Changes from previous:
2.56 Thu Oct 20 2011
- Updated for v5.15.4
2.56 Tues Sept 20 2011
- Updated for v5.15.3
- Corrected version of Module::CoreList shipped
with v5.10.0
- Corrected Errno being shipped with v5.6.0 and v5.6.1
- Corrected VMS::Filespec being shipped with v5.12.3
- Corrected DB_File being shipped with v5.8.2
- Added http://perlpunks.de/corelist to SEE ALSO
2.55 Sat Aug 20 2011
- Updated for v5.15.2
2.54 Wed Jul 20 2011
- Updated for v5.15.1
2.53 Thu Jul 14 2011
- Added release date option '-r' to corelist
2.52 Mon Jul 4 2011
- [rt.cpan.org #69108] Extraneous modules listed
reported by Dee Newcum
2.51
- Updated for v5.15.0
2.50
- Updated for v5.12.4
2.49
- Updated for v5.14.0
2.46 Sun Mar 20 2011
- Resolved RT #59511
- Updated for v5.13.11
2.45 Sun Feb 20 2011
- Bumped the version number and data for the 5.13.10 release
2.44 Sat Jan 22 2011
- Merged v5.12.3 and v5.13.9
2.43 Sat Jan 22 2011
- Updated for v5.12.3
2.42_01 Thu Jan 20 2011
- Updated for v5.13.9
2.42 Mon Dec 20 2010
- Updated for v5.13.8
2.41 Sat Nov 20 2010
- Updated for v5.13.7
2.40 Wed Oct 20 2010
- Updated for v5.13.6
Changes from previous:
0.006007 - 12 Sep 2011
- Depend on B::Hooks::OP::Check version 0.19, which fixes a serious bug in
how it interacts with other modules that hook ops.
- Initialize immediately upon loading the module, so that "was Devel::Declare
loaded soon enough" errors in string eval can be fixed by loading the
module earlier without having to also actually use the module earlier.
- Adjust toke_scan_str logic to always show a positive effective length of
string source.
- Return undef from toke_scan_str if string was unterminated.
- Detect and croak if unwanted reallocation occurs during toke_scan_str.
- Avoid memory leak in toke_scan_str.
- Give Devel::Declare::Context::Simple a version number.
- Add MYMETA.{json,yml} to MANIFEST.SKIP and .gitignore.
0.006006 - 23 Aug 2011
- Increase default linestr size to avoid reallocations (Zefram).
Changes from previous:
0.15 - 2011.04.17 - SAPER #PerlQA2011
- [DIST] CPAN-RT#54456: Set INSTALLDIRS to "site" when installed on
Perl 5.11+ (thanks to Todd Rinaldo).
- [DOC] Document a known bug under Perl 5.8.4 and 5.8.5.
- [TESTS] Fixed tests to pass under Perl 5.8.4 and 5.8.5.
0.14 - 2011.04.16 - SAPER #PerlQA2011
- [CODE] Updated from bleadperl:
- XSLoader::load() with no arguments can use caller to find
a default package (Nicholas Clark).
- Avoid defining a full XSLoader::bootstrap_inherit post 5.6,
as it's not needed (Nicholas Clark).
- Small optimisation: for the generated XSLoader.pm, avoid a
runtime lexical which is constant (Nicholas Clark).
- [TESTS] Updated from bleadperl, solving RT-CPAN #54132, #61332.
- [TESTS] Fixed tests for old Perls.
- [TESTS] Added t/00-load.t and t/01-api.t, to provide basic tests
when the main ones are skipped.
Change log:
This is the changelog file for the POCO C++ Libraries.
Release 1.4.2p1 (2011-09-24)
============================
- On Linux, the RTLD_DEEPBIND option is no longer passed to dlopen().
This change was introduced in 1.4.2 to solve a specific problem one customer
was having. Unfortunately, it leads to problems with RTTI.
- It's now possible to pass flags (SHLIB_GLOBAL, SHLIB_LOCAL) to
Poco::SharedLibrary::load() (and the constructor implicitly calling load()),
controlling the mode flags (RTLD_GLOBAL, RTLD_LOCAL) passed to dlopen().
On platforms not using dlopen(), these flags are ignored.
- fixed SF# 3400267: Path_WIN32.cpp bug
Release 1.4.2 (2011-08-28)
==========================
- added Poco::DateTimeFormat::ISO8601_FRAC_FORMAT
- added new Poco::DateTimeFormatter and Poco::DateTimeParser format specifier:
%s for seconds with optional fractions of a second
- fixed a problem with ioctl() on BSD platforms (including OS X) where the
second argument to ioctl() is unsigned long instead of int, causing bad
things on a OS X 64-bit kernel.
- fixed a potential endless loop when enumerating IPv6 network addresses
(reported by Laurent Carcagno)
- new compile-time config option on Windows to set thread names in
debugger. Enable with -DPOCO_WIN32_DEBUGGER_THREAD_NAMES. Available
only in debug builds.
- Cipher can now create Base64 and HexBinary encoded output without linefeeds
(suitable for use in cookies, etc.)
- added Poco::Path::popFrontDirectory()
- improved VxWorks support
- IPv6 fixes: added proper scope id handling in IPAddress, SocketAddress
and related classes.
- Added Poco::Net::ServerSocket::bind6() which allows control over the
IPPROTO_IPV6/IPV6_V6ONLY socket option.
- Removed Poco::MD2Engine class due to licensing issues (the
license for the MD2 code from RSA only allows non-commercial
use). Note that the MD4 and MD5 code from RSA does not have
this issue.
- fixed a Net HTTP client testsuite issue where some tests might
have failed due to prematurely aborted connections by
the HTTPTestServer.
- Poco::Net::SocketAddress: when there is more than one address
returned by a DNS lookup for a name, IPv4 addresses will be
preferred to IPv6 ones.
- NetworkInterface::list() now also returns IPv4 interfaces on Windows when
built with -DPOCO_HAVE_IPv6
- XMLWriter: fixed a bug with attribute namespaces (no namespace prefix
written if attribute namespace is the same as element namespace)
- fixed SF# 3378588: Mismatched new[]/delete (in RSAEncryptImpl and
RSADecryptImpl)
- fixed SF# 3212954 (OpenSSLInitializer::uninitialize() crash) and
SF# 3196862 (Static OpenSSLInitializer instance causes Windows
deadlocks) by removing the static Poco::Crypto::OpenSSLInitializer
instance. Automatic OpenSSL initialization is now done through
Poco::Crypto::Cipher, Poco::Crypto::CipherKey,
Poco::Crypto::X509Certificate, Poco::Net::Context classes; however,
it is still recommended to call Poco::Crypto::initializeCrypto()
and Poco::Crypto::uninitializeCrypto() early at application
startup, and late at shutdown respectively (or
Poco::Net::initializeSSL()/Poco::Net::uninitializeSSL() if the
NetSSL library is used) to avoid multiple full OpenSSL init/uninit
cycles during application runtime.
- Poco::Logger now also support a symbolic log level "none"
(for use with setLevel()) that disables logging completely
for that Logger (equivalent to setLevel(0)).
- Added experimental Android support, using the existing gmake-based
build system.
- fixed SF# 3288584: DateTimeFormatter link error
- fixed SF# 3187117: Typo in InflatingInputStream doc
- fixed SF# 3309731: _WIN32_WCE comparison should be with 0x600 not 600
- fixed SF# 3393026: RegularExpression.h identical enum value
- fixed SF# 3274222: AtomicCounter's postfix operators aren't atomic on Windows
- fixed SF# 3317177: Handle leak on windows
- fixed SF# 3181882: Poco::URI::getPathEtc() double-encodes query
- fixed SF# 3379935: ThreadPool Start Bug
- fixed SF# 3354451: Poco::Format::parsePrec never sets the precision to zero
- fixed SF# 3387258: _MAX_PATH used but unknown in Path_WIN32
- fixed a problem in RSAKeyImpl where direct access to the RSA in a EVP_PKEY
would no longer work in recent OpenSSL versions. Using EVP_PKEY_get1_RSA()
fixes the issue.
- added Poco::Crypto::EncryptingInputStream,
Poco::Crypto::EncryptingOutputStream, Poco::Crypto::DecryptingInputStream
and Poco::Crypto::DecryptingOutputStream.
- fixed SF# 3148126: HTTPSClientSession destructor (!) throws an IOException
- fixed SF# 3178098: Add constructor to Poco::TemporaryFile to specify directory
- fixed SF# 3175310: Absolute path when device
- fixed SF# 3301207: Guided tour example contradicts apidoc (API doc was wrong)
- Poco::Net::HTTPMessage::setContentLength() and
Poco::Net::HTTPMessage::getContentLength() now use std::streamsize
instead of int. This enables 64-bit Content-Length support at
least on 64-bit platforms.
- fixed SF# 3177530: TemporaryFile::tempName() + glob bug on xp
- fixed SF# 3177372: FileChannel documentation inconsistency
- added %E format specifier to Poco::PattermFormatter (epoch time in seconds
since midnight, January 1 1970)
- On Windows, Poco::Util::ServerApplication now supports a /description
command line argument for specifying a service description
(together with /registerService) - added
Poco::Util::WinService::setDescription() and
Poco::Util::WinService::getDescription()
- fixed SF# 3155477: Incorrect URI path handling
- fixed SF# 3309736: Extended Exception macros to set default exception code
new macro is named POCO_DECLARE_EXCEPTION_CODE
- added getter functions for modulus and exponents to Poco::Crypto::RSAKey.
- added Poco::Net::SocketAddress::operator == () and
Poco::Net::SocketAddress::operator != ()
- fixed SF# 3182746: IPAddress.cpp IPv6 bug on big-endian
- fixed SF# 3196961: Unix daemon fails to loadConfiguration() if
started from cwd
- fixed SF# 3393700: NotificationCenter may call a removed observer and crash.
- Reworked implementation of the events framework (Poco::BasicEvent
and friends). The framework is now completely multithreading
save (even in the case that an event subscriber object unsubscribes
and is deleted while an event is being dispatched). Also, the
restriction that any object can only register one delegate for
each event has been removed. For most cases, dispatching events
should be faster, as dispatching an event now needs less dynamic
memory allocations.
- fixed SF# 3178109: getNodeByPath() changes:
getNodeByPath() and getNodeByPathNS() have been moved to
Poco::XML::Node. Furthermore, when invoked on a Poco::XML::Document,
the behavior has changed so that the document element is now
included when traversing the path (previously, traversal would
start at the document element, now it starts at the document).
The path expression can now start with a double-slash, which
results in a recursive search for the path's first element in
the DOM tree.
- fixed SF# 3382935: String data being truncated using ODBC, and
SF# 2921813: Wrong implementation of the ODBC string binding
Release 1.4.1p1 (2011-02-08)
============================
- Poco::Mutex is now a recursive mutex again on Linux
(this was caused by an unfortunate feature test for
PTHREAD_MUTEX_RECURSIVE which did not work on Linux
as PTHREAD_MUTEX_RECURSIVE is an enum value and not
a macro)
- Poco::Net::SecureSocketImpl::abort() now only shuts
down the underlying socket connection and does not free
the SSL object, due to multithreading issues.
Release 1.4.1 (2011-01-29)
==========================
- fixed SF# 3150223: Poco::BinaryReader cannot read std::vector correctly
- fixed SF# 3146326: SharedMemory issue
- made Poco::Net::HTTPSession::abort() virtual
- added Poco::Net::SecureStreamSocket::abort() to immediately close
a SSL/TLS connection without performing an orderly SSL/TLS shutdown.
- fixed SF# 3148126: HTTPSClientSession destructor (!) throws an IOException.
Added try/catch block to Poco::Net::SecureSocketImpl destructor.
- added additional constructor to Poco::Net::HTTPSClientSession, taking
both a socket and a session object.
- Poco::Net::HTTPSession::abort() now also can be used with a
Poco::Net::HTTPSClientSession.
- fixed SF# 3148045: make clean and distclean issues
- changed Data library names on Unix/Linux platforms to
match the names on Windows (PocoSQLite -> PocoDataSQLite,
PocoMySQL -> PocoDataMySQL, PocoODBC -> PocoDataODBC)
- added additional options to configure script
- added additional documentation to Poco::Net::HTTPClientSession
- Poco::Net::HTTPClientSession::receiveResponse() closes the connection
if an exception is thrown while reading the response header.
This ensures that a new connection will be set up for the next request
if persistent connections are used.
- improved Poco::Net::MultipartDecoder performance by reading directly from
streambuf
- improved performance of Poco::Base64Encoder, Poco::Base64Decoder,
Poco::HexBinaryEncoder and Poco::HexBinaryDecoder by working directly with the
given stream's streambuf.
- improved performance of MessageHeader::read() by reading directly from
streambuf instead of istream.
- it is now possible to specify additional MIME part header fields
for a MIME part through the Poco::Net::PartSource class.
- upgraded SQLite to release 3.7.4
- added experimental VxWorks support for VxWorks 5.5.1/Tornado 2.2 and
newer. Please see the VxWorks Platform Notes in the reference documentation
for more information. Currently, the VxWorks is untested; full support
will be available in release 1.4.2.
- fixed SF# 3165918: Poco::DynamicAny fails to convert from string to float
- fixed SF# 3165910: Poco::Net::MessageHeader does not accept HTTP conforming
header
- made Poco::Task::cancel() virtual so that tasks can implement custom
cancellation behavior.
- added optional argument to Poco::Util::WinRegistryKey constructor
to specify additional flags (in addition to KEY_READ and KEY_WRITE)
for the samDesired argument of RegOpenKeyEx() or RegCreateKeyEx().
- improved Poco::BasicEvent::notify() performance by avoiding an
unnecessary heap allocation.
- added additional well-known port numbers to Poco::URI: rtsp, sip, sips, xmpp.
- added Poco::Net::MediaType::matchesRange()
- improved invalid socket handling: a Poco::Net::InvalidSocketException is
now thrown instead of an assertion when an operation is attempted
on a closed or otherwise uninitialized socket.
Changes from previous:
Changes for 0.72 Wed Jun 22 12:29:59 BST 2011
=================================================
* Added IPC::Open3 support for capturing STDOUT/STDERR
on MSWin32, prefer this over IPC::Run
Changes for 0.70 Mon Jan 31 20:34:09 GMT 2011
=================================================
* Apply a patch from Petya Kohts, RT #65276, with
changes for run_forked:
1) fix for the typo in the name of the signal
2) changed default for clean_up_children (which
seems to be the behavior expected by the majority of the users)
3) added detection (and forwarding to the caller) of the case
when run program is killed by signal
4) kill_gently is now used in cases when parent died
and when the executed program times out
5) added options which allow to execute some user code
in the beginning and at the end of the child
Changes for 0.68 Fri Jan 7 21:08:58 GMT 2011
=================================================
* Apply blead patch from Peter Acklam
Changes for 0.66 Tue Nov 23 12:10:24 GMT 2010
=================================================
* Apply documentation patch from Dan Dascalescu [RT # 63250]
* Apply another documentation patch from Dan Dascalescu [RT #63251]
* Fix an issue with _split_like_shell_win32() raised by tunakermit [RT #62961]
Changes for 0.64 Tue Oct 19 19:01:05 BST 2010
=================================================
* Only enable the can_run instances behaviour when
global variable $INSTANCES is set
Changes for 0.62 Tue Oct 19 15:51:30 BST 2010
=================================================
* Amended can_run behaviour to run all found instances
of the given binary in a list context
devel/p5-PEO-Component-Syndicator because recent versions of
chat/p5-POE-Component-IRC require this module.
POE::Component::Syndicator is a base class for POE components which
need to handle a persistent resource (e.g. a connection to an IRC
server) for one or more sessions in an extendable way.
Kitware is pleased to announce the release of CMake 2.8.6. This release features several notable changes, including a new AUTOMOC target property that helps make it easier to build Qt projects and support for the Intel Fortran compiler in VisualStudio10. Additionally, VisualStudio 9 and 10 generators for Itanium platform have been added, and appear as the “Visual Studio 9 2008 IA64” and “Visual Studio 10 IA64” generators on Windows builds of CMake.
This release also contains an experimental generator in the Windows build targeting VisualStudio 11. It will remain "Experimental" until VisualStudio 11 itself is finalized. If you have the developer preview of VisualStudio 11, we would appreciate any feedback you have. Please keep in mind that this is very new, and has only been tested enough to be useful to include for the community to test drive and provide feedback on.
Other improvements include the movement of many files from KDE's overrides of CMake module files, which have been accepted into upstream CMake thanks to the hard work of Alex Neundorf and Stephen Kelly. There have been many Xcode generator improvements, including support for iOS projects and the ability to switch between simulator and device builds.
Originally this was an attempt to upgrade version 5.0.0 to version
5.1-RELEASE or even 5.2-DEVELOPMENT, but it turns out that those
versions require a GNAT Ada compiler based on gcc 4.7, which hasn't
had its first release yet. This is mainly due to an change in the
project management API, but using the 4.7 source files fail to
compile due to the new SPARK restrictions. Therefore GPS must
remain at 5.0.x until such time as GNAT-AUX is based on gcc 4.7.
This is a bug fix release.
The list of bug fixes is unknown, but it's confirmed the bug on the
project dialog, library tab has been fixed and thus those patches
are removed.
The Makefile was updated to allow GPS users to take advantage of
the numerous Python scripts, the Python console, and the python-GTK
bindings. The option is present, but it has been removed from the
option list because the pkgsrc version of Python cause GPS to core
dump due to missing symbols in their dynamically-loaded libraries.
Version 2.6 and version 2.7 were both tested, and fail in different
ways. For comparison, the FreeBSD version of GPS builds and operates
fine with Python 2.7, although at times similar "undefined symbol"
messages appear it that error log. Once the issues with Python are
fixed, this new "python" option in options.mk will be re-enabled.
(accept any of python-2.x).
* 0.2.5
** Language:
- Make a very small attempt to parse C++ code, amounting to accepting
identifiers containing ::, tilde, and template invocations. Use the
option -c++. This is not likely to be very useful in practice.
- Added metavariable metavariable type.
- Add disjunctions on identifiers in some contexts
- Pretend that & indicates a pointer in -c++ mode
- Support for new and delete
- Allow arrays in smpl parameter declarations. Thanks to Jesper Louis
Andersen for pointing out the problem.
- Field list metavariables
- Add the ability to add __attributes__ (NB, no matching on attributes)
- Slightly improved the error message for transformation on multiple
variable declaration. Thanks to Jonathan Nieder for pointing out the
problem.
** Features:
- support transformations on variables (only) in declarations that declare
multiple variables
- allow #endif XXX in C code
- relax_include_path now applied to non local includes too, in which case
it tries to find a unique file with a suffix of the provided name.
this is useful for directories that are intended to be symbolic links.
- support matching and removing #undef
- support for iteration in ocaml, requires use of -no_show_diff
- calls to likely and unlikely propagate test expression status to their
arguments
- reuse typedefs and macros from complete parsing when reparsing the
transformed code
- better explanation when presenting glimpse tokens
- optimization for an if branch that is just { ... }
- spatch -control_flow_to_file file.c generates a file file.dot
- include files with names ending in .cpp if -c++ option is given
- removed use of the no longer supported sexplib function Conv.hashtbl_of_sexp
- add information about which Fedora packages are needed
** Bugfix:
- improved parsing of expressions to allow ... to the right of operators in
more places
- Fix check_config for Python 2.7 on Fedora 14 (Reported-by: Michael Stefaniuc)
- Check for ocamlfind in configure (Reported-by: Paul E. McKenney)
- Postpone use of ocamlfind at runtime to report fewer errors
- Add support for Python 2.4 binding with the provided pycaml library
- Allow '@' in script code. Thanks to Laurent Reveillere for noticing the
problem.
- Remove an optimization of x or true to true in asttoctl2 for the case
where x may make a modification, eg the case of -thing ?-thing
- Allow adding comments before a function.
- Introduction of newlines in function arguments needs to take into account
that the code added by SmPL can include newlines. Thanks to Thomas
Gleixner for finding the problem.
* 0.2.4
** Language:
- scripts now have names and can declare metavariables that are seen by
SmPL as identifier metavariables, see demos/{python,ocaml}tococci.cocci
- declaration (match variable declarations) and field (match structure
field declarations) metavariables
- matching of array initializations and of enum declarations
** Features:
- compatability with Python 2.7. Thanks to Richard W.M. Jones for pointing
out the problem
- python and ocaml string representations of parameter lists and expression
lists
- try simpler patterns on glimpse failure
- graceful and immediate failure when a virtual rule is defined on the
command line, but not in the rule
- spatch returns -1 when a virtual rule is requested that is not supported
- no transformation when using * with the option -no_show_diff
- expanded +++ file name for match output
- struct *^* and enum *^* metavariable types
- allow an assignment to match a variable initialization, even if multiple
variables are declared at the same time.
- add the option -recursive_includes
- added the option -use_idutils. The requires the previous use of the
script idutils_index.sh
** Bugfix:
- detect used after metavariables in simple statement metavariable match
- allow inheriting position variables over rules that make transformations
but depend on virtual rules that are not defined on the command line
- treat top-level tokens in decl visitor. Thanks to Peter Tummeltshammer
for noticing a resulting problem.
- improve typing of the result of an assignment expression
- eliminate some recursive calls in C code parsing to avoid stack overflows
- better pretty printing of #define when the whole line is removed
- better pretty printing of asm code containing
"::[input]"r"(&coherence_data[i])"
- allow pretty printing of C code with nameless bit fields
- avoid adding an error message to standard output if ocamlfind is not
found
- struct/union/enum-typed metavariables with a metavariable type name
- better management of whitespace between deleted lines. Thanks to
Andriy Gapon for noticing the problem.
- improved parsing recovery for badly parsed struct/enum typedefs
- drop expanded tokens in sgrep output. Thanks to Andriy Gapon for
noticing the problem.
- constant strings have array type, not pointer type. Thanks to Vasiliy
Kulikov for noticing the problem.
- improve indentation when adding code after a function call with indented
arguments.
- return type of sizeof converted to unsigned long. Thanks to Vasiliy
Kulikov for noticing the problem.
- improve spacing when adding structure initializers. Thanks to Vasiliy
Kulikov for noticing the problem.
Changes from previous:
==============================
2011-06-29 10:03:36 +0100 1.16
==============================
Resolve [rt.cpan.org #69175] "Test-POE-Server-TCP-1.14 stuck in test
on Strawberry Perl 5.12.3"
Made the explicit socket shutdown for 'cygwin' applicable to MSWin32
as well. I was able to reproduce the issue with:
This is perl 5, version 14, subversion 1 (v5.14.1) built for
MSWin32-x64-multi-thread
Version 1.6d of cxref released : Mon Oct 3 14:00:00 2011
--------------------------------------------------------
Bug fixes
Updated for latest version of autoconf.
Allow structure initialisers to have multiple components (e.g. a.b=1).
Remove gcc warning messages.
Change Makefile for better comptibility with FreeBSD.
Version 1.6c of cxref released : Mon May 31 10:00:00 2010
---------------------------------------------------------
Bug fixes
Handle the gcc __builtin_offsetof() and offsetof() functions.
Check that the lex/yacc programs actually exist at configure time.
Handle ASM statements with named identifiers in them.
Parsing changes
Removed the char_varying type.
Document changes
Update web page links
The build system of this package is completely fubar'd (mostly because this
has been designed to be built against the Boost source files and to never
be installed). I have made an attempt to make the build system and the
runtime finding of configuration systems slightly more sane, but can't
promise it will work everywhere yet. So far only tested in OS X 10.6.
Description:
Vera++ is a programmable tool for verification, analysis and
transformation of C++ source code.
Vera++ is a tool designed for programmers -- not only in the sense that
it helps with everyday development work, but also in the sense that all
its rules are implemented as scripts that can be modified or added by
the user. Thanks to this, programmers have complete freedom and control
over what the tool does and how it works.
Vera++ is mainly an engine that parses C++ source files and presents the
result of this parsing to scripts in the form of various collections --
the scripts are actually performing the requested tasks.
Changes from previous:
0.33 2011-09-28
- add conflict on MooseX::Method::Signatures 0.36 (ether)
0.32 2011-09-05
- bring the behavior of has_symbol for nonexistant scalars into line with
the xs version
- invalid package names (for instance, Foo:Bar) are not allowed
- invalid stash entry names (anything containing ::) are not allowed
Changelog:
0.5
- A Template is explicitly disallowed
from having a url that normalizes to relative outside
of the root. That is, if the Lookup is based
at /home/mytemplates, an include that would place
the ultimate template at
/home/mytemplates/../some_other_directory,
i.e. outside of /home/mytemplates,
is disallowed. This usage was never intended
despite the lack of an explicit check.
The main issue this causes
is that module files can be written outside
of the module root (or raise an error, if file perms aren't
set up), and can also lead to the same template being
cached in the lookup under multiple, relative roots.
TemplateLookup instead has always supported multiple
file roots for this purpose.
[ticket:174]
0.4.2
- Fixed bug regarding <%call>/def calls w/ content
whereby the identity of the "caller" callable
inside the <%def> would be corrupted by the
presence of another <%call> in the same block.
[ticket:170]
- Fixed the babel plugin to accommodate <%block>
[ticket:169]
0.4.1
- New tag: <%block>. A variant on <%def> that
evaluates its contents in-place.
Can be named or anonymous,
the named version is intended for inheritance
layouts where any given section can be
surrounded by the <%block> tag in order for
it to become overrideable by inheriting
templates, without the need to specify a
top-level <%def> plus explicit call.
Modified scoping and argument rules as well as a
more strictly enforced usage scheme make it ideal
for this purpose without at all replacing most
other things that defs are still good for.
Lots of new docs. [ticket:164]
- a slight adjustment to the "highlight" logic
for generating template bound stacktraces.
Will stick to known template source lines
without any extra guessing. [ticket:165]
0.4.0
- A 20% speedup for a basic two-page
inheritance setup rendering
a table of escaped data
(see http://techspot.zzzeek.org/2010/11/19/quick-mako-vs.-jinja-speed-test/).
A few configurational changes which
affect those in the I-don't-do-unicode
camp should be noted below.
- The FastEncodingBuffer is now used
by default instead of cStringIO or StringIO,
regardless of whether output_encoding
is set to None or not. FEB is faster than
both. Only StringIO allows bytestrings
of unknown encoding to pass right
through, however - while it is of course
not recommended to send bytestrings of unknown
encoding to the output stream, this
mode of usage can be re-enabled by
setting the flag bytestring_passthrough
to True.
- disable_unicode mode requires that
output_encoding be set to None - it also
forces the bytestring_passthrough flag
to True.
- the <%namespace> tag raises an error
if the 'template' and 'module' attributes
are specified at the same time in
one tag. A different class is used
for each case which allows a reduction in
runtime conditional logic and function
call overhead. [ticket:156]
- the keys() in the Context, as well as
it's internal _data dictionary, now
include just what was specified to
render() as well as Mako builtins
'caller', 'capture'. The contents
of __builtin__ are no longer copied.
Thanks to Daniel Lopez for pointing
this out. [ticket:159]
Changes from previous:
0.46 2011-01-23 16:45 UTC
+ Fix : [RT #64866] : Assertion failure with perl 5.13.9.
The real problem was that dispell() wasn't properly resetting
the magical sv flags, which became visible in the test suite
only with perl 5.13.9.
Thanks Joshua ben Jore for reporting.
0.45 2010-11-21 23:15 UTC
This is a maintenance release. The code contains no functional change.
Users of 0.44 can skip this update.
+ Doc : C++ compilers are officially NOT supported.
+ Doc : The minimum perl 5.10.0 requirement for uvar magic has been
made more explicit.
Thanks Peter Rabbitson for pointing this out and contributing a
patch.
+ Tst : Tune for perl 5.13.7.
+ Tst : Capture::Tiny will be used in t/17-ctl.t if and only if it can
capture a simple run.
0.44 2010-09-24 19:10 UTC
+ Fix : Broken linkage on Windows with gcc 3.4, which appears in
particular when using ActivePerl's default compiler suite.
For those setups, the Variable::Magic shared library will now
be linked against the perl dll directly (instead of the import
library). This (should) fix RT #51483.
Thanks Christian Walde for helping to reproduce this failure
and extra testing.
+ Rem : Support for development perls from the 5.11 branch but older
than the 5.11.0 release was removed. This could cause more
recent setups to fail.
+ Tst : Threads tests are now only run on perl 5.13.4 and higher.
They could segfault randomly because of what seems to be an
internal bug of Perl, which has been addressed in 5.13.4.
There is also an environment variable that allows you to
forcefully run those tests, but it should be set only for
author testing and not for end users.
Changes from previous:
0.31 [2010-10-10] Or the "Yay - an actual release!" release
- Same as 0.30_2
0.30_2 [2010-10-06] Or the "oh what a to do" release
- Added a bunch of folk to the acknowledgements
- Added some clarifying documentation to respond to RT#59293
- Marked a test that was failing under T::B 2.0 until we figure out
whether it should pass or not. See http://is.gd/fNOFb
0.30_1 [2010-10-04] Or the "Peter Rabbitson did all the work" release
- Added dates to changes file, as far as we can from backpan et al
- Fix for DB::args bug (thanks Peter Rabbitson)
- Fix for bizarre-copy bug (thanks Peter Rabbitson)
Changes from previous:
0.11 2011-10-04
- Don't accept package names that start with a digit. ( Jesse Luehrs )
- Rewrote some of the guts to use Module::Runtime, rather than reimplementing
its functionality. ( Jesse Luehrs )
0.10 2011-09-06
- Make sure the $@ localization doesn't hide errors - invalid module name
errors were being suppressed on perls before 5.14. ( Jesse Luehrs )
0.09 2011-09-05
- Fix is_class_loaded to ignore $ISA (but still look for @ISA) when trying to
determine whether a class is loaded. ( Jesse Luehrs )
- Lots of internals cleanup. ( Jesse Luehrs )
* The scripting part of the codebase is getting prepared for i18n/l10n.
* Interix, Cygwin and Minix ports got updated.
* Various updates to git-p4 (in contrib/), fast-import, and git-svn.
* Gitweb learned to read from /etc/gitweb-common.conf when it exists,
before reading from gitweb_config.perl or from /etc/gitweb.conf
* Various codepaths that invoked zlib deflate/inflate assumed that these
functions can compress or uncompress more than 4GB data in one call on
platforms with 64-bit long, which has been corrected.
* Git now recognizes loose objects written by other implementations that
use a non-standard window size for zlib deflation (e.g. Agit running on
Android with 4kb window). We used to reject anything that was not
deflated with 32kb window.
* Interaction between the use of pager and coloring of the output has
been improved, especially when a command that is not built-in was
involved.
* "git am" learned to pass the "--exclude=<path>" option through to underlying
"git apply".
* You can now feed many empty lines before feeding an mbox file to "git am".
* "git archive" can be told to pass the output to gzip compression and
produce "archive.tar.gz".
* "git bisect" can be used in a bare repository (provided that the test
you perform per each iteration does not need a working tree, of course).
* The length of abbreviated object names in "git branch -v" output
now honors the core.abbrev configuration variable.
* "git check-attr" can take relative paths from the command line.
* "git check-attr" learned an "--all" option to list the attributes for a
given path.
* "git checkout" (both the code to update the files upon checking out a
different branch and the code to checkout a specific set of files) learned
to stream the data from object store when possible, without having to
read the entire contents of a file into memory first. An earlier round
of this code that is not in any released version had a large leak but
now it has been plugged.
* "git clone" can now take a "--config key=value" option to set the
repository configuration options that affect the initial checkout.
* "git commit <paths>..." now lets you feed relative pathspecs that
refer to outside your current subdirectory.
* "git diff --stat" learned a --stat-count option to limit the output of
a diffstat report.
* "git diff" learned a "--histogram" option to use a different diff
generation machinery stolen from jgit, which might give better performance.
* "git diff" had a weird worst case behaviour that can be triggered
when comparing files with potentially many places that could match.
* "git fetch", "git push" and friends no longer show connection
errors for addresses that couldn't be connected to when at least one
address succeeds (this is arguably a regression but a deliberate one).
* "git grep" learned "--break" and "--heading" options, to let users mimic
the output format of "ack".
* "git grep" learned a "-W" option that shows wider context using the same
logic used by "git diff" to determine the hunk header.
* Invoking the low-level "git http-fetch" without "-a" option (which
git itself never did---normal users should not have to worry about
this) is now deprecated.
* The "--decorate" option to "git log" and its family learned to
highlight grafted and replaced commits.
* "git rebase master topci" no longer spews usage hints after giving
the "fatal: no such branch: topci" error message.
* The recursive merge strategy implementation got a fairly large
fix for many corner cases that may rarely happen in real world
projects (it has been verified that none of the 16000+ merges in
the Linux kernel history back to v2.6.12 is affected with the
corner case bugs this update fixes).
* "git stash" learned an "--include-untracked option".
* "git submodule update" used to stop at the first error updating a
submodule; it now goes on to update other submodules that can be
updated, and reports the ones with errors at the end.
* "git push" can be told with the "--recurse-submodules=check" option to
refuse pushing of the supermodule, if any of its submodules'
commits hasn't been pushed out to their remotes.
* "git upload-pack" and "git receive-pack" learned to pretend that only a
subset of the refs exist in a repository. This may help a site to
put many tiny repositories into one repository (this would not be
useful for larger repositories as repacking would be problematic).
* "git verify-pack" has been rewritten to use the "index-pack" machinery
that is more efficient in reading objects in packfiles.
* test scripts for gitweb tried to run even when CGI-related perl modules
are not installed; they now exit early when the latter are unavailable.
New Libraries
* Chrono: Useful time utilities
* Geometry: Geometry Library
* Phoenix: Define small unnamed function objects at the actual call site, and
more
* Ratio: Compile time rational arithmetic
More: http://www.boost.org/users/history/version_1_47_0.html
and accessing command line arguments. It was intially inspired by the user
friendly CLAP libary. The difference is that this library is templatized, so the
argument class is type independent. Type independence avoids
identical-except-for-type objects, such as IntArg, FloatArg, and StringArg.
While the library is not strictly compliant with the GNU or POSIX standards, it
is close.
TCLAP is written in ANSI C++ and is meant to be compatible with any
standards-compliant C++ compiler. It is known to work on Linux, MacOS X,
Windows, and Solaris platforms. The library is implemented entirely in header
files making it easy to use and distribute with other software. It is licensed
under the MIT License for worry free distribution.
TCLAP is now a mature, stable, and feature rich package. Unless I get really
bored, it probably won't see much further development aside from bug fixes and
compatibility updates. Please don't let any apparent project inactivity
discourage you from using the software!
This is a regular bugfix release.
setup: build inotify for sys.platform='linux*'
bundlerepo: add argument check before unlink
hgweb: properly check for bookmarks when drawing graph
http: handle push of bundles > 2 GB again (issue3017)
keyword: preserve file mode when overwriting
osutil: avoid accidentally destroying the True object in isgui (issue2937)
patch: correctly handle non-tabular Subject: line
patch: handle 'gitpatches' being empty, but not none
record: use command wrapper properly for qnew/qrefresh (issue3001)
setdiscovery: fix hang when #heads>200 (issue2971)
ui: also swap sys.stdout with self.fout in _readline
ui: write traceback to self.ferr
url: Remove the proxy env variables only when needed (issue2451)
url: parse fragments first (issue2997)
util: fix crash converting an invalid future date to string
util: wrap lines with multi-byte characters correctly (issue2943)
and... well, the kqueue support is built but not used!
Now, running a test program provided by Dmitry DTRT (with the exception
of two tests, which seem to be not working in macppc). Ride the previous
revision bump.
the Google Summer of Code 2011 program. Congratulations, Dmitry! :-)
This comes in the form of a new 'kqueue' build-time option. This should NOT
be an option, but it is for two reasons:
- I consider this still experimental for it to be enabled on everyone's
machines, and the code is not "official".
- The patch requires us to run automake/autoconf during the pre-configure
stage, and I don't want to add these dependencies on the default package
build.
When these patches get integrated upstream, we can get rid of the option
altogether.
Also note that I had to backport some of our custom changes to Makefile.in
files and some to configure to the original Makefile.am files and
configure.ac. This is to prevent our custom patches from being lost during
the automake/autoconf invocations we now do when 'kqueue' is enabled.
Bump PKGREVISION to 1.
* README now gives a quick into to using mr.
* Brought back the "deleted" parameter, which provides an easy way
to mark repositories that should be removed.
* Allow untrusted mrconfig files to set parameters to true/false.
So skip=true or deleted=true can be used in an untrusted mrconfig
file.
* Also allow order=N in an untrusted mrconfig file.
* Support bzr checkouts, which are updated with "bzr update", and
to which bzr automatically pushes commits. Closes: #643589
* Use bzr branch, not deprecated bzr clone when registering bzr
repositories. Closes: #643591
* Allow bzr branch|clone|get|checkout in untrusted mrconfig files.
* Avoid using sed -r in git-fake-bare, for OSX portability.
* git-fake-bare: handle fake bare repositories with core.bare not
set (Thanks, Julien Rebetez)
Changes from previous:
0.26 2011-06-06
- The default() method for class attributes always returned a value, even if
the default was a subroutine ref, which isn't how the method works for
regular attributes. This broke inlining with Moose HEAD.
0.25 2011-06-05
- Class attributes now have a definition context set, which means that errors
thrown from generated methods associated with these attributes say something
like "X at accessor MyClass::ClassAttribute (defined at path/to/file line
42)" rather than "generated method (unknown origin)".
0.24 2011-02-22
- This release provides forward compatibility with Moose 1.99+. It will still
work with Moose 1.23 as well. Partially based on work originally done by
Moritz Onken.
0.23 2011-02-13
- Fixed a bug where applying a role with class attributes didn't record the
role as actually being applied in the meta class (or role) to which it was
applied. Reported by Karen Etheridge. RT #59610.
- Applying multiple roles to a class lost all class attributes from those
roles. Fixed by Andrew Rodland. RT #59572.
0.22 2011-02-02
- Explicitly require namespace::clean 0.20 to avoid some bad interactions
between namespace::clean and Package::Stash.
0.21 2010-10-29
- Switch from Test::Exception to Test::Fatal.
0.20 2010-10-07
- A test file tried to load MooseX::Role::Parameterized, which was not listed
as a dep. Reported by Andreas Koenig. RT #61957.
0.19 2010-10-06
- Removed references to MooseX::ClassAttribute::Meta::Method::Accessor, which
was removed in 0.18. This caused lots of test failures if you hadn't
installed a previous version of this distribution.
0.18 2010-10-05
- Changes to work with (and require) Moose 1.15.
0.17 2010-09-26
- Changes to work with (and require) Moose 1.09.
Version 0.9.6
http://svn.edgewall.org/repos/babel/tags/0.9.6/
(Mar 17 2011, from branches/stable/0.9.x)
* Backport r493-494: documentation typo fixes.
* Make the CLDR import script work with Python 2.7.
* Fix various typos.
* Fixed Python 2.3 compatibility (ticket #146, #233).
* Sort output of list-locales.
* Make the POT-Creation-Date of the catalog being updated equal to
POT-Creation-Date of the template used to update (ticket #148).
* Use a more explicit error message if no option or argument (command) is
passed to pybabel (ticket #81).
* Keep the PO-Revision-Date if it is not the default value (ticket #148).
* Make --no-wrap work by reworking --width's default and mimic xgettext's
behaviour of always wrapping comments (ticket #145).
* Fixed negative offset handling of Catalog._set_mime_headers (ticket #165).
* Add --project and --version options for commandline (ticket #173).
* Add a __ne__() method to the Local class.
* Explicitly sort instead of using sorted() and don't assume ordering
(Python 2.3 and Jython compatibility).
* Removed ValueError raising for string formatting message checkers if the
string does not contain any string formattings (ticket #150).
* Fix Serbian plural forms (ticket #213).
* Small speed improvement in format_date() (ticket #216).
* Fix number formatting for locales where CLDR specifies alt or draft
items (ticket #217)
* Fix bad check in format_time (ticket #257, reported with patch and tests by
jomae)
* Fix so frontend.CommandLineInterface.run does not accumulate logging
handlers (#227, reported with initial patch by dfraser)
* Fix exception if environment contains an invalid locale setting (#200)
0.9.12
~~~~~~
This is a very big release. We've made huge improvements on three fronts:
1. Test failures are way nicer and easier to read
2. Matchers and ``assertThat`` are much more convenient to use
3. Correct handling of extended unicode characters
We've trimmed off the fat from the stack trace you get when tests fail, we've
cut out the bits of error messages that just didn't help, we've made it easier
to annotate mismatch failures, to compare complex objects and to match raised
exceptions.
Testing code was never this fun.
Changes
-------
* ``AfterPreproccessing`` renamed to ``AfterPreprocessing``, which is a more
correct spelling. Old name preserved for backwards compatibility, but is
now deprecated. Please stop using it.
(Jonathan Lange, #813460)
* ``assertThat`` raises ``MismatchError`` instead of
``TestCase.failureException``. ``MismatchError`` is a subclass of
``AssertionError``, so in most cases this change will not matter. However,
if ``self.failureException`` has been set to a non-default value, then
mismatches will become test errors rather than test failures.
* ``gather_details`` takes two dicts, rather than two detailed objects.
(Jonathan Lange, #801027)
* ``MatchesRegex`` mismatch now says "<value> does not match /<regex>/" rather
than "<regex> did not match <value>". The regular expression contains fewer
backslashes too. (Jonathan Lange, #818079)
* Tests that run with ``AsynchronousDeferredRunTest`` now have the ``reactor``
attribute set to the running reactor. (Jonathan Lange, #720749)
Improvements
------------
* All public matchers are now in ``testtools.matchers.__all__``.
(Jonathan Lange, #784859)
* ``assertThat`` can actually display mismatches and matchers that contain
extended unicode characters. (Jonathan Lange, Martin [gz], #804127)
* ``assertThat`` output is much less verbose, displaying only what the mismatch
tells us to display. Old-style verbose output can be had by passing
``verbose=True`` to assertThat. (Jonathan Lange, #675323, #593190)
* ``assertThat`` accepts a message which will be used to annotate the matcher.
This can be given as a third parameter or as a keyword parameter.
(Robert Collins)
* Automated the Launchpad part of the release process.
(Jonathan Lange, #623486)
* Correctly display non-ASCII unicode output on terminals that claim to have a
unicode encoding. (Martin [gz], #804122)
* ``DocTestMatches`` correctly handles unicode output from examples, rather
than raising an error. (Martin [gz], #764170)
* ``ErrorHolder`` and ``PlaceHolder`` added to docs. (Jonathan Lange, #816597)
* ``ExpectedException`` now matches any exception of the given type by
default, and also allows specifying a ``Matcher`` rather than a mere regular
expression. (Jonathan Lange, #791889)
* ``FixtureSuite`` added, allows test suites to run with a given fixture.
(Jonathan Lange)
* Hide testtools's own stack frames when displaying tracebacks, making it
easier for test authors to focus on their errors.
(Jonathan Lange, Martin [gz], #788974)
* Less boilerplate displayed in test failures and errors.
(Jonathan Lange, #660852)
* ``MatchesException`` now allows you to match exceptions against any matcher,
rather than just regular expressions. (Jonathan Lange, #791889)
* ``MatchesException`` now permits a tuple of types rather than a single type
(when using the type matching mode). (Robert Collins)
* ``MatchesStructure.byEquality`` added to make the common case of matching
many attributes by equality much easier. ``MatchesStructure.byMatcher``
added in case folk want to match by things other than equality.
(Jonathan Lange)
* New convenience assertions, ``assertIsNone`` and ``assertIsNotNone``.
(Christian Kampka)
* New matchers:
* ``AllMatch`` matches many values against a single matcher.
(Jonathan Lange, #615108)
* ``Contains``. (Robert Collins)
* ``GreaterThan``. (Christian Kampka)
* New helper, ``safe_hasattr`` added. (Jonathan Lange)
* ``reraise`` added to ``testtools.compat``. (Jonathan Lange)
HDF5 is a data model, library, and file format for storing and
managing data. It supports an unlimited variety of datatypes, and
is designed for flexible and efficient I/O and for high volume and
complex data. HDF5 is portable and is extensible, allowing applications
to evolve in their use of HDF5. The HDF5 Technology suite includes
tools and applications for managing, manipulating, viewing, and
analyzing data in the HDF5 format.
This package provides C++ wrappers for the HDF5 C library.
Changes from previous:
0.98 Wed, 23 Feb 2011 14:38:02 +1100
Bug Fixes
* subtest() should not fail if $? is non-zero. (Aaron Crane)
Docs
* The behavior of is() and undef has been documented. (Pedro Melo)
0.97_01 Fri Aug 27 22:50:30 PDT 2010
Test Fixes
* Adapted the tests for the new Perl 5.14 regex stringification.
(Karl Williamson) [github 44]
Doc Fixes
* Document how to test "use Foo ()". (Todd Rinaldo) [github 41]
Feature Changes
* subtest() no longer has a prototype. It was just getting in the way.
[rt.cpan.org 54239]
* The filehandles used by default will now inherit any filehandle
disciplines from STDOUT and STDERR IF AND ONLY IF they were applied
before Test::Builder is loaded. More later. [rt.cpan.org 46542]
Changes from previous:
0.3.0
- Reverted feature to match POD head section =head2, =head3 and =head4
because of backward compatibility issue (RT bug #70942)
0.2.9 2011-09-03
- This command does not generate warnings anymore:
perl -e 'use Getopt::Euclid; print $Getopt::Euclid::VERSION;'
- Required and optional arguments can now be put in a POD head section
=head2, =head3 and =head4 for more flexibility
0.2.8 2011-09-02
- Support for optional defaults (RT bug #61438, patch from Paolo Medeo)
- Extended the grammar of recognized POD headers
- Usage error message now mentions the --man option in addition to --help
Changes from previous:
0.22 2011-05-09
- Fix issues where the metaclass gets reinitialized after the call to
'extends' but before 'make_immutable'. This could happen if a role
used an extension which provided an application_to_class metarole,
since the role application would then apply a metarole to the class,
and metarole application currently causes metaclass reinitialization
in Moose. (ugh.)
0.21 2011-04-29
- Allow this module to work with constructors with names other than
'new'. If you're extending a class with a constructor named something
other than 'new', you should declare this when calling extends, as in:
extends 'Foo' => { -constructor_name => 'create' };
This will ensure that calling 'create' will also call Moose's
constructor.
0.20 2011-03-22
- fix warning when passing inline_constructor => 0 with no superclass new
method (rafl).
0.19 2011-03-02
- don't die if superclass doesn't have a 'new' method
0.18 2011-02-09
- forward compat for Moose 2.0
0.17 2010-11-08
- convert to Test::Fatal, and dep on it
0.16 2010-10-05
- Make it work with Moose 1.15 (Dave Rolsky).
Allow maximum number of targets in initiator to be set in mk.conf
with ISCSI_INITIATOR_MAX_TARGETS
Changes:
Provide function in libiscsi to return the compile-time limit on number of
targets. Use this function in the initiator, to ensure that the same limit
is used throughout (i.e. stop you attempting to compile the initiator with a
different limit to the library - the initiator uses the library for most of
its functionality).
If truncating the number of targets in the initiator, truncate it to the
correct amount, not half of what it should be.
Obey RFC3720 when determining targets. SendTargets=All MUST be supported
on a discovery session, and MUST NOT be supported on an operational session.
Previously, SendTargets=All was used both in the initial discovery session
and the later session. (from Daisuke Aoyama - author of istgt).
The GPRBuild Solaris targets were too specific and didn't recognize
GNAT-AUX on Solaris (actually OpenIndiana). This patch makes sure
GPRBuild-AUX can use GNAT-AUX and any other gcc-based compiler.
so it can be closed without affecting anybody else. gcc < 3.4 used to
have a separate argument to control this behaviour which was apparently
removed.
Untested, please report any problems to GNATS.
Changelog:
2011-06-13 release 1.0.22 of cvsd
changes since 1.0.21:
+ don't log EINTR on select() any more, not even in debug mode
+ fix for cvsd-buildroot to also work on multiarch setups
+ log address and port with bind() failures
+ Debian packaging updates
2010-09-08 release 1.0.21 of cvsd
changes since 1.0.20:
+ handle failure to bind() as a fatal error now
2010-09-05 release 1.0.20 of cvsd
changes since 1.0.19:
+ correctly listen on IPv4 and IPv6 addresses with recent Glibc
versions by not depending on the order of address records returned by
getaddrinfo() and work regardless of net.ipv6.bindv6only sysctl
2010-08-17 release 1.0.19 of cvsd
changes since 1.0.18:
+ cvsd-buildroot: ignore commented out lines in CVSROOT/passwd files
+ cvsd-buildroot: set an umask for generated files
+ some documentation updates
+ change init script dependency on $remote_fs (for /usr) from Should
to Required (thanks lintian)
+ Debian packaging improvements
2010-01-14 release 1.0.18 of cvsd
changes since 1.0.17:
+ use simpler shell semantics in cvsd-buildroot to fix a problem
with bash 4
+ fix call to uname in the cvsd-buginfo script
2009-12-30 release 1.0.17 of cvsd
changes since 1.0.16:
+ update to automake 1.11
+ some small spelling fixes in documentation
+ changed references to home page and contact email addresses to use
arthurdejong.org
+ Debian packaging improvements
Drastically improved memory handling for certain use cases
Added a new rendering backend to speed up Canvas operations on Windows systems
Bookmark and password changes now sync almost instantly when using Firefox Sync
The 'http://' URL prefix is now hidden by default
Added support for text-overflow: ellipsis
Added support for the Web Timing specification
Enhanced support for MathML
The WebSocket protocol has been updated from version 7 to version 8
Added an opt-in system for users to send performance data back to Mozilla
to improve future versions of Firefox
Fixed several stability issues
Fixed several security issues
* Use a monotonically increasing timer
* Immediately expire items when given a negative expiration time
* Fix memcached-tool to print about all slabs
* Properly daemonize memcached for debian
* Don't permanently close UDP listeners on error
* Allow memcached-init to start multiple instances (not recommended)
* Issue 214: Search for network libraries before searching for libevent
* Issue 213: Search for clock_gettime in librt
* Issue 115: accont for CAS in item_size_ok
* Fix incredibly slim race for maxconns handler. Should no longer hang ever
* Issue 183: Reclaim items dead by flush_all
* Issue 200: Don't fire dtrace probe as the last thing in a function
* Several SSL bufferevent fixes
* A Solaris sendfile bugfix
* A few other IOCP and rate-limiting bufferevent fixes
* Memory leak fixes
* Coverity fixes
* Portability and build fixes
* and more...
- The makefile's substitution of PREFIX and SH accidentally got rolled
into patch-ac at some point and thereby hardcoded. Since all patch-ac
was doing was setting these, make the substitution work without a
patch and remove patch-ac entirely.
- Commit some LP64 fixes I've had sitting in my tree since 2008 for
some reason.
- Fix some pkglint.
- PKGREVISION -> 4.
XXX: This is not right, but the way this thing is built doesn't give
XXX: too many choices at the moment. Should be fixed properly sometime.
XXX: This way the package is at least installable...
(for some reason there are two separate copies of it) to match the
POSIX definition.
Fixing this "right" would be a lot of work and does not seem
worthwhile, especially since this package looks to be more or less
dead upstream and is probably a good candidate for removal.
Revision history for Scope-Upper
0.16 2011-09-03 23:00 UTC
+ Add : uplevel($code, @args, $cxt) executes $code with arguments @args
in the upper context pointed by $cxt.
This is an XS version of the well-known uplevel() routine from
Sub::Uplevel. There are a few differences between both
implmentations that are listed in the documentation.
The XS version is roughly 10 times faster than the pure-Perl
version.
0.15 2011-08-24 14:20 UTC
+ Fix : Localizing subroutines in an higher scope will now correctly
update the method cache.
0.14 2011-02-27 00:00 UTC
+ Fix : [RT #64997] : Compatibility with perl 5.13.10.
Thanks Dave Mitchell for the notice.
+ Tst : Lengthy tests have been ported to Test::Leaner, making the
whole test suite about 50% faster.
0.13 2010-12-20 01:00 UTC
+ Fix : [RT #61904] : Stack corruption when using unwind() under
Devel::NYTProf.
Thanks Sergey Aleynikov for contributing a fix.
+ Fix : [RT #63378] : Compatibility with perl 5.13.8.
Thanks Andreas J. König for bisecting the issue.
+ Tst : Threads tests are now only run on perl 5.13.4 and higher.
They could segfault randomly because of what seems to be an
internal bug of Perl, which has been addressed in 5.13.4.
There is also an environment variable that allows you to
forcefully run those tests, but it should be set only for
author testing and not for end users.
Target "errorcheck" ("configure") cannot work in parallel.
Fix bug appeared in parallel builds (make -j).
Thanks to Alexander Nasonov.
Support for IRIX64 and Haiku was implemented.
PDF presentation update
Minor clean-ups
to kernel.org issues.
1.7.6.3:
* "git -c var=value subcmd" misparsed the custom configuration when
value contained an equal sign.
* "git fetch" had a major performance regression, wasting many
needless cycles in a repository where there is no submodules
present. This was especially bad, when there were many refs.
* "git reflog $refname" did not default to the "show" subcommand as
the documentation advertised the command to do.
* "git reset" did not leave meaningful log message in the reflog.
* "git status --ignored" did not show ignored items when there is no
untracked items.
* "git tag --contains $commit" was unnecessarily inefficient.
Also contains minor fixes and documentation updates.
1.7.6.2:
Junio C Hamano (3):
whitespace: have SP on both sides of an assignment "="
Revert "Merge branch 'cb/maint-quiet-push' into maint"
Git 1.7.6.2
Pang Yan Han (1):
update-ref: whitespace fix
Thomas Rast (1):
Documentation: clarify effects of -- <path> arguments
Ragel 6.7 - May 22, 2011
========================
-The C vim file now supports L,l on the end of literal numbers, other syntax
highlighting fixes.
-Added an explicit cast when modifying _trans to eliminate compiler warnings
on Windows/VC++
-Fix for ruby 1.9 compatibility.
-Fix for include directories (-I option) on Windows/VC++
-Can now rename data variable in C# code generator.
-Fix for non-char alphtype in C# code generator.
-Fix for signedness of wide char in C code generator. when comparing the wide
type against a literal we need to pick the right signedness for the literal.
-Fixed arithmetic overflow in generated C# code. The C# _acts and _nacts vars
should not be typed based on the size of the array elements they point to.
Fixes a bug reported by Attila Sztupák.
-Made the -L option work for Ruby.
-Enhanced ragel.m4 (from Diego).
-Applied GO patch from Justine Tunney.
-Applied D2 patch from Johannes Pfau.
-Applied Ocaml patch from ygrek.
v0.7.1 Added support for :before_method and :on_exception callback types.
Added support for registering multiple callbacks of a given type.
v0.7.0 Updated thrift gem dependency to 0.7.0
=== 2.3.0 / 2011-09-06
* 2 minor enhancements:
* Add -g flag to parser compile if DEBUG
* Lexer now embeds line number in yacc_value for keywords, helping fix up line numbers
* 3 bug fixes:
* Fix method line numbers when no args and no parens (quix)
* Fixed line numbers on return/break/next w/ result expr. (pjnz)
* Fixed some lexing state in order to parse: 'f (1), 2' as 'f(1, 2)'. (invernizzi)
= 0.10.0 ()
* Add Expectation#throws to allow a stubbed method to use Kernel#throw.
* Updates for versions of Test::Unit up to and including v2.3.3 (including
patch by Jens Fahnenbruck).
* Updates for versions of MiniTest up to and including v2.5.1.
* Since the singleton method added by Mocha masks the underlying instance
method, there's no need to move it out the way and then back again. This
fixes Github issue #20, because the original method is left unchanged -
https://github.com/floehopper/mocha/issues/20 (thanks to Nick Lewis).
* Handle stubbing of a singleton method, leaving the original method unchanged
after the test.
* When stubbing an instance method that was originally defined as a singleton
method, the original method should still exist after the test.
* Fixed mis-print in Mocha::ObjectMethods#unstub documentation (patch by Gleb
Pomykalov).
* Improved test coverage around stubbing of methods defined in different ways
- this makes use of the newly extracted introspection gem (although this
means some tests are now failing in Ruby v1.9.2).
* Added configuration for Travis continuous integration.
* Make the gemspec the canonical reference and stop generating it from the
Rakefile.
* Use the built-in Bundler rake tasks for packaging the gem.
* Use the "release" rake task provided by Bundler instead of using the
Rake::XForge::Release functionality.
* Extract Object#__metaclass__ into a new metaclass gem.
* Run rake tasks without `bundle exec`.
* Avoid deprecation warning for rdoc rake task.
* Remove the `use_test_unit_gem` MOCHA_OPTION which hasn't worked since we
switched to bundler - we can now run the tests specifying a different
Gemfile instead.
* Use multiple Gemfiles seems to run Travis CI builds against multiple version
of test-unit & minitest.
2010-12-28 Andrew McMillan <andrew@morphoss.com>
* Release 0.46
* [classBrowser] Fix checking of ordering fieldname.
* [Validation] Remove references to outdated ereg() functions.
* [AWLUtlities] Rename regex checking function more appropriately.
* [Session] Don't use ereg() for parsing regex.
* [XMLDocument] Use __construct style for constructor.
2010-12-24 Andrew McMillan <andrew@morphoss.com>
* [AwlQuery] Make the quote() method public/static.
2010-12-23 Andrew McMillan <andrew@morphoss.com>
* Convert everything except PgQuery to use AwlQuery
* [AwlQuery] Add minor missing functionality so we can switch.
* [vComponent] Improve parameter splitting to allow for escapes.
* [classEditor] Avoid uninitialised variable warning.
2010-11-27 Andrew McMillan <andrew@morphoss.com>
* [vComponent] Add __toString() methods.
* [Session] If someone is made inactive then their long-term login should cease.
2010-10-15 Andrew McMillan <andrew@morphoss.com>
* [AWLUtilities] Add an awl_version() function.
2010-10-10 Andrew McMillan <andrew@morphoss.com>
* [vComponent] It is OK to SetComponents() to an empty set.
cgdb-0.6.5 (01/01/2010)
* Turn autosourcereload to on by default. This matches GDB's behavior.
* CGDB should successfully build on Mac OS X 10.5.
Thanks Kyle McKay for sending in this patch.
* Add support for :up and :down in the cgdb status bar.
Thanks Benoit Pierre for reporting this.
* Enhanced tgdb in a2 mode. It now understands GNAT's
pre-instance-choice, instance-choice, post-instance-choice extension.
* Made searching and :commands a little more vi-like. If you abort a
search, you can still recall the last search with 'n' or 'N'. If you
backspace to the beginning of the line, a :command or /search is
aborted.
* Added some new cgdb commands.
:c is now a synonym for :continue
:f is now a synonym for :finish
:n is now a synonym for :next
:r is not a synonym for :run
:s is now a synonym for :step
:start is a new command that will send "start" to the debugger
:kill is a new command that will send "kill" to the debugger
* Shortcut mode has been removed from cgdb. If you liked it, you can
put these commands in your cgdbrc file,
map r :run<CR>
map c :continue<CR>
map f :finish<CR>
map n :next<CR>
map s :step<CR>
and you will get the same functionality. Do not forget that,
F5 is the same as 'run'.
F6 is the same as 'continue'.
F7 is the same as 'finish'.
F8 is the same as 'next'.
F10 is the same as 'step'.
are always available.
* Added a feature to CGDB which allows it to ask readline what key
sequence it uses for the Home and End keys. That way, if the terminfo
database is wrong, but the OS vendor modified /etc/inputrc, CGDB will
have a working Home and End key.
* CGDB fully supports the use of maps now. The feature is similar in
nature to the vim map feature. It supports map, unmap, imap
and iunmap. The map/unmap work in CGDB mode and the imap/iunmap
work in GDB mode.
* Added the cgdbmodekey configuration option. This allows the user to
change the key that switches them from GDB mode to CGDB mode. The
default value is the Escape key. This is useful for users that want
to use readline in vi-mode.
* Fixed a bug in CGDB where the Insert and Delete keys were not
recognized properly all of the time.
cgdb-0.6.4 (04/28/2007)
* CGDB can now be compiled with -jN option. It would fail previously.
* CGDB should now build on freebsd with out any patches.
* Added a new highlighting group, SelectedLineNr. From the documentation:
This represents the group that is used when CGDB is displaying the
currently selected line. This is the line that the cursor is on.
* CGDB shuts down now when it receives C-d (control d), just like GDB
does.
* Changed the GDB prompt from "(tgdb)" to "(gdb)". I think users
will be more comfortable with "(gdb)".
* Fixed a serious bug in CGDB. After the user entered a command
CGDB would not accept any more commands. This would only
happen sometimes, if at all. This is due to the fact that CGDB
had a bug in it that was sensitive to the amount of data that
the read system call returns to it.
* Fixed a bug in CGDB where it would not show a file in the
source viewer if there was a space in the path.
* Removed the escdelay option. It has been replaced by the
timeout, timeoutlen, ttimeout and ttimeoutlen options. Please
see the CGDB manual for more documentation on these options.
* Fixed a bug in libkui. If the user quickly typed ESC followed
by an 'o' while in the GDB window, the 'o' character would not
get processed by CGDB until another character was hit.
* CGDB has improved it's -d (Set debugger to use) command line
option. CGDB used to force the user to have a space between the
-d and the debugger to use, like "cgdb -d /usr/bin/gdb". Now the
user can also start CGDB like "cgdb -d/usr/bin/gdb".
* CGDB now honor's the '--' command line option, which will allow the
user to explicitly say when they want to end passing arguments to
CGDB and begin passing arguments to the debugger.
* Fixed another bug in the configure script. If the getopt.h header
file is not available on the system, then CGDB will fail to configure.
It should finish, and most likely build successfully.
* Fixed yet another bug in the auto tools scripts. cgdb-0.6.3 would
improperly install gdbmi_driver if make was invoked like
'make DESTDIR=/foo install'.
* Fixed another bug in the configure script. If the curses.h header
file is not available on the system, then CGDB will fail to configure.
It would previously configure and then end up with a compile error.
cgdb-0.6.3 (06/03/2006)
* CGDB will no longer crash if you attempt to do a regular expression
search with / when there is no file in the source viewer.
* CGDB now uses a psuedo terminal (PTY) to start GDB on, instead of
using a pipe. Jens Frederich reported that on a Mac, typing Ctrl-C
while the inferior was running would cause GDB to shut down, thus
causing CGDB to shut down. This patch fixes that problem. At the
same time, this patch caused GDB to ask the user several other
questions regarding shared breakpoints, and quiting while the
inferior is running. Thus, CGDB acts more as if you were just using
GDB.
* CGDB is now maintained in a subversion repository instead of a
CVS repository. Although this might not appear to effect end users,
it is possible that there will be problems associated with this
switch. For this reason, I have documented this here.
* Fixed another bug in the configure script. CGDB should now be able
to configure itself on a Mac. Thanks to Jens Frederich for finding
and reporting this problem.
* CGDB no longer packages the generated auto tool files in CVS. However,
the distributions will have these files. This means you will need
autoconf, autoheader and automake installed to compile CGDB from CVS.
* CGDB now compiles according to the c89 standard. This should help with
porting to older systems.
* The file cgdb.txt is now installed into $PREFIX/share/cgdb/cgdb.txt. This
is the file CGDB will now use to display the manual when :help is typed.
This prevents CGDB from copying the data into the users home directory.
cgdb-0.6.2 (04/09/2006)
* CGDB should run on solaris 2.5 again. I don't think it would since
CGDB .6.0.
* Fixed a bug in the configure script that was probably added in .6.0
when the large readline integration was done. It was possible for
readline to choose the wrong termcap library, because of a bad
autoconf macro that CGDB was using. This could potentially cause the
screen to get corrupted.
* Fixed a bug in the syntax highlighter for C strings. A line like
char *s = "\""; would cause the highlighter to break and most of the
file after that would be displayed in red (string literal color).
* The message 'CGDB had unexpected results, ...' will no longer be
displayed when CGDB can not parse a breakpoint. This message was being
displayed when a watchpoint was issued by the user. It does not fix any
bugs though, just stops the message from being displayed.
cgdb-0.6.1 (03/06/2006)
* An improvement for tab completion support. With certain versions of GDB,
an extra line "server complete ..." would be added to the list of
possible completions. This is no longer added. I believe this would
only happen with really old versions of GDB.
* Improved tab completion support in readline for quoted data. For
instance tab completing C++ member functions like b 'foo::b<tab>
will now produce b 'foo::bar()'.
* Improved tab completion support. This comes in 2 forms. First, CGDB
honors the readline variable rl_completion_query_items. If the number
of completions is greater than this variable, then the user will be
asked if they want the completion to happen. Second, after each page of
completions, the user will be asked if they want more completions to be
displayed or not.
* Thanks to Steve Folta for sending in a patch which added support in CGDB
for configuring the syntax colors. The colors used to be hard coded and
could not be changed. Now just use the ":highlight" command to change
the default colors. There is a new section in the manual called
"CGDB highlighting groups" that describes the new functionality.
* Again, thanks to Steve Folta for getting CGDB to run on terminals with
background colors besides black. CGDB used to paint the terminal
background color black regardless of what it was. Now it respects the
background color. This feature only works when linked to ncurses. If you
have curses, CGDB will force the background black.
* CGDB broke it's terminal resizing functionality in cgdb-0.6.0. This is
now working again. The bug was introduced while bringing readline into
the same address space as CGDB.
* The man page now get's installed when 'make install' is done.
cgdb-0.6.0 (01/30/2006)
* There is a new manual for CGDB. The text version is located in
cgdb/doc/cgdb.txt. The info version is available at cgdb/doc/cgdb.info.
The text, html, pdf, info and possibly other versions are avialable at
http://cgdb.sourceforge.net/documentation.shtml
* There is a new CGDB man page. It is located in cgdb/doc/cgdb.1.
* Fix a crash in CGDB if the user would type ':e' or ':edit' in the
source window, and the source window was empty. This crash is considered
not very significant, since I can't understand why the user would try to
refresh the file, if it didn't exist.
* Removed the :q! and :quit! commands, since they don't do anything
different than :q or :quit.
* Fix a bug regarding the CGDB source viewer displaying breakpoints. If you
used the file dialog to open a file, and the file was an absolute
path, CGDB would display the file. If it was the first time CGDB
encountered this file, then when you set a breakpoint, CGDB would not
display the line as having a breakpoint in the source window, even though
the breakpoint was actually being set. This is now fixed.
* CGDB now supports getting the initial line, as well as the initial
file that GDB is on when it first starts the inferior. It used to only
determine the initial file.
* CGDB now supports when the terminal does not have control flow turned
on. If control flow (XON/XOFF) is off, the ^s and ^q will go directly
to readline instead of freezing and restarting the terminal.
* CGDB now allows the user to use the \ character for line continuation
purposes. GDB handles this and now CGDB also handles this. Before this,
CGDB would break if the user attempted to do line continuation.
* CGDB no longer needs to do a fork to communicate data back and forth to
readline, it runs readline in it's own process context. This is a huge
improvement because it means that when you run CGDB you should only see 1
process executing instead of the usual 2.
* CGDB now supports tab completion. This has been a long awaited missing
feature of CGDB.
* CGDB now depends on readline 5.1.
* Fix a bug in the c/c++ syntax highlighter. If a file had "\\" as a
string literal, the CGDB syntax tokenizer would be built incorrectly
resulting in incorrect file highlighting.
cgdb-0.5.3 (08/08/2005)
* CGDB now can detect the version of readline that it is going to link against.
By doing this, it can alert the user at configure time if the readline version
is OK to build CGDB with. If it is not OK, configure will fail.
* CGDB has changed it's user interface in a way that is not backwards compatible.
The 't' key, now sets a temporary breakpoint. The 't' key used to put CGDB into
insert mode in the TTY window. Now 'I' does that.
* CGDB is now capable of setting a temporary breakpoint thanks to Marcel Lanz.
He sent in a patch with the functionality. Hitting 't' set's a temporary breakpoint.
* It was reported that the "Windows" Key would cause CGDB to shut down.
This has been fixed.
* Added a new 'arrowstyle' option, see README for documentation
* Add the ^l command to the Source window. CGDB will now clear and redraw
the screen when this command is received.
* It is now required to have the GNU history library in order
to configure CGDB successfully. This was always the case, but the
configure checks wouldn't catch it.
* Get readline configure check to only look for libreadline. Not libedit, ...
Future version of CGDB will depend on functionality that only readline will
contains.
cgdb-0.5.2 (05/21/2005)
* Attempt to get autoconf/automake machinery working with command,
make DESTDIR=... install. This should simplify the job of package maintainers.
* Added the --with-readline-prefix autoconf option. This makes configuring CGDB
much easier when readline is in a non standard place.
* Fixed major bug in CGDB-0.5.1 that prevented it from running on Solaris 9,
FreeBSD 5.2.1 and probably many others. CGDB would hang when trying to read
input. This bug made it impossible to use CGDB on those platforms. libkui is
the cause of this problem and was a replacement input library for CGDB in
cgdb-0.5.1. The bug is now fixed.
cgdb-0.5.1 (04/02/2005)
* fixed crash when user tries to set a breakpoint on a file that's not
a source file. This would happen if you typed ':help' to get into the
help, and then tried to set a breakpoint using the spacebar.
* obsolete the b and f commands. These were responsible for moving between
the previous and the next source file. They made CGDB unstable and I don't
think they were in the spirit of a vi/vim clone.
* fixed an annoying bug that appeared while searching for a string. If you
canceled the source view regex search by hitting ESC, 2 bad things would happen.
You would end up on the line that you cancelled at
The currently displayed line would stay the line you can cancelled at, even when
you start moving around again.
* reworked the signal handling code when GDB dies. Now CGDB handles the SIGCHLD
and passes the signal notification to TGDB. This fixes a race condition that
occurred, even though the race condition appeared to be harmless.
* Several memory runtime fixes found thanks to valgrind. This could potentially
fix crash's that I just never ran into.
* Fix a bug in the c/c++ syntax highlighter. If a file had a dos file format
and the C++ style comment // was used, then the file displayed would be
completely unreadable. This is because the trailing '\r' would get grouped with
the comment.
* Fix CGDB command line interface. If the user invoked CGDB like this,
'cgdb --args ./main one two three', then CGDB would hang or crash. This
was because it use to add extra command line options to GDB on the end.
Now it prepends the options to GDB, so that the --args command line
can still be honored.
* Added new logger interface. This traps all errors to a file or
a file descriptor. There is now a mechanism to write warnings and
errors to a file for inspection by the user.
* Removed deprecated libinput from the various subdirectory. It has been
replaced by libkui.
* Added 'set asr' or autosourcereload option to CGDB. It is off by default.
It is capable of reloading a source file when CGDB detects that a file has
been modified. To turn this feature on do ':set asr'.
* The :edit [:e] colon command has been implemented. It forces CGDB to
reload the current file in the source viewer.
cgdb-0.5.0 (07/14/2004)
* "progs" directory is no longer installed in a potentially
read-only directory. It is installed now in top_builddir, instead of
top_srcdir.
* Added libkui into CGDB. This is the fisrt of the 4 new libraries
designed to make CGDB flexible and modular. Because of this, the version
number has been bumped up to .5.0.
* Fixed crash in TGDB when buffering commands. If the user hit
'n' <Enter> <Enter> ... fast, while GDB was not responding, or working
on a command, the buffering code in TGDB would enter a bad command into
a queue, causing it to crash when it tries to execute the command.
* Fix crash in CGDB when parsing command line parameters. I think this crash
showed up when I starting compiling CGDB with 3.3 instead of 2.95. It would
crash when trying to pass the --pid argument to GDB through CGDB.
cgdb-0.4.2 (05/10/2004)
* CGDB now compiles again on AIX 4.3. Fixed Makefile.am in various/util/src
to not have a space between the -L and the path.
cgdb-0.4.1 (04/26/2004)
* Removed a memory leak from CGDB. When debugging applications that output
a lot of text, CGDB would reportedly grow well over 400MB. Thanks to the
sender of this patch, who decided to keep themselves anonymous.
* Removed all instances of PATH_MAX and replaced with FSUTIL_PATH_MAX.
PATH_MAX is not defined on HURD, and thus CGDB wouldn't compile in that
environment.
* Fixed TGDB bug. Tab completion was broken during cgdb-0.4.0 release.
== 1.2.11.2 / 2011-09-07
- Remove empty dirents after constructing tree (fixes non-contiguous
allocation table usage).
- Fix fallback Symbol#to_proc to match activesupport definition in case
we get loaded first (github #2, lazylester).
- Use method_defined? for fallback guards to support newer versions of
ruby (jocker).
- Add guard on FileTime#initialize to skip for newer versions of ruby.
Missing required methods, but optimization no longer relevant
anyway (github #4, sagmor).
0.1.10 release date: 2011-08-22
New feature
* Add encoding and unicode_errors option to packer and unpacker. When this
option is specified, (un)packs unicode object instead of bytes. This enables
using msgpack as a replacement of json. (tailhook)
* Various codepaths that invoked zlib deflate/inflate assumed that these
functions can compress or uncompress more than 4GB data in one call on
platforms with 64-bit long, which has been corrected.
* "git unexecutable" reported that "unexecutable" was not found, even
though the actual error was that "unexecutable" was found but did
not have a proper she-bang line to be executed.
* Error exits from $PAGER were silently ignored.
* "git checkout -b <branch>" was confused when attempting to create a
branch whose name ends with "-g" followed by hexadecimal digits,
and refused to work.
* "git checkout -b <branch>" sometimes wrote a bogus reflog entry,
causing later "git checkout -" to fail.
* "git diff --cc" learned to correctly ignore binary files.
* "git diff -c/--cc" mishandled a deletion that resolves a conflict, and
looked in the working tree instead.
* "git fast-export" forgot to quote pathnames with unsafe characters
in its output.
* "git fetch" over smart-http transport used to abort when the
repository was updated between the initial connection and the
subsequent object transfer.
* "git fetch" did not recurse into submodules in subdirectories.
* "git ls-tree" did not error out when asked to show a corrupt tree.
* "git pull" without any argument left an extra whitespace after the
command name in its reflog.
* "git push --quiet" was not really quiet.
* "git rebase -i -p" incorrectly dropped commits from side branches.
* "git reset [<commit>] paths..." did not reset the index entry correctly
for unmerged paths.
* "git submodule add" did not allow a relative repository path when
the superproject did not have any default remote url.
* "git submodule foreach" failed to correctly give the standard input to
the user-supplied command it invoked.
* submodules that the user has never showed interest in by running
"git submodule init" was incorrectly marked as interesting by "git
submodule sync".
* "git submodule update --quiet" was not really quiet.
* "git tag -l <glob>..." did not take multiple glob patterns from the
command line.
require EXPR only accepts Class/Name.pm style module names, not
Class::Name. How frustrating! For that, we provide load_class
'Class::Name'.
It's often useful to test whether a module can be loaded, instead
of throwing an error when it's not available. For that, we provide
try_load_class 'Class::Name'.
Finally, sometimes we need to know whether a particular class has
been loaded. Asking %INC is an option, but that will miss inner
packages and any class for which the filename does not correspond
to the package name. For that, we provide is_class_loaded 'Class::Name'.
* suppress warning in perl < 5.10
Changes 0.10:
* support new style: $pm->start(sub { ... })
* collect zombie processes without any delay when spawn_interval is nonzero
* do not sleep spawn_interval seconds when a worker process exits non-zero and
if err_repawn_interval < spawn_interval
Changes 0.09:
* support for slow startup (with the new spawn_interval parameter) and slow
shutdown (by passing arrayrefs as values of the trap_signals hashref)
* Exclude the example scripts from getting installed.
Changes 0.7.8:
* Make $VERSION compatible with the most perl versions possible
Changes 0.7.7:
* Small distribution fixes
This is a regular bugfix release (slightly early due to travel plans).
commands: clarify that 'hg heads foo' shows heads on branch foo
dispatch: don't rewrap aliases that have the same definition
graphlog: attempt to fix index overrun (issue2912)
http: pass user to readauthforuri() (fix 4a43e23b8c55)
http: strip credentials from urllib2 manager URIs (issue2885)
parsers: avoid pointer aliasing
subrepo: fix cloning of repos from urls without slash after host (issue2970)
ui: pass ' ' to raw_input when prompting
url: really handle urls of the form file:///c:/foo/bar/ correctly
=== 2.12.2 / 2011-08-24
* 2 minor enhancements:
* Added recursive chmod to docs task to ensure group write-ability (kleb)
* Improve error message when no manifest found. (Apparently some linux packagers remove manifest files)
* 3 bug fixes:
* Added missing minitest/autorun require (NOT reported by fedoraproject.org)
* Fix for RSpec support broken in 0dc6b2f (ged)
* Fix for the edgiest of edge cases (-Ku in ruby19... ummm, yeah)
=== 2.2.0 / 2011-08-23
* 2 minor enhancements:
* Moved Keyword, Environment, and StackState inside of RubyParser
* Added proper dsym and dsym->sym support.
* 3 bug fixes:
* Added extra (failing) tests for call/iter line number checking (quix)
* Fixed line numbers for certain call/iter edge cases
* Fixed parsing of: alias :"<<" :">>".
=== 2.12.1 / 2011-08-22
* 2 bug fixes:
* Fixed a test that fails on non-privileged (rvm/multiruby) setups
* RbConfig is used instead of Config to remove a 1.9.3 warning.
Now it requires rubygems 1.8 at least.
=== 4.6.2 / 2011-08-24
* 1 minor enhancement:
* Added Autotest Tips section to Readme
* 1 bug fix:
* Fixed mri 1.9.3 change to Find.find with autotest/restart and missing files.
=== 3.9.3 / 2011-08-23
* Bug fixes
* Add US-ASCII magic comments to work with <tt>ruby -Ku</tt>. Issue #63 by
Travis D. Warlick, Jr.
* Image paths at HTTPS URLs are now turned into +<img>+ tags. Pull
Request #60 by James Mead
* Markup defined by RDoc::Markup#add_special inside a <tt><tt></tt> is no
longer converted.
Experimental version released on August 24th, 2011.
The biggest change in this release is the ability for Kyua to run test
programs implemented using different frameworks. What this means is
that, now, a Kyua test suite can include not only ATF-based test
programs, but also "legacy" (aka plain) test programs that do not use
any framework. I.e. if you have tests that are simple programs that
exit with 0 on success and 1 on failure, you can plug them in into a
Kyua test suite.
Other than this, there have been several user-visible changes. The most
important are the addition of the new 'config' and 'debug' subcommands
to the 'kyua' binary. The former can be used to inspect the runtime
configuration of Kyua after parsing, and the latter is useful to
interact with failing tests cases in order to get more data about the
failure itself.
Without further ado, here comes the itemized list of changes:
* Generalized the run-time engine to support executing test programs
that implement different interfaces. Test programs that use the ATF
libraries are just a special case of this. (Issue 18.)
* Added support to the engine to run "plain" test programs: i.e. test
programs that do not use any framework and report their pass/fail
status as an exit code. This is to simplify the integration of legacy
test programs into a test suite, and also to demonstrate that the
run-time engine is generic enough to support different test
interfaces. (Issue 18.)
* Added the 'debug' subcommand. This command allows end users to tweak
the execution of a specific test case and to poke into the behavior of
its execution. At the moment, all this command allows is to view the
stdout and stderr of the command in real time (which the 'test'
command currently completely hides).
* Added the 'config' subcommand. This command allows the end user to
inspect the current configuration variables after evaluation, without
having to read through configuration files. (Issue 11.)
* Removed the test_suites_var function from configuration files. This
was used to set the value of test-suite-sepecific variables, but it
was ugly-looking. It is now possible to use the more natural syntax
'test_suites.<test-suite-name>.<variable> = <value>'. (Issue 11.)
* Added a mechanism to disable the loading of configuration files
altogether. Needed for testing purposes and for scriptability.
Available by passing the '--config=none' flag.
* Enabled detection of unused parameters and variables in the code and
fixed all warnings. (Issue 23.)
* Changed the behavior of "developer mode". Compiler warnings are now
enabled unconditionally regardless of whether we are in developer mode
or not; developer mode is now only used to perform strict warning
checks and to enable assertions. Additionally, developer mode is now
only automatically enabled when building from the repository, not for
formal releases. (Issue 22.)
* Fixed many build and portability problems to Debian sid with GCC 4.6.3
and Ubuntu 10.04.1 LTS. (Issues 20, 21, 26.)
0.27 August 22, 2011
* Remove MooseX::Role::Parameterized::Meta::Parameter; instead,
the default for parameters of (is => 'ro') is set by the
"parameter" sugar. This removes an unnecessary metaclass and
removes weird edge cases caused by it. Please please please
report any breakage!
0.26 March 20, 2011
* make tests stop relying on a specific format for anon class/role names (doy)
0.25 March 4, 2011
* Move the reinitialization hook from Role to Trait so hardcore hackers
(trait janitors) get the benefit too (doy)
0.24 March 1, 2011
* Made MXRP more extensible by hooking into the role metaobject
reinitialization (doy)
0.23 December 25, 2010
* Merry Christmas!
* Make MooseX::Role::Parameterized::Extending more instructive. Hopefully.
0.22 November 26, 2010
* The test suite now uses Test::Fatal instead of Test::Exception (Karen
Etheridge).
* Fix Test::More dependency (reported by Father Chrysostomos) [rt.cpan.org #63222]
0.21 November 15, 2010
* "package" arg can now tell generate_role to use a specific package (rjbs)
0.20 November 2, 2010
* Minor test refactoring to fix blead support [rt.perl.org #78244]
0.12 Fri, 04 Feb 2010 10:39:00 +0000
* Bump namespace::clean dep to 0.20 to pull in the bugfix for
Package::Stash::XS 0.19
pkgsrc change: handle p5-Class-MOP now being part of p5-Moose
0.106 Sat, 05 Mar 2011 21:50:10 -0600
* stop shipping two copies of the pod tests
0.105 Mon, 27 Sep 2010 15:06:30 -0400
* avoid warnings with latest Moose
pkgsrc change: handle p5-Class-MOP now being part of p5-Moose
Also see Moose::Manual::Delta for more details of, and workarounds
for, noteworthy changes.
2.0202 Tue, Jul 26, 2011
[BUG FIXES]
* Be more consistent about how type constraint messages are handled.
2.0201 Fri, Jul 22, 2011
[BUG FIXES]
* Moose::Util::does_role shouldn't call ->does on things that don't inherit
from Moose::Object.
* Make ->does initialize the metaclass, so that calling it as a class method
on a class which sets up inheritance via some method other than extends
works properly (this fixes an issue with MooseX::Types).
* Make Dist::CheckConflicts a runtime requirement, so moose-outdated always
works.
2.0200 Mon, Jul 18, 2011
[OTHER]
* No changes from 2.0105 (other than a few minor documentation tweaks).
2.0105-TRIAL Mon, Jun 27, 2011
[ENHANCEMENTS]
* Moose::Util::does_role now respects overridden ->does methods. (doy)
2.0104-TRIAL Mon, Jun 20, 2011
[OTHER]
* Include changes from 2.0010.
2.0103-TRIAL Mon, Jun 20, 2011
[DEPRECATIONS]
* Several things that have been deprecated for a while have been removed. See
the 2.0000 section in Moose::Manual::Delta for details.
[NEW FEATURES]
* New Moose::Util::TypeConstraints::union function for creating union type
constraints without having to rely on the string type constraint parsing.
This also allows for creating unions of anonymous type constraints.
(kentnl)
[OTHER]
* Include changes from Moose 2.0009.
2.0102-TRIAL Sat, Jun 18, 2011
[ENHANCEMENTS]
* The native Array trait now has a 'first_index' method, which works just
like the version in List::MoreUtils. (Karen Etheridge)
* Clean up some internal code to help out extensions.
[OTHER]
* Include changes from Moose 2.0008.
2.0101-TRIAL Mon, Jun 06, 2011
[OTHER]
* Various packaging issues.
2.0100-TRIAL Mon, Jun 06, 2011
[DEPRECATIONS]
* Using a hand-optimized type constraint is now deprecated. In keeping with
our release policy, this won't actually start warning until the 2.0200
release.
[NEW FEATURES]
* Type constraints can now provide inlined versions, which should make
inlined code which uses type constraints (such as accessors) faster. This
replaces the existing hand-optimized constraint feature. (Dave Rolsky)
[ENHANCEMENTS]
* Remove a lot of cases where generated methods closed over meta objects.
Most simple cases should now only close over simple data types and
coderefs. This should make deparsing simpler.
2.0010 Mon, Jun 20, 2011
[BUG FIXES]
* Fix regression in 2.0009 and 2.0103 when applying roles during init_meta in
an exporter that also re-exports Moose or Moose::Role. (t0m, ilmari)
2.0009 Sun, Jun 19, 2011
[BUG FIXES]
* duck_type type constraints now report reasonable errors when given
something which isn't an instance of an object. (t0m)
* Moose::Util::apply_all_roles now works even if the applicant is a non-Moose
class. (perigrin)
* When an object is reblessed, triggers are called on attributes that are
set during the reblessing. (Karen Etheridge).
[OTHER]
* Better error message if Moose->init_meta is called with a 'metaclass'
option when that metaclass hasn't been loaded. (jasonmay)
2.0008 Thu, Jun 16, 2011
[BUG FIXES]
* The 'accessor' native delegation for hashrefs now allows setting the value
to undef. (sugoik, doy)
[ENHANCEMENTS]
* Various generated methods have more useful context information. (doy)
2.0007 Sun, May 15, 2011
[BUG FIXES]
* Make sure weak attributes remain weak when cloning. (doy, rafl)
2.0006 Mon, May 09, 2011
[BUG FIXES]
* Revert the List::MoreUtils version bump, as it breaks backwards
compatibility. The dependency will be bumped with Moose 2.0200.
2.0005 Mon, May 09, 2011
[BUG FIXES]
* Only sort the alias keys when determining caching.
2.0004 Mon, May 09, 2011
[BUG FIXES]
* Bump the List::MoreUtils dep to avoid buggy behavior in old versions.
* Sort the list of roles and the alias and excludes parameters when
determining caching, since their order doesn't matter.
2.0003 Mon, May 09, 2011
[BUG FIXES]
* Applying multiple role objects (rather than role names) at once no longer
skips every other role. (rjbs)
* Caching of anon classes now works more sanely in the presence of role
application parameters - alias and excludes options are taken into account,
and caching is disabled entirely if other parameters exist. Asking for
caching (instead of just not weakening) when parameters are given will
begin warning in Moose 2.0200. (doy, autarch)
2.0002 Thu, Apr 28, 2011
[ENHANCEMENTS]
* Provide definition context information for constructors and destructors, so
that they will show up as something other than "generated method (unknown
origin)". Also, add the package that accessors are defined in to their
definition context.
* Use Devel::PartialDump in type constraint error messages, if it is
installed.
[BUG FIXES]
* Stop hiding warnings produced by throwing errors in DEMOLISH methods.
* The 'reset' native delegation for Counter attributes will now also respect
builders (previously, it only respected defaults).
2.0001 Fri, Apr 22, 2011
[ENHANCEMENTS]
* Anonymous classes and roles now have a unified implementation in
Class::MOP::Package. This means that anonymous packages are now also
possible. (Shawn M Moore, doy)
[BUG FIXES]
* No longer call XSLoader from multiple places, as this causes issues on
older perls. (doy, ribasushi)
* Moose::Meta::Role->create now accepts the 'roles' parameter, as it was
documented to. (Chris Weyl)
* Depend on Eval::Closure 0.04, which fixes some issues in mod_perl
environments. (doy, mateu)
2.0000 Mon, Apr 11, 2011
[API CHANGES]
* The RegexpRef type constraint now accepts regular expressions blessed into
other classes, such as those found in pluggable regexp engines.
Additionally the 'Object' constraint no longer rejects objects implemented
as a blessed regular expression. (David Leadbeater)
[OTHER]
* Moose::Manual::Support now explicitly states when major releases are
allowed to happen (January, April, July, or October).
1.9906-TRIAL Mon, Apr 04, 2011
[OTHER]
* Update conflicts list.
* Minor pod updates.
1.9905-TRIAL Mon, Mar 28, 2011
[NEW FEATURES]
* The Moose::Meta::Role::Attribute class now has an original_role method
which returns the role which first defined an attribute. See the docs for
details. (Dave Rolsky)
* Moose::Util::MetaRole will make sure that the class to which you're
applying metaroles or base class roles can actually have them applied. If
not (it's not a Moose class, it has a non-Moose metaclass, etc.), then it
gives a useful error message. Previously, this would just end up dying in
the MetaRole code without a useful message. (Dave Rolsky)
[BUG FIXES]
* When a role had its own applied_attribute metaclass (usually from MetaRole
application), that metaclass would get lost when that role participated in
role composition. It was also lost if that role was consumed by some other
role. Both of these cases have been fixed. Attributes are always applied
with the applied_attribute metaclass of the role which originally defined
them. (Dave Rolsky)
1.9904-TRIAL Fri, Mar 04, 2011
[BUG FIXES]
* Reinitializing anonymous roles used to accidentally clear out the role's
stash in some circumstances. This is now fixed. (doy)
* The Int type constraint now rejects integers with trailing newlines.
(Matthew Horsfall)
1.9903-TRIAL Mon, Feb 28, 2011
[BUG FIXES]
* Reverse the order that Moose::Exporter 'also' exports are dispatched. When
trying to re-export from a package that itself exported a modified set of
Moose sugar, you'd get the original Moose sugar instead of the overrides.
There are also now tests for this. (perigrin)
* Don't initialize lazy attributes with defaults in the constructor (for
immutable classes). (mo)
* When reinitializing meta objects for classes and roles, we failed to
preserve roles and role applications. This led to weird bugs. Many MooseX
modules end up reinitializing your class or role. (Dave Rolsky)
1.9902-TRIAL Mon, Jan 03, 2011
[OTHER]
* Fix generation of CCFLAGS.
* Add a bit more Dist::Zilla functionality.
1.9901-TRIAL Mon, Jan 03, 2011
[OTHER]
* Fix some indexing issues.
* Fix a few issues with the conflict checking stuff.
1.9900-TRIAL Sat, Jan 01, 2011
[OTHER]
* The entire Class::MOP distribution has been merged with Moose. In the
future, the Class::MOP code itself will be merged into Moose, and
eventually the Class::MOP namespace will disappear entirely. For the
current release, we have simply changed how Class::MOP is
distributed. (Dave Rolsky).
* Switched to Dist::Zilla for development. However, we still have a minimal
Makefile.PL in the repository that can be used for development. (Dave
Rolsky)
[API CHANGES]
* Roles now have their own default attribute metaclass to use during
application to a class, rather than just using the class's
attribute_metaclass. This is also overridable via ::MetaRole, with the
applied_attribute key in the role_metaroles hashref (doy).
* The internal code used to generate inlined methods (accessor, constructor,
etc.) has been massively rewritten. MooseX modules that do inlining will
almost certainly need to be updated as well.
[ENHANCEMENTS]
* We now load the roles needed for native delegations only as needed. This
speeds up the compilation time for Moose itself. (doy)
1.25 Fri, Apr 1, 2011
[BUG FIXES]
* Reinitializing anonymous roles used to accidentally clear out the role's
stash in some circumstances. This is now fixed. (doy) (backported from
1.9904)
1.24 Tue, Feb 24, 2011
[BUG FIXES]
* Reverse the order that Moose::Exporter 'also' exports are dispatched. When
trying to re-export from a package that itself exported a modified set of
Moose sugar, you'd get the original Moose sugar instead of the overrides.
There are also now tests for this. (perigrin) (backported from 1.9903)
1.23 Sun, Feb 13, 2011
[PACKAGING FIX]
* The 1.22 release had a bad MANIFEST. This has been fixed.
1.22 Sun, Feb 13, 2011
[BUG FIXES]
* When reinitializing meta objects for classes and roles, we failed to
preserve roles and role applications. This led to weird bugs. Many MooseX
modules end up reinitializing your class or role. (Dave Rolsky) (backported
from 1.9903)
1.21 Wed, Nov 24, 2010
[ENHANCEMENTS]
* The Support manual has been updated to reflect our new major/minor version
policy. (Chris Prather)
* The Contributing manual has been updated to reflect workflow changes based
on this new support policy. (doy)
[BUG FIXES]
* The role attribute metaclass did not inherit from Class::MOP::Object,
which could cause errors when trying to resolve metaclass compatibility
issues. Reported by Daniel Ruoso. (doy)
* The lazy_build feature was accidentally removed from all the docs. Now
it's listed in Moose.pm again. (Chris Prather)
1.20 Fri, Nov 19, 2010
[BUG FIXES]
* When using native delegations, if an array or hash ref member failed a
type constraint check, Moose ended up erroring out with "Can't call method
"get_message" on unblessed reference" instead of generating a useful error
based on the failed type constraint. Reported by t0m. RT #63113. (Dave
Rolsky)
1.19 Tue, Nov 2, 2010
[BUG FIXES]
* There was still one place in the code trying to load Test::Exception
instead of Test::Fatal. (Karen Etheridge)
1.18 Sun, Oct 31, 2010
[ENHANCEMENTS]
* Type constraint objects now have an assert_coerce method which will either
return a valid value or throw an error. (rjbs)
* We now warn when an accessor for one attribute overwrites an accessor for
another attribute. RT #57510. (Dave Rolsky)
[BUG FIXES]
* The native Array and Hash delegation methods now coerce individual new
members if the _member type_ has a coercion. In other words, if the array
reference is defined as an ArrayRef[DateTime], and you've defined a
coercion from Int to DateTime, then pushing an integer via a delegation
method will coerce the integer to a DateTime object. Reported by Karen
Etheridge. RT #62351. (Dave Rolsky)
* An attribute using native delegations did not always properly coerce and
type check a lazily set default value. (doy and Dave Rolsky)
* Using a regexp to define delegations for a class which was not yet loaded
did not actually work, but did not explicitly fail. However, it caused an
error when the class was loaded later. Reported by Max Kanat-Alexander. RT
#60596. (Dave Rolsky)
* Attempting to delegate to a class or role which is not yet loaded will now
throw an explicit error. (Dave Rolsky)
* Attempting to set lazy_build in an inherited attribute was ignored. RT
#62057. (perigrin)
[OTHER]
* The Moose test suite now uses Test::Fatal instead of
Test::Exception. (rjbs)
1.17 Tue, Oct 19, 2010
[BUG FIXES]
* Make native delegation inlining work with instance metaclasses where slot
access is an do {} block, like Kioku. This fixes the use of native
delegations together with Kioku. (Scott, doy)
1.16 Mon, Oct 18, 2010
[ENHANCEMENTS]
* Almost every native delegation method which changes the attribute value
now has an explicitly documented return value. In general, this return
value matches what Perl would return for the same operation. (Dave Rolsky)
* Lots of work on native delegation documentation, including documenting
what arguments each native delegation method allows or requires. (Dave
Rolsky)
* Passing an odd number of args to ->new() now gives a more useful warning
than Perl's builtin warning. Suggested by Sir Robert Burbridge. (Dave
Rolsky)
* Allow disabling stack traces by setting an environment variable. See
Moose::Error::Default for details. This feature is considered
experimental, and may change in a future release. (Marcus Ramberg)
* The deprecation warning for using alias and excludes without a leading
dash now tells you the role being applied and what it was being applied
to. (mst).
[BUG FIXES]
* A number of native trait methods which expected strings as arguments did
not allow the empty string. This included Array->join, String->match,
String->replace, and String->substr. Reported by Whitney Jackson. RT
#61962. (Dave Rolsky)
* 'no Moose' no longer inadvertently removes imports it didn't create
itself. RT #60013. (Florian Ragwitz, doy)
* Roles now support passing an array reference of method names to method
modifier sugar functions. (doy)
* Native traits no longer use optimized inlining routines if the instance
requests it (in particular, if inline_get_slot_value doesn't return
something that can be assigned to). This should fix issues with
KiokuDB::Class. (doy)
* We now ignore all Class::MOP and Moose classes when determining what
package called a deprecated feature. This should make the deprecation
warnings saner, and make it possible to turn them off more easily. (Dave
Rolsky)
* The deprecated "default is" warning no longer happens if the attribute has
any accessor method defined (accessor, reader, writer). Also, this warning
only happens when a method that was generated because of the "default is"
gets called, rather than when the attribute is defined. (Dave Rolsky)
* The "default default" code for some native delegations no longer issues a
deprecation warning when the attribute is required or has a builder. (Dave
Rolsky)
* Setting a "default default" caused a fatal error if you used the builder
or lazy_build options for the attribute. Reported by Kent Fredric. RT
#59613. (Dave Rolsky)
1.15 Tue, Oct 5, 2010
[API CHANGES]
* Major changes to Native Traits, most of which make them act more like
"normal" attributes. This should be mostly compatible with existing code,
but see Moose::Manual::Delta for details.
* A few native traits (String, Counter, Bool) provide default values of "is"
and "default" when you created an attribute. Allowing them to provide
these values is now deprecated. Supply the value yourself when creating
the attribute.
* New option 'trait_aliases' for Moose::Exporter, which will allow you to
generate non-global aliases for your traits (and allow your users to
rename the aliases, etc). (doy)
* 'use Moose' and 'use Moose::Role' now accept a '-meta_name' option, to
determine which name to install the 'meta' name under. Passing 'undef'
to this option will suppress generation of the meta method entirely. (doy)
* Moose now warns if it overwrites an existing method named "meta" in your
class when you "use Moose". (doy)
[ENHANCEMENTS]
* Native Trait delegations are now all generated as inline code. This should
be much faster than the previous method of delegation. In the best case,
native trait methods will be very highly optimized.
* Reinitializing a metaclass no longer removes the existing method and
attribute objects (it instead fixes them so they are correct for the
reinitialized metaclass). This should make the order of loading many
MooseX modules less of an issue. (doy)
* The Moose::Manual docs have been revised and updated. (Dave Rolsky)
[BUG FIXES]
* If an attribute was weak, setting it to a non-ref value after the object
was constructed caused an error. Now we only call weaken when the new
value is a reference.
* t/040_type_constraints/036_match_type_operator.t failed on 5.13.5+. Fixed
based on a patch from Andreas Koenig.
0.33 Thu 4 Aug 2011
- Updated can_xs to fix a bug in it
0.32 Fri May 20 2011
- Production release, no other changes
0.31_02 Mon 21 Mar 2011
- More accurate detection of XS support (ADAMK)
0.31_01 Mon 21 Mar 2011
- Updating copyright year (ADAMK)
- Teak documentation of all() and none() (WYANT)
- Memory leak fixed for apply() and XS version restored (ARC)
- Memory leak fixed for indexes() and XS version restored (ARC)
- Memory leak fixed for part() and XS version restored (ARC)
0.30 Thu 16 Dec 2010
- Change the way we localise PERL_DL_NONLAZY to false to remove
a warning that some people were seeing. The new approach is taken
from the way that List::Util does it.
0.29 Wed 8 Dec 2010
- Removed an erroneous Test::NoWarnings dependency
0.28 Tue 7 Dec 2010
- Switching to a production release
- Restored the regression test for RT #38630 from 0.23.
As apply() was disabled in 0.27_04 this test will only act
to validate the future XS restoration of apply().
- Adding uniq warning tests, disabled initially
0.27_04 Mon 6 Dec 2010
- The behaviour of any/all/none/notall has changed when
passed a null list to treat a null list as a legitimate list.
Instead of returning C<undef> the functions now return the
following: any {} == false, all {} == true, none {} == true,
notall {} == false.
Resolves#40905: Returning undef when none is passed an empty
- Disabled the leaking XS versions of part(), apply() and indexes()
0.27_03 Mon 6 Dec 2010
- General house cleaning
0.27_02 Wed 1 Dec 2010
- Reduced test suite peak memory consumption by 5-10 meg
- Added the 'distinct' alias for the uniq function, for people that
like their chained map/grep/sort pipelines with a SQL'ish flavour.
- Expanded test suite for the any() group of functions.
- The any() group of functions now strictly always return scalar
boolean true, false and undef to match the XS version.
0.27_01 Wed 1 Dec 2010
- Refactored the split test scripts into a common test module
to be shared between both the Perl and XS versions.
- Reapply fix for http://rt.cpan.org/Ticket/Display.html?id=39847
"minmax error: unpredictable results with lists of 1 element"
0.26 Tue 23 Nov 2010
- No changes
- Some parts of the CPAN cloud were confusing my 0.24
release with the older deleted 0.24. Bumping version
past Tassilo's to clarify things.
0.11 2011-06-19
- Allow an empty hash for the -deprecations parameter.
0.10 2010-10-25
- The test suite now uses Test::Fatal instead of Test::Exception. (Karen
Etheridge)
0.09 2010-10-17
- Added a compilation test, because otherwise all test files could not end up
doing skip_all, which may make smokers and test harnesses unhappy.
0.08 2010-10-15
- Include Test::Requires in prereq list. Reported by Todd Rinaldo. RT #62173.
0.07 2010-10-15
- The use of regular expressions in ignores didn't really work in 0.06.
- Added missing dep on List::MoreUtils.
- Replaced Test::Warn with Test::Output in the tests, and made the tests
actually test what I think they should be testing.
0.06 2010-10-14
- The -ignore parameter now accepts regular expressions as well as package
names.
0.05 2010-10-14
- Fixed what looked like a bug in -ignore handling, although I couldn't seem
to write a test that triggered it.
- Removed hard dep on Test::Warn for the benefit of Moose.
String eval is often used for dynamic code generation. For instance,
Moose uses it heavily, to generate inlined versions of accessors
and constructors, which speeds code up at runtime by a significant
amount. String eval is not without its issues however - it's
difficult to control the scope it's used in (which determines which
variables are in scope inside the eval), and it can be quite slow,
especially if doing a large number of evals.
This module attempts to solve both of those problems. It provides
an eval_closure function, which evals a string in a clean environment,
other than a fixed list of specified variables. It also caches the
result of the eval, so that doing repeated evals of the same source,
even with a different environment, will be much faster (but note
that the description is part of the string to be evaled, so it must
also be the same (or non-existent) if caching is to work properly).
Release 8.13 16-Aug-2011
------------------------
This is mainly a bug-fix release. There has been a lot of internal refactoring.
The Unicode tables have been updated. The only new feature in the library is
the passing of *MARK information to callouts. Some additions have been made to
pcretest to make testing easier and more comprehensive. There is a new option
for pcregrep to adjust its internal buffer size.
fork(2) calls slow down as the parent process uses more memory due to
the need to copy page tables. In many common uses of fork(), where it
is followed by one of the exec family of functions to spawn child
processes (Kernel#system, IO::popen, Process::spawn, etc.), it's
possible to remove this overhead by using the use of special process
spawning interfaces (posix_spawn(), vfork(), etc.)
The posix-spawn library aims to implement a subset of the Ruby 1.9
Process::spawn interface in a way that takes advantage of fast process
spawning interfaces when available and provides sane fallbacks on
systems that do not.
Major changes include:
The address bar now highlights the domain of the website you're visiting
Streamlined the look of the site identity block
Added support for the latest draft version of WebSockets with a prefixed API
Added support for EventSource / server-sent events
Added support for window.matchMedia
Added Scratchpad, an interactive JavaScript prototyping environment
Added a new Web Developer menu item and moved development-related items into it
Improved usability of the Web Console
Improved the discoverability of Firefox Sync
Reduced browser startup time when using Panorama
Fixed several stability issues
Fixed several security issues
[0.21]
- When using the tie() fallback ensure we do not obliterate a
foreign tie()
- Better document how to disable the tie() fallback
[0.20_01] (the "mst made me do it" release)
- Only invoke the deleted sub stashing if we run udner a debugger
(avoid runtime penalty of Sub::Name/Sub::Identify)
- Spellfixes (RT#54388)
- When B::Hooks::EndOfScope is not available, switch to a simple
tie() of %^H. While it can not 100% replace B::H::EOS, it does
everything n::c needs
[0.20]
- Bump Package::Stash dependency to 0.22 to pull in a bugfix in
Package::Stash::XS 0.19.
[0.19]
- Port to the new Package::Stash 0.18 API and depend on it.
- Don't rely on package::stash's remove_package_symbol implementation
(doy).
2.3.14:
Security fix with Ruby 1.8.x not affected Ruby 1.9.x:
Ruby on Rails has provided a high performance replacement for
ERB::Util.h since version 2.0.0. Due to a bug in the Ruby 1.8 Regular
Expression code this replacement version will fail to escape certain
malformed unicode strings. This malformed output will then be
interpreted as HTML by some browsers on some operating systems.
Major changes in this release:
* Added sequential version numbering
* Added a optional configure script - the Makefile still works
for most systems.
* Improvements to the "annotate" algorithm: only search
primary ancestors and ignore branches.
* Update the "scrub" command to remove traces of login-groups
and subrepositories.
* Added the --type option to the "fossil tag find" command.
* In contexts where only a check-in makes sense, resolve
branch and tag names to checkins only, never events or other
artifacts.
* Improved display of file renames on a diff. A rebuild is
required to take full advantage of this change.
* Update the built-in SQLite to version 3.7.7.
2.3.2 2011-08-15
----------------
A bug fix release.
Improvements
------------
[ui][console] Added some newlines to improve readability.
Fixes
-----
[ui][console] Worked --verbose again.
Re-supported Ruby 1.8.6. [Reported by James Mead]
Thanks
------
James Mead
=== 2.1.0 / 2011-08-15
* 2 minor enhancements:
* Added new accessor canonicalize_conditions to toggle conditional canonicalization (on by default). (confused)
* Awesome cleanup: Replaced call to append_block by block_append. (Confusion)
* 2 bug fixes:
* Fixed handling last line of =begin/=end. (raybaxter)
* Fixed source line numbers after heredocs. (jbarreneche)
=== 3.0.6 / 2011-08-16
* 5 minor enhancements:
* Added ParseTreeTestCase#add_19tests
* Added a bunch of FAILING 1.9 tests stolen from presidentbeef's repo! yay!
* Added add_19tests and add_18tests to segregate version specific parsing.
* Moved pt_testcase.rb to sexp_processor
* Segregated a bunch of 1.8 specific tests using add_18tests.
=== 2.12.0 / 2011-08-15
* 2 minor enhancements:
* Forcibly copy file specified in --main to index.html. (ewong)
* need_tar now defaults to false.
* 4 bug fixes:
* Don't attempt to require RSpec 1 in an RSpec 2 environment. (monde)
* Fixed test/packaging issue when wrapping hoespec in namespace. (damphyr)
* Fixed tests for jruby. (sferik)
* Hoe.load_plugins removes bad entries so plugin? returns false for unloaded
plugins.
