing regression test for RT #20480: misparse of '.1234' as '.'
Fix POD spelling errors (ran through Test::Spelling)
Stop using unlocalized $_ in _process_next_char (from part of RT #16952)
Add failing test for "/=" which gets parsed as two operators, "/" and "="
Fix parsing for *= and /= (two separate bugs)
Refactor duplicated file.code/file.dump testing into a subroutine
Work on PPI::Token::Number coverage
Fixes and tests for tokenizing invalid binary, octal and hex digits.
Note binary, oct, hex fixes in changelog
Fix a trivial thinko in the dor patch that broke pragma() and thus
Tiny POD fix for t/99_pod.t failure
Got test coverage of PPI::Token::Number to 97%
Fix: -.0 is a number (Unknown needed to know about "-.")
Two vaguely-related parse errors:
Break out Octal, Hex and Binary into their own PPI::Token::Number subclasses
Document the new ::Number subclasses in the PDOM class list
Break out Number::Float and Number::Version into their own subclasses.
Remove some duplicated code, fix some documentation
Remove an unneeded 'use Test::Object'
Implement exponential notation numbers
Implement the literal() method for PPI::Token::Number and subclasses.
More $token->literal() tests, this time for expected failures
Trivial doc spelling corrections
Localize all filehandle globs
Add more t/data files for rountrip.t to test
Change all unlocalized $_ assignments to lexical vars
Minor test tweaks, disentangled from uncommitted underscore patch
Oops, paste-o
Add incomplete heredoc to test code
seperator -> separator
Failing test to distinguish
Add a failing test: statement is misclassified as ::Compound
Fixed incorrect location() for PPI::Structure instances
Failing test: hash of hashes, internal hash interpreted as a block
Failing test for RT 23253
expand the RT 23253 failing test a little bit
Typo: modifieds -> modifiers
Fix pragma() to allow 'use utf8;'
Added failing unit tests:
Fix regressions:
Fix for $^\w and %^H
Fix t/data/08_regressions/14b_minus.code
Correct a misconception on my part -- returning undef from
Document PPI::Token::Quote::literal()
Added literal() to PPI::Token::DashedWord
My previous $^X fix accidentally broke $^WIDE_SYSTEM_CALLS and other
Trivial fix for RT #30469 (calling length() on PPI::Token gives error)
Failing test for "sub foo: Attr {}". If you add a space before the
Add a failing test for "label :" which Perl interprets the same as
Implement support for "label\s*:" syntax interpreted as PPI::Token::Label
Workaround for
Simple code cleanup: use -1 to endex end-of-array
Unicode BOM support
Christian Walde (94):
RT #90792: PPI::XSAccessor overwrites indexing of actual source modules on CPAN
RT# 69026: PPI docs mention PPI::Token::Bareword, but it does not exist
add link to the github repo
clarify that some variables called $rest have nothing to do with earlier scoped instances
prevent a parse of "a " x $n becoming pathological with increasing $n (label detection)
prevent a parse of "'a'" x $n becoming pathological with increasing $n
prevent a parse of "a " x $n becoming pathological with increasing $n (bareword extraction)
golf PPI::Lexer::_add_element
speed up parsing of __DATA__ segments, by processing line-, not char-wise
allow parsing of files of any size, now that we know how to fix performance
prevent a parse of '$a ' x $n becoming pathological with increasing $n (symbol extraction)
prevent a parse of 'or ' x $n becoming pathological with increasing $n (forced bareword detection)
prevent a parse of '$#a ' x $n becoming pathological with increasing $n (array index tokenization)
prevent a parse of '-a ' x $n becoming pathological with increasing $n (dashed word tokenization)
prevent a parse of '<<"a" ' x $n becoming pathological with increasing $n (heredoc start tokenization)
prevent a parse of '-a::a ' x $n becoming pathological with increasing $n (charwise word tokenization)
prevent a parse of '{}<>' x $n becoming pathological with increasing $n (glob detection)
prevent a parse of '1x ' x $n becoming pathological with increasing $n (x operator detection)
prevent a parse of 'sub() ' x $n becoming pathological with increasing $n (prototype tokenization)
create mgc version of $CURLY_SYMBOL_MGC for coming optimizations
prevent a parse of '*{' x $n becoming pathological with increasing $n (magic vs glob cast)
prevent a parse of '${' x $n becoming pathological with increasing $n (magic scalar detection)
prevent a parse of '@{' x $n becoming pathological with increasing $n (magic array detection)
prevent a parse of '%{' x $n becoming pathological with increasing $n (magic hash detection)
prevent a parse of '$${' x $n becoming pathological with increasing $n (magic deref detection)
prevent a parse of '${' x $n becoming pathological with increasing $n (magic control value detection)
replace old curly_symbol with new mgc curly_symbol
prevent a parse of 'v1a ' x $n becoming pathological with increasing $n (v-string detection)
prevent a parse of 'a<<' x $n becoming pathological with increasing $n (heredoc/bit-shift detection)
prevent a parse of 'sub a:a();' x $n becoming pathological with increasing $n (parametrized attribute tokenization)
prevent a parse of 'q a' x $n becoming pathological with increasing $n (gap operator separator detection)
prevent a parse of 'q{}' x $n becoming pathological with increasing $n (quote brace detection)
remove _scan_for_character as it was never used
Make some more subs constant that can be for speed.
reinstate some unix newlines that got lost earlier
make LICENSE permanent in the repo since i don't have the automation
generate README via MI plugin
updates for dev release 1.217_01
updates for dev release 1.218
add .gitignore
updates for dev release 1.219_01
updates for release 1.220
little bit of a readability improment in PPI::Token::HereDoc
typo fix in docs
refactor ppi_token_heredoc.t for easier adding of TODO blocks
add tests for END/DATA segments at end of document
add test for preventing recognition of bareword with ' separator after truncated decimal number as eq
extract check_with into helper module
add tests for methods of Token::QuoteLike::Regexp
add tests for Token::Regexp
add tests for Token::Structure
tests for float+op
ensure Document->save does not mangle newlines on windows
keep readme generation from mangling newlines on windows
updates for dev release 1.221_01
don't plan tests for a file that's skip_all
document Statement::Package->version as being part of the API
updates for dev release 1.221_02
prepare for next release
add some more coverage for untested subs
todo test for PPI::Normal->register to prevent duplicate registrations
add todo test for mistaken firstidx usage in ::Node
_previous_significant_tokens always returns an array reference
tiny Changes restructure
prevent PPI::Normal->register from registering duplicates
make remove_child actually return undef on failure to find child to remove
fix up incomplete Token::Unknown tests
compress fail explanations in token_unknown.t for comparability
test list constructor as well in token_unknown.t
add a few canary tests in token_unknown.t that will require future fixes
factor out _cast_or_op in Token::Unknown
report line numbers for TODOs in token_unknown.t to help find passing ones
add tests for use + versions from marpa
tests for require + quote literal
fix test name for "method op method" in token_operator.t
_previous_significant_tokens does not return fake tokens anymore
allow Token::Whitespace to throw useful error on unexpected input
parse 1.eq 1 as float + op, not concatenation
update .gitignore and set up MANIFEST.SKIP
add notes about development processes
updates for release 1.222
updates for release 1.224
remove unused dependency Test::Warn
mark quick fix as needing to be verified
updates for release 1.226
updates for release 1.228
Revert "mark quick fix as needing to be verified"
Revert "returning from function if $word isn't defined before using it causes a failure."
prevent possible regex on undefined scalar in __current_token_is_forced_word
updates for release 1.232
updates for release 1.234
make test pragma warning code enable -w to match warnings policy
prevent Node->child from proceeding without a valid argument
updates for release 1.236
Damyan Ivanov (1):
Add "use utf8;" to t/15_charsets.t so perl knows the test text is utf8-encoded
David Steinbrunner (12):
typo fixes
typo fix
typo fix
typo fix
typo fix
typo fix
typo fix
typo fix
typo fix
typo fixes
typo fixes
typo fixes
Elliot Shank (68):
Fix Token::Quote::Double::simplify().
Fix stupid compilation issue in Token::Quote::Double::simplify().
Implement Statement::Include::module_version().
Failing tests for Token::QuoteLike::Words::literal().
Implement Token::QuoteLike::Words::literal().
Fix double variable declaration and t/02_api.t for
Spelling mistake in inline test label in Statement::Include.
Implement line_number(), column_number() and
Make Changes use native line-endings.
Whoops. Got burnt by t/02_api.t again; this time for the
Update Changes to mention PPI::Token::BOM.
Add support for "state" variables.
Change PPI::Document::File to handle "0" as a filename
Clean up a couple double sigils in Token::Quote::Double
Implement Token::Word::method_call().
Handle a bunch more definitely false cases for
Implement descendant_of() and ancestor_of() on PPI::Element.
Add tests for Statement::include::version() based upon
In PPI::Token::Word::method_call(), handle indirect notation
Had a useless condition in Token::Word::method_call().
Document the return value of PPI::Node::remove_child().
Create PPI::Statement::specialized() so that client code doesn't have
Fix PPI test counts. Add Structure::complete() to t/02_api.t.
Fix inheritence description in PPI::Statement::Break. Looked like a
Fix test labels for Statement::specialized().
Create test for Statement::Compound::type() which reveals a parsing
Add two new tests to t/data/05_lexer/05_compound_loops.*:
Add support for the smart match operator.
Initial support for #line directives in PPI::Element::location(). Can
Support for line numbers from #line directives in
Small documentation addition to
Implement PPI::Statement::Variable::symbols() as the guts of
Trailing whitespace in PPI::Statement::Variable.
Add support for UNITCHECK blocks.
Partial support for switch statements, i.e. the "given" part. No
Make PPI::Structure::Given contain Expressions, not Statements.
Finish initial switch statement support-- when statements now get
Whoops. Forgot about default statements for switch statements.
Handle file names from #line directives.
Implement logical_line_number() and logical_filename() to
Fix minor POD typo in PPI::Statement.
Foreach loop variables can be declared via "our" and "local" (PPI::Lexer).
Fix INHERITANCE section in the POD for several PPI::Token::* classes.
Implement PPI::Statement::Include::arguments(). I'm not so sure of
Small documentation change to PPI::Statement::Include.
Fix count in test name in ppi_statement_compound.t.
Document that PPI::Statement::When covers "default" as well.
Fix PPI::Transform::document(). It didn't pick up its $class
Add PPI regression test for a for loop iterating over the results of
That spiritual thing is "souls", not "soles" in PPI docs.
When I updated the PPI::Statement::Compound::type() docs, I didn't
Fix line_number(), column_number(), visual_column_number(),
Update PPI Changes based upon location fixes in PPI::Element.
Fix the documentation fo PPI::Element::location() with respect to the
Fix location methods on PPI::Element in the cases where location()
Change PPI::Statement::Include::version() back to returning a string.
Restore the PPI::Statement::Include functionality to get a version
Double quote the code heredocs for the logical location tests in
Update PPI Changes.
Set svn:eol-style on all of the PPI files.
The rename in PPI from "Switch" to "Given" wasn't complete in cases
Add "default" to the PPI::Statement::When documentation in
Add failing test to PPI::Token::Magic for parsing "%-".
Fix mention of remove() in the PPI::Element::delete() documentation.
Stop PPI::Dumper from causing locations to be reindexed. Locations
Add failing test to PPI::Statement::Package for RT #52259.
Add Tom's regex stuff to PPI/Changes.
Small comment fix in PPI::Statement::Compound.
Gabor Szabo (1):
add failing unit test for Token::Comment
Graham Ollis (1):
do not rely on '.' being in @INC
Guillaume Aubert (4):
Configure Travis-CI and Coveralls.io.
Add tests for PPI::Token::HereDoc.
Add failing tests for heredocs without a trailing newline.
recognize heredoc even if they have no newline at the end
Joel Maslak (1):
Fix bug where Solaris 80 bit long doubles on x86 hardware causes
Julian Fondren (2):
r17642 at tiamat: julian | 2007-03-14 14:57:07 +0900
r17644 at tiamat: julian | 2007-03-15 13:51:13 +0900
Karen Etheridge (3):
fix typo in documentation
eliminate some uses of List::MoreUtils
remove use of Test::NoWarnings; only check for warnings in author mode.
Kent Fredric (1):
t/marpa.t: Fix test reliance on '.' in @INC
Matthew Horsfall (4):
speed up adding an element to an unlabeled statement ( "'a'" x 50000 )
Make subs constant that can be for speed.
Add tests for possible PPI::Statement::Package->version functionality
Add ->version method to PPI::Statement::Package.
Mike (1):
RT 36540 -- support upper case in hex and binary numbers
Mike O'Regan (95):
fix method names in docs
arn't --> aren't
assuption --> assumption
definately --> definitely
existance --> existence
seperate/seperator --> separate/separator
RT 67264: fix spelling of Tom Christiansen
RT 30863: spelling fix
RT 51693: fix pod markup containing '>'
RT 45014: parse '12.34..56.78' parsed as version string + '..' + float
remove inline tests from PPI::Toke::Pod
RT 75039: don't allow '=CUT' to terminate POD
add .t file for PPI::Token::Operator
remove inline tests from PPI::Lexer
RT 68176 and 71705 -- support all augmented assignment operators
Parse x3 as x operator plus a number.
add passing regression tests for (1)x6, which was fixed prior to 1.215
tests and fix for methods/functions like ->x86_convert() parsing as containing the x operator
fix for '1=>x' being parsed as x operator
remove Test::Inline tests and normalize formatting
RT 75308: Andy Stauner's fix for PPI::Node::find_first
add missing metion of 'goto' as a PPI::Statement::Break
Update Changes file for 1.216_01
Issue #40: remove referenes to rt.cpan.org
minimal test for while {} continue {}
Perl::Critic issue 192: hash constructor mistaken for block after ||
Perl::Critic issue 390: hash constructor mistaken for block in ?:
Issue #37: Capture variables above $9 misparsed
issue #31 - DESTROY and AUTOLOAD don't parse as subs without 'sub'
don't reuse variable names in scope
don't rely on error global after eval
remove unneeded initializations
dead code
remove unused variables
use regex with split
trailing comma instead of semicolon
fix punctuation and formatting
use three-arg open(), not two-arg
use strict
change foreach to map
don't warn about variable used to keep document from going out of scope
Issue #55 make prototype() method behave as documented, make
Issue #50 - illusion of use of $1 after failed match
Separate all keywords from immediately-following single quotes.
Debian change for #95
fix spelling of "string"
simplify() docs: fix spelling and be explicity about type change
replace cut and paste with correct descriptions
tabify
typo fixes
Remove references to CVS, SVN, and SourceForge.
Fix repeated word in documentation
Fix module spelling in documentation
Note that PPI::Token::Quote::Double::simplify does
add dev tool artifacts, as in the coverage branch
Noun article fixes in docs and comments
fix use of its vs it's
clone() method is public from Element, don't call it internal
typos
document the PPI::Exception class
note that PPI::Document->new can also throw exceptions
use xdigit character class -- fix #100
add tests for Token::Unknown
fill out t\ppi_statement_package.t
add Statement::Sub tests
tests for Statement::Include and Token::Number::Version
tests for Token::Operator
PPI::Test::pragmas for test files instead of boilerplate
Token::Quotelike::Words test refactoring and coverage
split up the old t::lib::PPI
factor out the pause() function
factor out find_files()
factor out quotable()
Merge t/20_tokenizer_regression.t into t/21_exhaustive.t
Sub attribute parsing unit tests
skip more test cases that do not pass on 5.6.2
add plan on perl < 5.8 to make recent Test::More happy
prepare next set of changes
Remove undocumented, non-working 'timeout' feature.
Explain PPI warnings philosophy.
consolidate, expand PPI::Token::Symbol unit tests
missing %^H; misspelled $\
parse left side of => as bareword even if it looks like a keyword or op
remove source code escapes in the output of QuoteLike::Words->literal
expand $'x to $::main::x in Symbol->canonical as with $::x
extend _cast_or_op
handle "1*{2}"; make bitwise handling more consistent
parse x as the first element of code as a word, not an operator
recognize the implied end of a package statement that includes a block
recognize package names that look like operators as strings, not ops
parse package names that look like v10 as strings, not versions
parse things like v49use as a single bareword, not v-string + keyword
parse x64 as a word, not x operator + number
parse subroutine attributes correctly
Issue #65 Prevent sub names like v10 from being version strings
Olivier Mengué (2):
POD fix: missing title for arguments method
t/14_charsets.t: add a failing test as TODO
Reini Urban (1):
PPI::Dumper fails with strict hashpairs (cperl 5.27)
Salvatore Bonaccorso (1):
New upstream version 1.236
Shmuel Fomberg (2):
Fixed: Quote to QuoteLike
fixed - dor to xor
Steffen Müller (5):
tiny typo in docs
add failing test
doc typo
Allow ENV{X_TOKENIZER} to be set for swapping out the tokenizer during tests
Initial stab at a simple client for the Lacuna Expanse JSON-RPC API
Thomas Sibley (3):
Remove generated README during `make distclean`
Catch "Misplaced _ in number" warnings from 5.6
prevent a warning when Module::Install::ReadmeFromPod isn't loaded
Tom Wyant (21):
Parse the curlys in 'use constant 1.1 {One => 1}' as a PPI::Structure::Constructor, not a PPI::Structure::Block, to be consistent with the parsing of 'use constant {One => 1}'. Added tests for both of these behaviors to t/ppi_statement_include.
Removed rev 8860 changes from t/ppi_statement_include.t, put equivalent code in lib/PPI/Lexer.pm (in a =begin testing section) and t/ppi_lexer.t (as real code).
Once more tweak, this one on whitespace in t/ppi_statement_include.t to get it back the way it was _really_ supposed to be.
Add ability to recognize the 'curly bracket' variables such as ${^MATCH}. These are made into PPI::Token::Magic objects. Tests were added to PPI::Token::Magic to demonstrate the recognition of both 'normal' and 'curly bracket' variables. See the end of perlvar for a description of what was attempted.
Add ability to detect the canonical ways to force curly brackets to be interpreted as either a hash constructor (unary +) or block (semicolon after open curly), as documented in perlref. Tests added to the inline tests for _curly(). I am afraid I succumbed to the desire to collapse the lookahead class determination into a list, the same way some of the lookbehind determination was collapsed into a hash. I will be glad to undo this if desired.
Add back what looks to me like a missing =head2 to document document().
Recognize the second set of curlys in ${block}{...} and @{block}{...} as subscripts, representing a dereference or a hash slice respectively.
Added detection of %- and %+ (in PPI::Token::Unknown), and $#+ and $#- (in PPI::Token::Magic), with test for one of the latter.
PPI mis-parses "pack'H*'" as Token::Word "pack'H" etc. This proposed patch also includes 'unpack', 'gt', 'ge', 'lt', and 'le', plus tests. The implementation has been made hash-based, but I fear that fully fixing the problem will require the name of every built-in in the hash, and then some.
Tweak wording of documentation
Have PPI::Token::_QuoteEngine::Full correctly parse monstrosities of
Add private methods to PPI::Token::_QuoteEngine::Full to access stuff
Bypass the 'Calculate locations for the rest' loop in
When a subroutine attribute spans more than one line, the
RT #61305: return { foo => 1 } should parse curlys as hash constructor,
43933: Recognize <*> as being a glob operation, not a readline.
RT #64247: bless {} probably contains a hash constructor, not a block
Back out commit 13236, as documented in RT ticket 43933. The original
RT #65199 - Cast can trump braces in PPI::Token::Symbol->symbol()
Actually found in the wild: someone created a method named 'sub', and
RT #67517: Form feed appears to be white space
bowtie (1):
modernizations for Makefile.PL
brian d foy (1):
* The example to find named subroutines in PPI::Document is a bit
-----------------------------------------------------------------------
No new revisions were added by this update.
--
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/pkg-perl/packages/libppi-perl.git
More information about the Pkg-perl-cvs-commits
mailing list