version 2.3.3 (2015.07.07):
* Fixed a localization bug parsing floating point numbers.
* Added Visual Studio 2012 projects.
* Added Visual Studio 2008 projects.
* Removed Visual Studio 2005 and 2003 projects.
* Modified to build against Xerces-C 3.1.3.
version 2.3.2 (2015.07.07):
* XQilla can now be built with Visual Studios 2013.
version 2.3.1 (2015.05.18):
* Modified to build against Xerces-C 3.1.2.
* Added Visual Studio 2010 projects.
* Fixed a bug that could cause incorrect query results when combining
nested expressions and functions.
* Fixed a crash that could happen when XQilla is built with Visual
Studio 2010 or higher.
version 2.3.0 (2011.09.09):
* Added support for XQuery 3.0 version declarations.
* Updated to support XQuery 3.0 higher order functions, including
partial function application, function coercion, and function item
SequenceType checking.
* Implemented XQuery 3.0 FLWOR clause ordering.
* Implemented XQuery 3.0 FLWOR count clause.
* Implemented XQuery 3.0 extended typeswitch syntax.
* Implemented XQuery 3.0 namespace test and constructor.
* Implemented XQuery 3.0 private functions.
* Updated to the latest W3C XQuery Full Text syntax.
* Implemented Unicode word bounding and sentence bounding for XQuery
Full Text tokenization.
* Implemented XQuery Full Text case options.
* Implemented XQuery 1.0 second edition semantics for module import.
* Added the xqilla:explain() function to return the query plan of an
expression.
* Added xqilla:words() and xqilla:sentences() to perform Unicode word
and sentance boundary detection.
* Added xqilla:drop().
* Rewrote a number of the built-in XQuery functions in XQuery itself.
* Implemented constant time indexing into materialized sequences.
* Improved the efficiency of the tail call optimization code.
* Optimized boolean operators.
* Optimized DOMDocument::evaluate() calls.
* Fixed various bugs in static typing.
* Fixed various memory leaks in the DOM 3 API, and regular expressions.
* Fixed a bug that caused a crash when using huge integers.
* Modified to build against Xerces-C 3.1.1
* Added Visual Studio 2010 support.
* Build fix for AIX 5.3.
* Build fix for the clang++ compiler.
* Build fix for GCC 4.6.
libopenmpt 0.3.1 (2017-09-28)
Changelog since libopenmpt 0.2.6774-beta20.
libopenmpt 0.3.1
[Bug] Windows: libopenmpt resource did not compile for release versions.
libopenmpt 0.3.0 (2017-09-27, not released)
[New] New error handling functionality in the C API, which in particular allows distinguishing potentially transient out-of-memory errors from parse errors during module loading.
[New] New API openmpt::module::get_selected_subsong() (C++) and openmpt_module_get_selected_subsong() (C).
[New] Faster file header probing API openmpt::probe_file_header() and openmpt::probe_file_header_get_recommended_size (C++), and openmpt_probe_file_header(), openmpt_probe_file_header_without_filesize(), openmpt_probe_file_header_from_stream() and openmpt_probe_file_header_get_recommended_size() (C).
[New] New API openmpt::could_open_probability() (C++) and openmpt_could_open_probability() (C). This fixes a spelling error in the old 0.2 API.
[New] openmpt123: openmpt123 can now open M3U, M3U8, M3UEXT, M3U8EXT and PLSv2 playlists via the --playlist option.
[New] openmpt123: openmpt123 now supports very fast file header probing via the --probe option.
[New] Libopenmpt now supports building for Windows 10 Universal (Windows Store 8.2) APIs with MSVC, and also for the older Windows Runtime APIs with MinGW-w64.
[New] New API header libopenmpt_ext.h which implements the libopenmpt extension APIs also for the C interface.
[New] The Reverb effect (S99 in S3M/IT/MPTM, and X99 in XM) is now implemented in libopenmpt.
[New] For Amiga modules, a new resampler based on the Amiga’s sound characteristics has been added. It can be activated by passing the render.resampler.emulate_amiga ctl with a value of 1. Non-Amiga modules are not affected by this, and setting the ctl overrides the resampler choice specified by OPENMPT_MODULE_RENDER_INTERPOLATIONFILTER_LENGTH or openmpt::module::RENDER_INTERPOLATIONFILTER_LENGTH. Support for the MOD command E0x (Set LED Filter) is also available when the Amiga resampler is enabled.
[Change] libopenmpt versioning changed and follows the more conventional major.minor.patch as well as the recommendations of the SemVer scheme now. In addition to the SemVer requirements, pre-1.0.0 versions will also honor API and ABI stability in libopenmpt (i.e. libopenmpt ignores SemVer Clause 4).
[Change] The output directories of the MSVC build system were changed to bin/vs2015-shared/x86-64-win7/ (and similar) layout which allows building in the same tree with different compiler versions without overwriting other outputs.
[Change] The emscripten build now exports libopenmpt as ‘libopenmpt’ instead of the default ‘Module’.
[Change] Android: The build system changed. The various Android.mk files have been merged into a single one which can be controlled using command line options.
[Change] The Makefile build system now passes std=c++11 to the compiler by default. Older compilers may still work if you pass STDCXX=c++0x to the make invocation.
[Change] The Makefile option ANCIENT=1 is gone.
[Change] The optional dependencies on libltdl or libdl are gone. They are no longer needed for any functionality.
[Regression] Compiling client code using the C++ API now requires a compiler running in C++11 mode.
[Regression] Support for GCC 4.1, 4.2, 4.3, 4.4, 4.5, 4.6, 4.7 has been removed.
[Regression] Support for Clang 3.0, 3.1, 3.2, 3.3 has been removed.
[Regression] Support for Emscripten versions older than 1.31.0 has been removed.
[Regression] Support for Android NDK versions older than 11 has been removed.
[Regression] Visual Studio 2008, 2010, 2012, 2013 support has been removed.
[Regression] Dynamic run-time loading of libmpg123 is no longer supported. Libmpg123 must be linked at link-time now.
[Regression] xmp-openmpt: xmp-openmpt now requires XMPlay 3.8 or later and compiling xmp-openmpt requires an appropriate XMPlay SDK with XMPIN_FACE >= 4.
[Regression] Support for libmpg123 older than 1.13.0 has been removed.
[Regression] Un4seen unmo3 support has been removed.
[Bug] C++ API: openmpt::exception did not define copy and move constructors or copy and move assignment operators in libopenmpt 0.2. The compiler-generated ones were not adequate though. libopenmpt 0.3 adds the appropriate special member functions. This adds the respective symbol names to the exported ABI, which, depending on the compiler, might or might not have been there in libopenmpt 0.2. The possibly resulting possible ODR violation only affects cases that did crash in the libopenmpt 0.2 API anyway due to memory double-free, and does not cause any further problems in practice for all known platforms and compilers.
[Bug] The C API could crash instead of failing gracefully in out-of-memory situations.
[Bug] The test suite could fail on MacOSX or FreeBSD in non-fatal ways when no locale was active.
[Bug] libopenmpt_stream_callbacks_fd.h and libopenmpt_stream_callbacks_file.h were missing in Windows development packages.
[Bug] libopenmpt on Windows did not properly guard against current working directory DLL injection attacks.
[Bug] localtime() was used to determine the version of Schism Tracker used to save IT and S3M files. This function is not guaranteed to be thread-safe by the standard and is now no longer used.
[Bug] Possible crashes with malformed IT, ITP, AMS, MDL, MED, MPTM, PSM and Startrekker files.
[Bug] Possible hangs with malformed DBM, MPTM and PSM files.
[Bug] Possible hangs with malformed files containing cyclic plugin routings.
[Bug] Excessive loading times with malformed ITP / truncated AMS files.
[Bug] Plugins did not work correctly when changing the sample rate between two render calls.
[Bug] Possible NULL-pointer dereference read during obscure out-of-memory situations while handling exceptions in the C API.
[Bug] libmodplug: libmodplug.pc was wrong.
[Bug] Cross-compiling libopenmpt with autotools for Windows now properly sets -municode and -mconsole as well as all required Windows system libraries.
[Bug] foo_openmpt: Interpolation filter and volume ramping settings were confused in previous versions. This version resets both to the defaults.
[Bug] libmodplug: The CSoundFile::Read function in the emulated libmodplug C++ API returned the wrong value, causing qmmp (and possibly other software) to crash.
Support for SoundTracker Pro II (STP) and Digital Tracker (DTM) modules.
Increased accuracy of the sample position and sample rate to drift less when playing very long samples.
Various playback improvements for IT and XM files.
Channel frequency could wrap around after some excessive portamento / down in some formats since libopenmpt 0.2-beta17.
Playback improvements for S3M files made with Impulse Tracker and Schism Tracker.
ParamEq plugin emulation didn’t do anything at full gain (+15dB).
All standard DMO effects are now also emulated on non-Windows and non-MSVC systems.
Added libopenmpt_stream_callbacks_buffer.h which adds openmpt_stream_callbacks support for in-memory buffers, possibly even only using a truncated prefix view into a bigger file which is useful for probing.
Avoid enabling some ProTracker-specific quirks for MOD files most likely created with ScreamTracker 3.
Tremolo effect only had half the intended strength in MOD files.
Pattern loops ending on the last row a pattern were not executed correctly in S3M files.
Work-around for reading MIDI macros and plugin settings in some malformed IT files written by old UNMO3 versions.
Improve tracker detection in IT format.
Playback fixes for 8-channel MED files
Do not set note volume to 0 on out-of-range offset in XM files.
Better import of some slide commands in SFX files.
Sample 15 in “Crew Generation” by Necros requires short loops at the beginning of the sample to not be ignored. Since we need to ignore them in some (non-ProTracker) modules, we heuristically disable the old loop sanitization behaviour based on the module channel count.
Both normal and percentage offset in PLM files were handled as percentage offset.
MT2 files with instruments that had both sample and plugin assignments were not read correctly.
Some valid FAR files were rejected erroneously.
Support for VBlank timing flag and comment field in PT36 files.
Improved accuracy of vibrato command in DIGI / DBM files.
STM: Add support for “WUZAMOD!” magic bytes and allow some slightly malformed STM files to load which were previously rejected.
Detect whether “hidden” patterns in the order list of SoundTracker modules should be taken into account or not.
Tighten heuristics for rejecting invalid 669, M15, MOD and ICE files and loosen them in other places to allow some valid MOD files to load.
Improvements to seeking: Channel panning was not always updated from instruments / samples when seeking, and out-of-range global volume was not applied correctly in some formats.
seek.sync_samples=1 did not apply PTM reverse offset effect and the volume slide part of combined volume slide + vibrato commands.
If the order list was longer than 256 items and there was a pattern break effect without a position jump on the last pattern of the sequence, it did not jump to the correct restart order.
Makefile has now explicit support for FreeBSD with no special option or configuration required.
openmpt123: Improved section layout in man page.
libmodplug: Added all missing C++ API symbols that are accessible via the public libmodplug header file.
Autotools build system now has options --disable-openmpt123, --disable-tests and --disable-examples which may be desireable when cross-compiling.
Windows binary packages now ship with libmpg123 included.
Update DEPENDS and test dependencies.
-------------------------------------------
version 0.0200 at 2017-09-29 01:03:32 +0000
-------------------------------------------
Change: b0ceabb33df843b11871a21209e5fd4bf1831fb9
Author: Naveed Massjouni <naveed@vt.edu>
Date : 2017-09-28 21:02:24 +0000
v0.0200
Change: 86a1ec89187cdd64286444448e5a3cab788859b8
Author: Naveed Massjouni <naveedm9@gmail.com>
Date : 2017-03-02 10:25:46 +0000
Merge pull request #5 from PerlDancer/sawyer/feature/reentrant-schema
Make schema() possibly reentrant
Change: d319f5adbd6accadde265a294176b3454fc950a0
Author: Sawyer X <xsawyerx@cpan.org>
Date : 2017-02-16 00:52:33 +0000
Add test for feature
Change: 2e6d0056b2342debce14ecfd41f362e2ef9eff81
Author: Sawyer X <xsawyerx@cpan.org>
Date : 2017-02-16 00:52:33 +0000
Add support for schema configuration in schema() call:
What troubles me is that this module is not reentrant. You will
always get the same schema object and you can only configure it once.
Configuring it once is great, but having to argue with the code over
at which point to do it is annoying.
(Example: Issue #12 in Dancer2-Plugin-DBIC.)
This now allows the schema() function to also receive parameters to
use instead of the default configuration. It tries todo it in a
non-intrusive way for the code with minimal code changes and without
altering anything outside this function (like the global
configuration).
Change: 21a44eb576168081e0ec5793671b26b586347e2d
Author: Sawyer X <xsawyerx@cpan.org>
Date : 2017-02-16 00:52:33 +0000
Retrieve the options before fetching the schema:
The previous logic would try to look up the schema and return it, and
if it didn't exist, it would fetch the options and try to create one.
This makes sense.
However, this also means that if we want to create a schema without
rummaging in the cache, you will have to check whether to even check
the cache.
Instead, we simply retrieve the relevant information for creation
before deciding what to do. Then we could either create a new one
with this configuration or return one.
Change: b406d522942e8596affe45f47b679d44d22e1f5a
Author: Sawyer X <xsawyerx@cpan.org>
Date : 2017-02-16 00:52:33 +0000
Refactor: Move creation of schemas outside schema():
The schema() function is in charge of finding the right plugin, but
also creates them. Ths makes it difficult to have logic that allows
creating without saving or retrieval. We just split the creation into
its own function.
I also added comments to make it a bit clearer.
Change: bc90045822e7d3fc0f4e8f0d10a45044bcba9773
Author: Naveed Massjouni <naveedm9@gmail.com>
Date : 2017-02-13 10:04:52 +0000
Merge pull request #4 from PerlDancer/sawyer/fixes
Various fixes and cleanups
Change: bcee651ab7580e899ad906be872c4874f333d156
Author: Sawyer X <xsawyerx@cpan.org>
Date : 2017-02-13 10:56:10 +0000
Revert "Cleanup: Explicitly use DBIx::Class::Schema::Loader:"
This reverts commit 4f80a6f9ef3c994ffda9991d7fe7839307b6abbf.
Change: 9e527d272a25e4be0a64454d8b258211f171211e
Author: Sawyer X <xsawyerx@cpan.org>
Date : 2017-02-10 11:12:43 +0000
Do not add DBIx::Class::Schema::Loader as hard prereq:
Currently DBIx::Class::Schema::Loader is loaded optionally, but
reported as a hard dependency. (See previous commit.)
This makes it recommended, but optional.
Change: 4f80a6f9ef3c994ffda9991d7fe7839307b6abbf
Author: Sawyer X <xsawyerx@cpan.org>
Date : 2017-02-10 11:12:39 +0000
Cleanup: Explicitly use DBIx::Class::Schema::Loader:
In Dancer2-Plugin-DBIC, DBIx::Class::Schema::Loader is mentioned as a
requirement. It's a lazy optional one. It's also hard to spot because
it's being loaded as a string using Module::Load.
Instead, this makes it a bit clearer by using 'require' directly.
It would be better to make this an optional requirement instead of a
default one.
Change: 1b545771aec22a89989c284f950526270ae272b5
Author: Sawyer X <xsawyerx@cpan.org>
Date : 2017-02-10 10:58:09 +0000
Cleanup: Replace die() with croak():
croak() is used in some places but not everywhere, so might as well
report from the caller everywhere.
Change: c6f2031977572bcfd4352e131eb274cbc0be14d0
Author: Sawyer X <xsawyerx@cpan.org>
Date : 2017-02-10 10:56:38 +0000
Cleanup: Simply conditions code
Change: d9042c990068eb816b21396573704012d66c2ead
Author: Sawyer X <xsawyerx@cpan.org>
Date : 2017-02-10 10:56:20 +0000
Cleanup: Fix extra or missing semicolons
Change: 30166bce0c49ca7e979c9bbd82b582850805df7d
Author: Sawyer X <xsawyerx@cpan.org>
Date : 2017-02-10 10:56:20 +0000
Add provides metadata:
The following plugin will add the provides metadata to the
distribution, which is considered a best practice.
Change: 35e6040f94c561ab29811fe1db8164f03492d87f
Author: Sawyer X <xsawyerx@cpan.org>
Date : 2017-02-10 10:56:20 +0000
Provide META.json file in distribution:
The following plugin adds the META.json file. This is considered the
more modern META data format.
1.321 2017-10-01 BOOK
- Various small improvements to the Tutorial (with help from Ron Savage)
- Minor code changes (with help from E. Choroba)
- Minor fix to eg/build-git
Enhance the reduce-depends.awk script to reduce a larger set of
dependencies into a single dependency. The patterns representing
intervals of version numbers (can be open-ended) are of the form:
pkg>lower
pkg>=lower
pkg<upper
pkg<=upper
pkg>lower<upper
pkg>lower<=upper
pkg>=lower<upper
pkg>=lower<=upper
These patterns are now condensed into a single dependency of the
same form. For example, given the following patterns:
pkg>=1.0
pkg>2.0
pkg<3.0
pkg<=4.0
pkg>=2.5<3.5
the reduced pattern becomes:
pkg>=2.5<3.0
Add the test script used to help with refactoring and adding the
new feature to the script.
This is a mostly complete rewrite of the script; change the
license to the standard 2-clause BSD license used by TNF.
2.4.0:
Add support for tornado async transport via gen.coroutine
Check if soap:address is defined in the service port instead of raising an exception
Update packaging (stop using find_packages())
Properly handle None values when rendering complex types
Fix generating signature for empty wsdl messages
Support passing strings to xsd:Time objects
Revision 0.3.6:
* End-of-octets encoding optimized at ASN.1 encoders
* The __getitem__/__setitem__ behavior of Set/Sequence and SetOf/SequenceOf objects aligned with the canonical Mapping and Sequence protocols in part
* Fixed crash in ASN.1 encoder when encoding an explicitly tagged component of a Sequence
1.3.1:
Validate for Python 3.6
1.3.0:
Drop some non-ASCII characters from HISTORY.rst
Added official support for Python 3.5
Removed confusingly placed lock from example
Corrected invalidation cache documentation
Updated to latest Travis-CI environment
3.22.21:
* Bugs fixed:
738893 ComboBox in appears-as-list mode: Finishing scroll with mouse causes...
786613 Adwaita: provide a generic sidebar icon and label styling
787172 EmojiChooser: Improve selection of section buttons
787279 EmojiChooser: Hovered emoji become impossible to see in HighContrast...
787280 HighContrast: Fix regression in vertical separator in titlebar headerbar
787410 entry: fix memory leak
* Translation updates
2.1.4:
Fix broken match function when using rebulk>=0.9.0.
2.1.3:
Add nzb as new container possible value
Add EAC3 as new audio_codec possible value
Add FullHD as new other possible value
Added python 3.6 support
Dropped python 2.6 support
Make container values consistent and always lowercase
Fix –type movie being ignored for movies that starts with numbers
Fix invalid language detection due the common words audio, true and unknown
Fix episode type detection when series name contains year followed by SEE pattern
Version 5.0.2:
Correct handling of import in child thread
Correct handling of “dis” module with Python 3.5.1
Correct handling of “multiprocess.process” module
Correct attempt to assign variable to an empty list
Improved README
Add hook for pythonnet package
Add hook for sqlite3 and improve win32file hook
Add FAQ entry
1.18.0:
- Include additional assets used for distribution packages in the source tarball
- Consider ``[`` and ``]`` as safe characters in path and query components
of URLs, i.e. they are not escaped anymore
- Disable codecov project coverage check
3.31.1:
This release improves the handling of deadlines so that they act better with the shrinking process. This fixes :issue:`892`.
This involves two changes:
1. The deadline is raised during the initial generation and shrinking, and then lowered to the set value for final replay. This restricts our attention to examples which exceed the deadline by a more significant margin, which increases their reliability.
2. When despite the above a test still becomes flaky because it is significantly faster on rerun than it was on its first run, the error message is now more explicit about the nature of this problem, and includes both the initial test run time and the new test run time.
In addition, this release also clarifies the documentation of the deadline setting slightly to be more explicit about where it applies.
v1.11
Support the modern standard types char16_t and char32_t. These work like wchar_t: they represent one unicode character, or when used as charN_t * or charN_t[] they represent a unicode string. The difference with wchar_t is that they have a known, fixed size. They should work at all places that used to work with wchar_t (please report an issue if I missed something). Note that with set_source(), you need to make sure that these types are actually defined by the C source you provide (if used in cdef()).
Support the C99 types float _Complex and double _Complex. Note that libffi doesn’t support them, which means that in the ABI mode you still cannot call C functions that take complex numbers directly as arguments or return type.
Fixed a rare race condition when creating multiple FFI instances from multiple threads. (Note that you aren’t meant to create many FFI instances: in inline mode, you should write ffi = cffi.FFI() at module level just after import cffi; and in out-of-line mode you don’t instantiate FFI explicitly at all.)
Windows: using callbacks can be messy because the CFFI internal error messages show up to stderr—but stderr goes nowhere in many applications. This makes it particularly hard to get started with the embedding mode. (Once you get started, you can at least use @ffi.def_extern(onerror=...) and send the error logs where it makes sense for your application, or record them in log files, and so on.) So what is new in CFFI is that now, on Windows CFFI will try to open a non-modal MessageBox (in addition to sending raw messages to stderr). The MessageBox is only visible if the process stays alive: typically, console applications that crash close immediately, but that is also the situation where stderr should be visible anyway.
Progress on support for callbacks in NetBSD.
Functions returning booleans would in some case still return 0 or 1 instead of False or True. Fixed.
ffi.gc() now takes an optional third parameter, which gives an estimate of the size (in bytes) of the object. So far, this is only used by PyPy, to make the next GC occur more quickly (issue 320). In the future, this might have an effect on CPython too (provided the CPython issue 31105 is addressed).
Add a note to the documentation: the ABI mode gives function objects that are slower to call than the API mode does. For some reason it is often thought to be faster. It is not!
version 2.2.1:
* gpg: Fix formatting of the user id in batch mode key generation
if only "name-email" is given.
* gpgv: Fix annoying "not suitable for" warnings.
* wks: Convey only the newest user id to the provider. This is the
case if different names are used with the same addr-spec.
* wks: Create a complying user id for provider policy mailbox-only.
* wks: Add workaround for posteo.de.
* scd: Fix the use of large ECC keys with an OpenPGP card.
* dirmngr: Use system provided root certificates if no specific HKP
certificates are configured. If build with GNUTLS, this was
already the case.
