AUTOFIX: Makefile.common:49: Replacing "${PKGSRC_COMPILER} == \"sunpro\"" with "${PKGSRC_COMPILER:Msunpro}".
The PKGSRC_COMPILER can be a list of chained compilers, e.g. "ccache
distcc clang". Therefore, comparing it using == or != leads to wrong
results in these cases.
AUTOFIX: Makefile:35: Replacing "${PKGSRC_COMPILER} == \"clang\"" with "${PKGSRC_COMPILER:Mclang}".
The PKGSRC_COMPILER can be a list of chained compilers, e.g. "ccache
distcc clang". Therefore, comparing it using == or != leads to wrong
results in these cases.
Thu Mar 28 2019 Andreas Schneider <asn@cryptomilk.org>
* cmocka version 1.1.5
* Added cmocka_set_skip_filter()
Thu Mar 28 2019 Andreas Schneider <asn@cryptomilk.org>
* cmocka version 1.1.4
* Added assert_float(_not)_equal()
* Added expect_any_always()
* Small bug fixes
(from Unix/NEWS)
============================================================================
Release Notes for cloc version 1.84
https://github.com/AlDanial/cloc
September 22, 2019
New Languages and File Types:
o APL
o Apache Thrift
o Imba
o IPL
o SaltStack
Updates:
o Improved handling of --git --diff
o Add .podspec as a Ruby file extension.
o New switch, --stat, to force a stat check of
directories. This enables traversal of directory
types whose nlink count does not match the count
of subdirectories (examples: CD-ROM, FAT, AFS).
o Force stat on top level directories. This is
companion logic to the new --stat switch.
o Add file extensions defined by the Linguist project.
o New switch, --exclude-content, to exclude files if
any of their lines match a regex.
o New switch, --timeout, to give control over timeouts
on direct counts (as opposed to diff timeouts).
o On Windows, lower batch size on git archive
commands to 100 files (the value is 1,000 on *nix).
o Better identification of Smarty .tpl template files.
Bug Fixes:
o On Windows, use double quotes for git archive entries.
o On Windows, double slash path separators with --by-file
and JSON output.
o On Windows, replace / with \\ in results only when
running with --by-file.
o Fix header line metrics for files/sec and lines/sec
when doing --diff.
o Python: handle cases of /* or */ embedded in strings
(this conflicted with internal logic which replaced
docstring triple quotes with C comments).
o Fix mishandling of --exclude-dir with --follow-link.
o Fix doubly-counted #! line
o Support --hide-rate with csv output.
============================================================================
Release Notes for cloc version 1.82
https://github.com/AlDanial/cloc
May 3, 2019
New Languages and File Types:
o Apex Class
o DIET
o Fennel
o FXML
o Jupyter notebook files (.ipynb)
o Python wheel files (.whl)
o SVG
o reStructuredText
Updates:
o Handle backslashed quotes in strings to improve --strip-str-comments
logic.
o Improve Perl v. Prolog identification.
o New switch --docstring-as-code to count Python docstrings as code.
o Partial support fot Lua nested comments.
o Renamed Skylark to Starlark.
o Modified Java filter to try to handle embedded comments better.
o Recognize groovy as one of the #! languages.
o Add .pyw as a file extension for Python.
o Update comment definitions for Velocity Template Language.
o Override file alignment logic with --diff when only two files are
given.
o Identify TeX v. Visual Basic for .cls file extension.
o Include result summation line with --csv output.
o Support --csv, --json, --md, --xml, --yaml output types for the
--ignored debug file.
o Force IO encoding to allow wider Unicode characters for creating
Windows exe's with PAR::Packer
============================================================================
Release Notes for cloc version 1.80
https://github.com/AlDanial/cloc
Oct. 5, 2018
This is a bug-fix release to solve mis-handling of git inputs for
straight counts in 1.78 (--git --diff works ok in 1.78),
https://github.com/AlDanial/cloc/issues/328.
New Languages:
o ReasonML
Updates:
o New SQL output style option with "--sql-style Named_Columns". This
includes the table "t" column names in each insert statement.
o New option --file-encoding to create all output files using the
given file encoding (for example "--file-encoding
o Add escript as script language for Erlang.
o Improved handling of custom Smarty definitions through
--read-lang-def.
o Better diff logic when comparing two files.
o Faster program start on the Unix-tuned version of cloc when
multiple processors are not used.
============================================================================
Release Notes for cloc version 1.78
https://github.com/AlDanial/cloc
Sept. 7, 2018
New Languages:
o EJS
o Gradle
o Igor Pro
o JSON5
o Nix
o R Markdown
o SparForte
o Xtend
Updates:
o Interpret "--diff-timeout 0" as allowing unlimited time to compute
diffs.
o Add .inl extension for C++
o Improve language rules for Assembly and JCL.
o Better handling of OCaml comments (don't just use Pascal rules).
o Handle unusual file names with --use-sloccount.
o Remove GNU grep specific -P option with --use-sloccount (fall back
to simpler BSD grep options).
o Handle OCaml nested comments.
o Remove debug print statement for counts of Rmd (R markdown) files.
o New switch --include-ext to only count files with the given extension.
o New switch --config to load command line switches from a file.
o New switch --hide-rate to suppress header line (makes the output
deterministic).
o Skip header line when combining diff reports.
o Allow any number (including just one) of diff reports to be summed.
o Exit early if --diff arguments are the same.
o New switch --strip-str-comments to (try to) avoid 'Complex regular
subexpression recursion limit' warning.
o Run cleanly under Perl 5.28 and 5.32 by backslashing braces in regex.
o Suppord Idris block comments.
o Set UTF-8 encoding on XML and XSL files generated by cloc.
o New switch --write-lang-def-incl-dup to include files with extension
collisions (but should be used with care because the resulting language
definition file will be rejected by cloc until all collisions have been
resolved).
Bug Fixes:
o Correctly handle --not-match-d and --follow-links together.
o Fix --git --diff handling of large file sets.
o Remove \$ from git pathspec filename-cleansing regex as git already
handles such filename mangling
o Remove debug die() invocation in make_file_list().
o Fix improperly formatted JSON emitted with --by-file-by-lang.
o Improve accuracy of C#/Smalltalk disambuguator (was biased toward
counting .cs files as Smalltalk).
o Improve accuracy of TypeScript/Qt Linguist disambuguator.
o More robust YAML output by quoting file and language names.
o Handle more unusual file names with --git --diff
Sat Oct 20, 2019: Released Check 0.13.0
based on hash 2b18886a9a9d3bab44917a550d12128ad7e2c197
* configure: optional build documentation
Issue #206 (GitHub)
* missing <unistd.h> in some files
Issue #196 and Issue #186 (GitHub)
* Varous documentation improvements
* END_TEST is now optional, as how START_TEST works has been redone
Issue #158
* Various CMake related changes:
- Support exporting Check to other projects with CMake 3
Issue #185
- Shared library support for Visual Studio
Issue #220
- Fix wrong library filename
Issue #226
- Add support for CMake package registry
Issue #227
- CMake build type can now be debug or release
Issue #228
- Add checkmk to CMake build.
(from docs/changes.txt), sorry for long lines.
Version 3.45 - May 2018
- New support for reading and writing unsigned long long datatypes.
This includes 'implicit datatype conversion' between the unsigned long
long datatype and all the other datatypes.
- Increased the hardcoded NMAXFILES setting for maximum number of
open files from 1000 to 10000.
- Bug fix to fits_calc_binning wrapper function, which wasn't filling
in the returned float variables.
- Fixed a parsing bug for image subsection and column binning range
specifiers that was introduced in v3.44.
Version 3.44 - April 2018
- This release primarily patches security vulnerabilities. We
strongly encourage this upgrade, particularly for those running
CFITSIO in web accessible applications.
In addition, the following enhancements and fixes were made:
- Enhancement to 'template' and 'colfilter' functionality. It is now
possible to delete multiple keywords using wildcard syntax. See
"Column and Keyword Filtering Specification" section of manual for
details.
- histo.c uses double precision internally for all floating point
binning; new double-precision subroutines fits_calc_binningd(),
fits_rebin_wcsd(), and fits_make_histd(); existing
single-precision histogram functions still work but convert values
to double-precision internally.
- new subroutine fits_copy_cols() / ffccls() to copy multiple columns
- Fix in imcompress.c for HCOMPRESS and PLIO compression of unsigned
short integers.
- Fix to fits_insert_card(ffikey). It had wrongly been capitalizing
letters that appeared before an '=' sign on a CONTINUE line.
Version 3.43 - March 2018
The NASA security team requires the following warning to all users of
CFITSIO:
=====
The CFITSIO open source software project contains vulnerabilities
that could allow a remote, unauthenticated attacker to take control
of a server running the CFITSIO software. These vulnerabilities
affect all servers and products running the CFITSIO software.
The CFITSIO team has released software updates to address these
vulnerabilities. There are no workarounds to address these
vulnerabilities. In all cases, the CFITSIO team is recommending an
immediate update to resolve the issues.
=====
- Fixed security vulnerabilities.
- Calls to https driver functions in cfileio.c need to be macro-
protected by the HAVE_NET_SERVICES variable (as are the http and
ftp driver function calls). Otherwise CMake builds on native
Windows will fail since drvrnet.o is left empty.
- Bug fix to ffmvec function. Should be resetting a local colptr
variable after making a call to ffiblk (which can reallocate Ftpr->
tableptr). Originally reported by Willem van Straten.
- Ignore any attempted request to not quantize an image before
compressing it if the image has integer datatype pixels.
- Improved error message construction throughout CFITSIO.
Version 3.42 - August 2017 (Stand-alone release)
- added https support to the collection of drivers handled in cfileio.c
and drvrnet.c. This also handles the case where http transfers are
rerouted to https. Note that this enhancement introduces a dependency
on the libcurl development package. If this package is absent, CFITSIO
will still build but will not have https capability.
- made fix to imcomp_init_table function in imcompress.c. It now writes
ZSIMPLE keyword only to a compressed image that will be placed in the
primary header.
- fix made to fits_get_col_display_width for case of a vector column
of strings.
Version 3.42 - March 2017 (Ftools release only)
- in ftp_open_network and in ftp_file_exist, added code to repeatedly
attempt to make a ftp connection if the ftp server does not respond
to the first request. (some ftp servers don't appear to be 100% reliable).
- in drvrnet.c added many calls to 'fclose' to close unneeded files,
to avoid exceeding the maximum allowed number of files that can be
open at once.
- made substantial changes to the ftp_checkfile and http_checkfile routines
to streamline the process of checking for the existence of a .gz or .Z
compressed version of the file before opening the uncompressed file
(when using http or ftp to open the file).
- modified the code in ftp_open_network to send "\r\n" as end-of-line
characters instead of just "\n". Some ftp servers (in particular,
at heasarc.gsfc.nasa.gov) now require both characters, otherwise the
network connection simply hangs.
- modified the http_open_network routine to handle HTTP 301 or 302 redirects
to a FTP url. This is needed to support the new configuration on
the heasarc HTTP server which sometimes redirects http URLS to a ftp URL.
Version 3.41 - November 2016
- The change made in version 3.40 to include strings.h caused problems on
Windows (and other) platforms, so this change was backed out. The reason
for including it was to define the strcasecmp and strcasencmp functions, so
as an alternative, new equivalent functions called fits_strcasecmp and
fits_strncasecmp have been added to CFITSIO.as a substitute. All the
previous calls to the str[n]casecmp functions have been changed to
now call fits_str[n]casecmp. In addition, the previously defined
ngp_strcasecmp function (in grparser.c) has been removed and the calls to
it have been changed to fits_strcasecmp.
- The speed.c utility program was changed to correctly call
the gettimeofday function with a NULL second arguement.
Version 3.40 - October 2016
- fixed a bug when writing long string keywords with the CONTINUE convention
which caused the CONTINUE'd strings to only be 16 characters long, instead
of using up all the available space in the 80-character header record.
- fixed a missing 'defined' keyword in fitsio.h.
- replaced all calls to strtok (which is not threadsafe) with a new ffstrtok
function which internally calls the threadsafe strtok_r function. One
byproduct of this change is that <strings.h> must also be included
in several of the C source code files.
- modified the ffphbn function in putkey.c to support TFORM specifiers that
use lowercase 'p' (instead of uppercase) when referring to a variable-length
array column.
- modified the lexical parser in eval.y and eval_y.c to support bit array
columns (with TFORMn = 'X') with greater than 256 elements. Fix to bitcmp
function: The internal 'stream' array is now
allocated dynamically rather than statically fixed at size 256.
This was failing when users attempted a row filtering of a bitcol
that was wider than 256X. In bitlgte, bitand, and bitor functions, replaced
static stream[256] array allocation with dynamic allocation.
- modified the ffiter function in putcol.c to fix a problem which could
cause the iterator function to incorrectly deal with null values. This
only affected TLONG type columns in cases where sizeof(long) = 8, as well
as for TLONGLONG type columns.
- Fix made to uncompress2mem function in zcomprss.c for case where output
uncompressed file expands to over the 2^32 (4Gb) limit. It now
checks for this case at the start, and implements a 4Gb paging
system through the output buffer. The problem was specifically
caused by the d_stream.avail_out member being of 4-byte type uInt,
and thus unable to handle any memory position values above 4Gb.
- fixed a bug in fpackutil.c when using the -i2f (integer to float) option
in fpack to compress an integer image that is scaled with non-default values
for BSCALE and BZERO. This required an additional call to ffrhdu to reset
the internal structures that describe the input FITS file.
- modified fits_uncompress_table in imcompress.c to silently ignore the
ZTILELEN keyword value if it larger than the number of rows in the table
- Tweak strcasecmp/strncasecmp ifdefs to exclude 64-bit MINGW
environment, as it does not lack those functions. (eval_l.c,
fitsio2.h)
- CMakeLists.txt: Set M_LIB to "" for MINGW build environment (in
addition to MSVC).
- Makefile.in: Add *.dSYM (non-XCode gcc leftovers on Macs) to
clean list. Install libs by name rather than using a wildcard.
- configure: Fix rpath token usage for XCode vs. non-XCode gcc on Macs.
Version 1.6, 2019-02-23
* New option --all (-A)
Produce graphs for all global functions in the program. Use this
option if your program contains functions which are not directly
reachable from main().
The output consist of separate flow graphs for each global function
defined in the program. These graphs will be placed after the graph
for main() (if it exists), and will be ordered lexicographically by
the function name.
* New option --no-main
This option has the same effect as '--all', except that the graph for
main() function (if it exists) is treated same way as all the other
graphs, i.e. it will not be placed at the top of output, but in its
place as per the lexicographic ordering of function names.
0.8:
Added
Python 3.8 support
Compatibility with Django Debug Toolbar 2.x
Removed
Python 3.3 support
Compatibility with Django Debug Toolbar 1.x
Fixed
Include dicts which aren’t tracked by garbage collector in summary diff
Fix formatting of Python 3 class names in summary diff
* Version 2.3.0 (released 2019-11-14)
** Mitre has assigned CVE-2019-12290 which was fixed by
the roundtrip feature introduced in 2.2.0 (commit 241e8f48)
** Update the data tables from Unicode 6.3.0 to Unicode 11.0
** Turn _idn2_punycode_encode, _idn2_punycode_decode
into compat symbols (Fixes#74)
ccache 3.7.6
Bug fixes
The opt-in “file_macro sloppiness” mode has been removed so that the input file path now is always included in the direct mode hash. This fixes a bug that could result in false cache hits in an edge case when “file_macro sloppiness” is enabled and several identical source files include a relative header file with the same name but in different directories.
Statistics files are no longer lost when the filesystem of the cache is full.
Bail out on too hard Clang option -MJarg (in addition to the previous bailout of -MJ arg).
Properly handle color diagnostics in the depend mode as well.
4.1.3:
Add magics for 3.5.8 and 3.5.9
Python 3.0 tolerance
Fix for unmarshaling Python 3.8 str from 3.2
Pypy 3.3, 3.5 3.6 and 3.6.9 magic numbers and support
improve distribution wheels, and eggs
pytest 5.3.0:
Deprecations
* The default value of junit_family option will change to xunit2 in pytest 6.0, given that this is the version supported by default in modern tools that manipulate this type of file.
In order to smooth the transition, pytest will issue a warning in case the --junitxml option is given in the command line but junit_family is not explicitly configured in pytest.ini.
For more information, see the docs.
Features
* The pytest team has created the pytest-reportlog plugin, which provides a new --report-log=FILE option that writes report logs into a file as the test session executes.
Each line of the report log contains a self contained JSON object corresponding to a testing event, such as a collection or a test result report. The file is guaranteed to be flushed after writing each line, so systems can read and process events in real-time.
The plugin is meant to replace the --resultlog option, which is deprecated and meant to be removed in a future release. If you use --resultlog, please try out pytest-reportlog and provide feedback.
* When sys.pycache_prefix (Python 3.8+) is set, it will be used by pytest to cache test files changed by the assertion rewriting mechanism.
This makes it easier to benefit of cached .pyc files even on file systems without permissions.
* Allow selective auto-indentation of multiline log messages.
Adds command line option --log-auto-indent, config option log_auto_indent and support for per-entry configuration of indentation behavior on calls to logging.log().
Alters the default for auto-indention from on to off. This restores the older behavior that existed prior to v4.6.0. This reversion to earlier behavior was done because it is better to activate new features that may lead to broken tests explicitly rather than implicitly.
* pytester learned two new functions, no_fnmatch_line and no_re_match_line.
The functions are used to ensure the captured text does not match the given pattern.
The previous idiom was to use re.match:
assert re.match(pat, result.stdout.str()) is None
Or the in operator:
assert text in result.stdout.str()
But the new functions produce best output on failure.
* Added tolerances to complex values when printing pytest.approx.
For example, repr(pytest.approx(3+4j)) returns (3+4j) ± 5e-06 ∠ ±180°. This is polar notation indicating a circle around the expected value, with a radius of 5e-06. For approx comparisons to return True, the actual value should fall within this circle.
* Added the pluginmanager as an argument to pytest_addoption so that hooks can be invoked when setting up command line options. This is useful for having one plugin communicate things to another plugin, such as default values or which set of command line options to add.
Improvements
* Use multiple colors with terminal summary statistics.
* Quitting from debuggers is now properly handled in doctest items.
* Improved verbose diff output with sequences.
Before:
After:
* Display untruncated assertion message with -vv.
* Fixed plurality mismatch in test summary (e.g. display “1 error” instead of “1 errors”).
* Config.InvocationParams.args is now always a tuple to better convey that it should be immutable and avoid accidental modifications.
* pytest.main now returns a pytest.ExitCode instance now, except for when custom exit codes are used (where it returns int then still).
* Align prefixes in output of pytester’s LineMatcher.
* Collection errors are reported as errors (and not failures like before) in the terminal’s short test summary.
* pytester.spawn does not skip/xfail tests on FreeBSD anymore unconditionally.
* The “[XXX%]” indicator in the test summary is now colored according to the final (new) multi-colored line’s main color.
* Added --co as a synonym to --collect-only.
* atomicwrites is now only used on Windows, fixing a performance regression with assertion rewriting on Unix.
* Now parametrization will use the __name__ attribute of any object for the id, if present. Previously it would only use __name__ for functions and classes.
* Improved failure reporting with pytester’s Hookrecorder.assertoutcome.
* The reason for a stopped session, e.g. with --maxfail / -x, now gets reported in the test summary.
* Improved cache.set robustness and performance.
Bug Fixes
* Fixed --setup-plan showing inaccurate information about fixture lifetimes.
* Fixed line offset mismatch of skipped tests in terminal summary.
* The PytestDoctestRunner is now properly invalidated when unconfiguring the doctest plugin.
This is important when used with pytester’s runpytest_inprocess.
* BaseExceptions are now handled in saferepr, which includes pytest.fail.Exception etc.
* pytester: fixed order of arguments in rm_rf warning when cleaning up temporary directories, and do not emit warnings for errors with os.open.
* Fixed result of getmodpath method.
Trivial/Internal Changes
* RunResult from pytester now displays the mnemonic of the ret attribute when it is a valid pytest.ExitCode value.
Changes for v1.10.2:
- Include all test directory contents in source distribution
Changes for v1.10.1:
- Trigger completers on –optional=PARTIAL_VALUE
- Complete console scripts installed from wheels
4.28 Tue Nov 19 13:55:39 CET 2019
- (libev) fix ev_port backend, thanks to David H. Gutteridge for
reminding me to actually release the fix.
- (libev) many bugfixes in linuxaio backend.
- (libev) experimental io uring interface.
- try to name ev.h more explicitly, to hopefully improve portability.
- opportunistically round up wait times for poll and epoll backend,
to avoid unnecessary loop iterations.
- add build dependency on ev_linuxaio.c.
- quickly (re)-ported to minix 3.3 before minix crashed again.
Version 6.1
Updated for the macOS 10.15.1 SDK (Xcode 11.2)
Fix reference counting in -[OC_PythonData length], which resulted in use-after-free.
Fix problems found in pyobjc-core by the clang static analyser
Version 6.0.1
Remove debug print accidently left in production
Surpress “-Wunguarded-availability” warnings in the extension AppKit._inlines
Version 6.0
Removed Python 2 support from the C extension in pyobjc-core
Reformatted code in pyobjc-core:
- Use “black” for Python code
- Use “clang-format” for Objective-C code
Updated bindings for macOS 10.15 (Xcode 11.0)
The userspace driver frameworks introduced in macOS 10.15 (DriverKit and related frameworks) will not be exposed through PyObjC. Please let me know if you have a good use case for using these frameworks with Python.
Add new framework wrappers for all other new frameworks in macOS 10.15:
AuthenticationServices
CoreHaptics
CoreMotion
DeviceCheck
ExecutionPolicy
FileProvider
FileProviderUI
LinkPresentation
OSLog
PencilKit
PushKit
QuickLookThumbnailing
Speech
SoundAnalysis
SystemExtensions
Add new framework wrappers for a number of older frameworks:
MetalKit (new in macOS 10.11)
Fix crash when creating NSData objects on macOS 10.15
Remove USE_RAKE. It was used to control dependency to ruby-rake package
befor ruby 2.0 days. Now all ruby*-base package contains rake command and
USE_RAKE is already not used in any where.
pytest 5.2.4:
Bug Fixes
- Fix incorrect discovery of non-test ``__init__.py`` files.
- Revert "The first test in a package (``__init__.py``) marked with ``@pytest.mark.skip`` is now correctly skipped.".
-----------------------------------------
version 1.78 at 2019-11-14 14:18:25 +0000
-----------------------------------------
- Release engineering for 1.78
- No functional changes, removed stray file from the dist. GH #37
-----------------------------------------
version 1.76 at 2019-11-10 21:56:25 +0000
-----------------------------------------
- Release engineering for 1.76
- Fix issues with patching utils/h2ph.PL on v5.8.x
- Whilst I was at it got it working for the v5.9.x series and converted
patch storage to base64
-----------------------------------------
version 1.74 at 2019-11-10 19:12:13 +0000
-----------------------------------------
- Release engineering for 1.74
- fix#35; restore file permission
- Revert "Fix issue where Configure was left not executable"
- This reverts commit 8c5111233d3120fc9e5c11045755a4d96d247848.
- Will apply skaji's solution as it is much more elegant.
-----------------------------------------
version 1.72 at 2019-11-10 17:14:02 +0000
-----------------------------------------
- Release engineering for 1.72
- Fix issue where Configure was left not executable
- Added test prereq Test::More as suggested by CPANTS.
- Signed-off-by: Chris 'BinGOs' Williams <chris@bingosnet.co.uk>
-----------------------------------------
version 1.70 at 2019-11-08 02:01:40 +0000
-----------------------------------------
- Release engineering for 1.70
- Added Base64 encoding of patches
-----------------------------------------
version 1.68 at 2019-11-07 10:11:25 +0000
-----------------------------------------
- Release engineering for 1.68
- Don't index package Errno
-----------------------------------------
version 1.66 at 2019-11-07 00:59:46 +0000
-----------------------------------------
- Release engineering for 1.66
- Fix issues with v5.7.x (h2ph.PL and Errno.pm.PL)
- Fix utils/h2ph.PL patching on v5.8.0
- Set files to be patched to 0644 if they are 0444, set back after.
- Issue 31: In an environment that uses the Lustre file system, it's
been seen that any files that are set to 0444 (read only) will cause
fatal filesystem errors when the patch is attempted. This makes using
perlbrew to install perl on these systems impossible This change
detects read-only files about to be patched, sets them to 0644 for
the duration of the patching, then sets them back to 0444 so that
they exist as expected during the actual build process.
0.42 2019-11-12
[CHANGES]
* Added Future::Queue
[BUGFIXES]
* Remove already-completed futures from the on_cancel chain of others
to avoid the list growing arbitrarily large in some situations; e.g
easily provoked by long loops in Future::AsyncAwait
version 1.29: Fri 8 Nov 09:18:01 CET 2019
Fixes:
- error code changed from libxml2 2.9.9 to 2.9.10
rt.cpan.org#130934 [Petr Pisar]
Improvements:
- skip tests with error messages for Haiku, because they are quite
different. [cpantesters]
7.2.0
-----
* New itertools
* :func:`distinct_combinations`
* :func:`set_partitions` (thanks to kbarrett)
* :func:`filter_except`
* :func:`map_except`
7.1.0
-----
* New itertools
* :func:`ichunked` (thanks davebelais and youtux)
* :func:`only` (thanks jaraco)
* Changes to existing itertools:
* :func:`numeric_range` now supports ranges specified by
``datetime.datetime`` and ``datetime.timedelta`` objects (thanks to MSeifert04 for tests).
* :func:`difference` now supports an *initial* keyword argument.
* Other changes
* Various documentation fixes (thanks raimon49, pylang)
7.0.0
-----
* New itertools:
* :func:`time_limited`
* :func:`partitions` (thanks to rominf and Saluev)
* :func:`substrings_indexes` (thanks to rominf)
* Changes to existing itertools:
* :func:`collapse` now treats ``bytes`` objects the same as ``str`` objects. (thanks to Sweenpet)
The major version update is due to the change in the default behavior of
:func:`collapse`. It now treats ``bytes`` objects the same as ``str`` objects.
This aligns its behavior with :func:`always_iterable`.
.. code-block:: python
>>> from more_itertools import collapse
>>> iterable = [[1, 2], b'345', [6]]
>>> print(list(collapse(iterable)))
[1, 2, b'345', 6]
6.0.0
-----
* Major changes:
* Python 2.7 is no longer supported. The 5.0.0 release will be the last
version targeting Python 2.7.
* All future releases will target the active versions of Python 3.
As of 2019, those are Python 3.4 and above.
* The ``six`` library is no longer a dependency.
* The :func:`accumulate` function is no longer part of this library. You
may import a better version from the standard ``itertools`` module.
* Changes to existing itertools:
* The order of the parameters in :func:`grouper` have changed to match
the latest recipe in the itertools documentation. Use of the old order
will be supported in this release, but emit a ``DeprecationWarning``.
The legacy behavior will be dropped in a future release. (thanks to jaraco)
* :func:`distinct_permutations` was improved (thanks to jferard - see also `permutations with unique values <https://stackoverflow.com/questions/6284396/permutations-with-unique-values>`_ at StackOverflow.)
* An unused parameter was removed from :func:`substrings`. (thanks to pylang)
* Other changes:
* The docs for :func:`unique_everseen` were improved. (thanks to jferard and MSeifert04)
* Several Python 2-isms were removed. (thanks to jaraco, MSeifert04, and hugovk)
Urwid 2.1.0
* Add support for Python 3.7 and 3.8, drop support for Python 3.3
* Add 24-bit (true color) support.
* Add TrioEventLoop
* Add support for input encoding in Terminal widget
* Add ability to specify LineBox title attribute
* Add custom checkbox symbol
* Add installation instruction to README
* Remove PollingListWalker class
* Change SelectableIcon default cursor_position to 0.
* Extended numerical editing: integers and floats
* Re-raise coroutine exceptions in AsyncioEventLoop properly
* Fixed locale issue
* Gate SIGWINCH behind GLib 2.54+
* Remove method Text._calc_line_translation()
* Fix colon in HalfBlock5x4Font
* Don't use deprecated inspect.getargspec() with python3
* Fix issue "Non-integer division in bargraph when using set_bar_width(1)"
* Fix misleading indentation in Screen._stop()
* Fix crash on click-Esc & Esc-click
* Use 'TimerHandle.cancelled()' if available
* Break rather than raising exception on shard calculation bug.
* Increase _idle_emulation_delay.
* Fix EOF detection for the Terminal widget on Python 3
* Fix the asyncio example, and make the raw Screen work without real files
* Unbreak python ./examples/treesample HOME END keys.
* Urwid.util: Fix bug in rle_append_beginning_modify
* Fix AttributeError on mouse click
* Fix ProgressBar smoothing on Python 3.x
* Fix asyncio event loop test on py3.4
* Handle case where MainLoop._topmost_widget does not implement mouse_event
* Implement `ellipsis` wrapping mode for StandardTextLayout
* Fix .pack call in Columns.column_widths
* Use ._selectable member for Edit widget
* Fix use of ignore_focus, for widgets inheriting from Text
* Remove some special handling for TreeListBox
* Make Columns and Pile selectable when any child widget is
* Implement get_cursor_coords for Frame widget
* Fix Frame mouse_event when footer is trimmed
* Fix Python 3.8 SyntaxWarning: 'str' object is not callable
* README: Use SVG build status badge
pytest 5.2.3:
Bug Fixes
- The first test in a package (``__init__.py``) marked with ``@pytest.mark.skip`` is now correctly skipped.
- Fix ``--trace`` when used with parametrized functions.
- Using ``request`` as a parameter name in ``@pytest.mark.parametrize`` now produces a more
user-friendly error.
