1) fix the PLIST to correspond with the files added+removed
2) fix the interpreter in some installed files
3) ignore work-directory references in 12 installed files. Yes, this
is wrong, and has been reported to parrot, issue #201.
PKGREVISION not bumped, since this would not create a package earlier.
Parrot 3.8.0 News:
- Core
+ New tools/release/auto_release.pl script automates most of
release
- Languages
+ Winxed
- Updated snapshot to version 1.2.0
- allowtailcall modifier in try
--debug command-line option, __DEBUG__ predefined constant
and __ASSERT__ builtin
- namespace, class, and ~ (bitwise not) operators
- Implicit nested namespace in namespace and class
declarations
- -X command-line arg
- Documentation
+ Improved release manager guide
- Tests
+ New Makefile target "resubmit_smolder" to resubmit test
results
+ New Makefile target "all_hll_test" runs the test suite of all
HLLs and libraries known to work on Parrot
+ New Makefile target "interop_tests" run language
interoperability tests, which runs as part of the normal "make
test" as well
The previous commit caused DragonFly to build the libraries with a
different file name than specified in the PLIST, causing the build
to fail on DragonFly.
This commit forces DragonFly to use library names without dots to
match the PLIST.
=== 3.11 / 2011/10-17
* Bug fixes
* Avoid parsing TAGS files included in gems. Issue #81 by Santiago Pastorino.
=== 3.10 / 2011-10-08
* Major enhancements
* RDoc HTML output has been improved:
* The search from Vladimir Kolesnikov Sdoc has been integrated.
The search index generation is a reusable component through
RDoc::Generator::JsonIndex
* The table of contents is now a separate page and now shows links to
headings and sections inside a page or class.
* Class pages no longer show the namespace and no longer have file info
pages.
* HTML output is HTML 5.
* Static files can be copied into RDoc using --copy-files
* RDoc supports additional documentation formats:
* TomDoc 1.0.0-rc1
* RD format
The default markup can be set via the <tt>--markup</tt> option.
The format of documentation in a particular file can be specified by the
+:markup:+ directive. If the +:markup:+ directive is in the first comment
it is used as the default for the entire file. For other comments it
overrides the default markup format.
The markup format can be set for rake tasks using RDoc::Task#markup
* RDoc can save and load an options file.
To create an options file that defaults to using TomDoc markup run:
rdoc --markup tomdoc --write-options
This will create a .rdoc_options file. Check it in to your VCS and
package it with your gem. RDoc will automatically load this file and
combine it with the user's options.
Some options are not saved. See RDoc::Options@Saved+Options for full
details.
* Minor enhancements
* RDoc autoloads everything. You only need to require 'rdoc' now.
* HTML headings now have ids matching their titles.
= Hello!
Is rendered as
<h1 id="label-Hello%21">Hello!</h1>
* Labels for classes or methods can be linked-to by adding an <tt>@</tt>
following the class or method reference. For example,
<tt>RDoc::Markup@Links</tt>
See RDoc::Markup@Links for further details.
* For HTML output RDoc uses +SomeClass.method_name+ and
+SomeClass#method_name+ for remote methods and attributes and
+::method_name+ and +#method_name+ for local methods.
* RDoc makes an effort to syntax-highlight ruby code in verbatim sections.
See RDoc::Markup@Paragraphs+and+Verbatim
* Added RDoc::TopLevel#text? and RDoc::Parser::Text to indicate a
parsed file contains no ruby constructs.
* Added <tt>rdoc-label</tt> link scheme which allows bidirectional links.
See RDoc::Markup for details.
* Added RDoc::Comment which encapsulates comment-handling functionality.
* Added RDoc::Markup::PreProcess::post_process to allow arbitrary comment
munging.
* RDoc::RDoc::current is set for the entire RDoc run.
* Split rdoc/markup/inline into individual files for its component classes.
* Moved token stream HTML markup out of RDoc::AnyMethod#markup_code into
RDoc::TokenStream::to_html
* "Top" link in section headers is no longer inside the heading element.
* RDoc avoids printing some warnings unless run with `rdoc --verbose`. For
Rails issue #1646.
* Finishing a paragraph with two or more spaces will result in a line break.
This feature is experimental and may be modified or removed.
* Bug fixes
* Performance of RDoc::RubyLex has been improved. Ruby Bug #5202 by Ryan
Melton.
* Clicking a link in the method description now works. Issue #61 by Alan
Hogan.
* Fixed RDoc::Markup::Parser for CRLF line endings. Issue #67 by Marvin
Gülker.
* Fixed lexing of percent strings like %r{#}. Issue #68 by eclectic923.
* The C parser now understands classes defined with
+rb_struct_define_without_accessor+ (like Range). Pull Request #73 by Dan
Bernier
* Fixed lexing of <code>a b <<-HEREDOC</code>. Issue #75 by John Mair.
* Added LEGAL.rdoc with references to licenses in other files. Issue #78 by
Dmitry Jemerov.
* Block parameters are displayed in Darkfish output again. Issue #76 by
Andrea Singh.
* The method parameter coverage report no longer includes parameter default
values. Issue #77 by Jake Goulding.
* The module for an include is not looked up until parsed all the files are
parsed. Unless your project includes nonexistent modules this avoids
worst-case behavior (<tt>O(n!)</tt>) of RDoc::Include#module.
changes in sbcl-1.0.53 relative to sbcl-1.0.52:
* enhancement: on 64-bit targets, in src/compiler/generic/early-vm.lisp,
the parameter n-fixnum-tag-bits may now vary from 1 (fixnum =
(signed-byte 63)) to 3 (fixnum = (signed-byte 61)) at build-time.
* enhancement: SB-EXT:WAIT-FOR allows waiting for arbitrary events.
* minor(?) incompatible(?) change: The default fixnum width on 64-bit
targets is now 63 bits (up from 61).
* enhancement: DESCRIBE now reports a lambda-list and source location
for complext setf-expanders.
* bug fix: PUSH, PUSHNEW, POP, REMF, INCF, DECF, DEFINE-MODIFY-MACRO,
GETF, LOGBITP, LDB, and MASK-FIELD now arrange for non-primary values
of multiple-valued places to be set to NIL, instead of signalling an
error (per a careful reading of CLHS 5.1.2.3).
* bug fix: floating-point traps now work on darwin/x86 and /x86-64.
* bug fix: repair crash in x86oid darwin signal handling emulation
when built with certain compilers.
* bug fix: SB-ROTATE-BYTE misrotated to the right when using constant
rotation arguments on x86-64. (lp#882151)
* bug fix: low-level control flow optimisations could result in bogus
code in functions with tail and non-tail calls to local functions on
x86oids. (lp#883500)
* bug fix: on SPARC/:sb-unicode, avoid crashing the assembler by trying
to emit literal characters > (code-char 4095), for comparisons with
constant characters.
* bug fix: ROOM reported only the low 32 bits of dynamic space usage
on 64 bit platforms. (lp#881445)
* bug fix: DELETE-FILE did not MERGE-PATHNAMES, making it possible to
delete the wrong file when using relative pathnames. (lp#882877)
* bug fix: optimized SEARCH of vectors-on-vectors mishandled zero-length
sequences and :KEY NIL.
changes in sbcl-1.0.52 relative to sbcl-1.0.51:
* enhancement: ASDF has been updated to version 2.017.
* enhancement: the --core command line option now accepts binaries with
an embedded core.
* enhancement: when built with :sb-core-compression, core files (regular
or executable) can be compressed with zlib. Use the :COMPRESSION
argument to SAVE-LISP-AND-DIE to specify a compression level.
* enhancement: --[no-]merge-core-pages determines whether the runtime
hints the operating system that identical core pages between SBCL
processes should share the same physical memory. Default is to only
enable this for compressed cores.
* optimization: SLEEP no longer conses.
* optimization: *PRINT-PRETTY* no longer slows down printing of strings
or bit-vectors when using the standard pretty-print dispatch table.
* bug fix: non-function FTYPE declarations no longer cause a compiler-error.
(lp#738464)
* bug fix: compiler-errors causes by MEMBER types in conjunction with with
AREF, CHAR, etc. (lp#826971)
* bug fix: compiler-errors causes by integer arguments with composed of
multiple ranges to ARRAY-IN-BOUNDS-P. (lp#826970)
* bug fix: ,@ and ,. now signal a read-time error for certain non-list
expressions. (lp#770184)
* bug fix: complex single float literals are correctly aligned when used
as arguments of arithmetic operators.
* bug fix: on 32-bit platforms, rounding of double floats larger than a
fixnum is correct. (reported by Peter Keller)
* bug fix: stray FD-HANDLERs are no longer left lying around after unwinds
from RUN-PROGRAM. (lp#840190, reported by Dominic Pearson; fix from Max
Mikhanosha)
* bug fix: redefining classes such that slots with custom allocation are
added or removed works again.
Please refer detail for:
http://svn.ruby-lang.org/repos/ruby/tags/v1_9_3_0/NEWShttp://svn.ruby-lang.org/repos/ruby/tags/v1_9_3_0/ChangeLog
Short summary from NEWS:
* Ruby's License is changed from a dual license with GPLv2
to a dual license with 2-clause BSDL.
* Encoding
* new encodings: CP950, CP951, UTF-16 and UTF-32
* change alias:
* SJIS is Windows-31J
* Regexps now support Unicode 6.0. (new characters and scripts)
* builtin classes
* ARGF
* new methods: ARGF.read_nonblock and so on.
* Array
* extended method: Array#pack supports endian modifiers
* String
* extended method: String#unpack supports endian modifiers
* new method: String#prepend and String#byteslice
* Bignum
* Multiplication algorithm for Bignums with a large number of digits over
150 BDIGITs is changed in order to reduce its calculation time.
Now such large Bignums are multiplied by using Toom-3 algorithm.
* File
* new constant: File::NULL and File::DIRECT
* IO
* extended method: IO#putc supports multibyte characters
* new methods: * IO#advise, IO.write and IO.binwrite
* Kernel
* move #__id__ to BasicObject
* extended method: Kernel#rand supports range argument
* Module
* new methods: Module#private_constant and Module#public_constant
* Random
* extended method: Random.rand supports range argument
* Time
* extended method: Time#strftime supports %:z and %::z
* Process
* Process#maxgroups and Process#maxgroups= now raise NotImplementedError if
the platform don't support supplementary groups concept
* Correct DESCR; this is 1.9.2 release minimum base package.
* Don't remove whole gem directory but keep its own gem directory only.
* Also make MESSAGE explicitly 1.9.2.
* rubygem: Avoid to use Gem::RequirePathsBuilder now.
* Bump PKGREVISION.
* Use 18, 19 instead of 1.9, 2.0 for RUBY_VERSION_DEFAULT.
* Add 193 for Ruby 1.9.3, too.
* If RUBY_VERSION_SUPPORTED contains single version of Ruby, make package
force depends to the version.
* Move RUBY_SITE_SUBDIR to Makefile.common.
* Change RUBY_VERSION_SUFFIX to RUBY_VERSION_FULL.
* Remove small code for NetBSD 1.x.
* Change RUBY_DLEXT and RUBY_SLEXT by ${_OPSYS_SHLIB_TYPE} instead of
${OPSYS}'s value.
GCC 4.6.2 is a bug-fix release containing fixes for regressions and
serious bugs in GCC 4.6.1, with over 110 bugs fixed since previous
release.
This is the list of problem reports (PRs) from GCC's bug tracking system
that are known to be fixed in the 4.6.2 release. This list might not be
complete (that is, it is possible that some PRs that have been fixed are
not listed here).
http://gcc.gnu.org/bugzilla/buglist.cgi?bug_status=RESOLVED&resolution=FIXED&target_milestone=4.6.2
I didn't check the others.
We should make consider just removing all these ONLY_FOR_PLATFORM
restrictions and just make the description or a +DISPLAY message
clear on what is supported.
cVS: ----------------------------------------------------------------------
Simplify building with the Android NDK.
Allow Android'd support_boehm=no to work.
Disable the -Wunused-by-set-variable warning supported by newer gcc versions.
Add shared_perfcounters option to enable minimal.
Fix configure check for darwin to support all variants
Fix mingw32 cross-build on a git no-branch checkout.
Add a membar to libgc's UNLOCK () on arm.
Pass HAVE_ARMV6 to libgc on darwin too.
"platform" in Python terms is different for Linux kernel 2.* Vs Linux
kernel 3.*. Add in support to pull in a different PLIST for Linux 3.*.
Fixes build under Ubuntu 11.10.
XXX Perhaps it would be cleaner to name the PLIST to match the python platform
name - since we already calculate that anyway, and that is exactly what drives
the contents of these PLISTs.
