pkgsrc changes:
- Update MASTER_SITES to MASTER_SITE_PYPI
- Update HOMEPAGE
- Get rid of PYTHON_VERSIONS_INCOMPATIBLE
(beets is now compatible with Python 3 too)
- Misc cosmetic fixes and simplifications
Changes:
1.4.3 (January 9, 2017)
-----------------------
Happy new year! This new version includes a cornucopia of new features from
contributors, including new tags related to classical music and a new
:doc:`/plugins/absubmit` for performing acoustic analysis on your music. The
:doc:`/plugins/random` has a new mode that lets you generate time-limited
music---for example, you might generate a random playlist that lasts the
perfect length for your walk to work. We also access as many Web services as
possible over secure connections now---HTTPS everywhere!
1.4.2 (December 16, 2016)
-------------------------
This is just a little bug fix release. With 1.4.2, we're also confident enough
to recommend that anyone who's interested give Python 3 a try: bugs may still
lurk, but we've deemed things safe enough for broad adoption. If you can,
please install beets with ``pip3`` instead of ``pip2`` this time and let us
know how it goes!
1.4.1 (November 25, 2016)
-------------------------
Version 1.4 has **alpha-level** Python 3 support. Thanks to the heroic efforts
of :user:`jrobeson`, beets should run both under Python 2.7, as before, and
now under Python 3.4 and above. The support is still new: it undoubtedly
contains bugs, so it may replace all your music with Limp Bizkit---but if
you're brave and you have backups, please try installing on Python 3. Let us
know how it goes.
1.3.19 (June 25, 2016)
----------------------
This is primarily a bug fix release: it cleans up a couple of regressions that
appeared in the last version. But it also features the triumphant return of the
:doc:`/plugins/beatport` and a modernized :doc:`/plugins/bpd`.
It's also the first version where beets passes all its tests on Windows! May
this herald a new age of cross-platform reliability for beets.
1.3.18 (May 31, 2016)
---------------------
This update adds a new :doc:`/plugins/hook` that lets you integrate beets with
command-line tools and an :doc:`/plugins/export` that can dump data from the
beets database as JSON. You can also automatically translate lyrics using a
machine translation service.
The ``echonest`` plugin has been removed in this version because the API it
used is `shutting down`_. You might want to try the
:doc:`/plugins/acousticbrainz` instead.
.. _shutting down: https://developer.spotify.com/news-stories/2016/03/29/api-improvements-update/
1.3.17 (February 7, 2016)
-------------------------
This release introduces one new plugin to fetch audio information from the
`AcousticBrainz`_ project and another plugin to make it easier to submit your
handcrafted metadata back to MusicBrainz.
The importer also gained two oft-requested features: a way to skip the initial
search process by specifying an ID ahead of time, and a way to *manually*
provide metadata in the middle of the import process (via the
:doc:`/plugins/edit`).
Also, as of this release, the beets project has some new Internet homes! Our
new domain name is `beets.io`_, and we have a shiny new GitHub organization:
`beetbox`_.
1.3.16 (December 28, 2015)
--------------------------
The big news in this release is a new :doc:`interactive editor plugin
</plugins/edit>`. It's really nifty: you can now change your music's metadata
by making changes in a visual text editor, which can sometimes be far more
efficient than the built-in :ref:`modify-cmd` command. No more carefully
retyping the same artist name with slight capitalization changes.
This version also adds an oft-requested "not" operator to beets' queries, so
you can exclude music from any operation. It also brings friendlier formatting
(and querying!) of song durations.
1.3.15 (October 17, 2015)
-------------------------
This release adds a new plugin for checking file quality and a new source for
lyrics. The larger features are:
* A new :doc:`/plugins/badfiles` helps you scan for corruption in your music
collection. Thanks to :user:`fxthomas`. 🐛`1568`
* :doc:`/plugins/lyrics`: You can now fetch lyrics from Genius.com.
Thanks to :user:`sadatay`. 🐛`1626` 🐛`1639`
* :doc:`/plugins/zero`: The plugin can now use a "whitelist" policy as an
alternative to the (default) "blacklist" mode. Thanks to :user:`adkow`.
🐛`1621` 🐛`1641`
1.3.14 (August 2, 2015)
-----------------------
This is mainly a bugfix release, but we also have a nifty new plugin for
`ipfs`_ and a bunch of new configuration options.
1.3.13 (April 24, 2015)
-----------------------
This is a tiny bug-fix release. It copes with a dependency upgrade that broke
beets.
1.3.12 (April 18, 2015)
-----------------------
This little update makes queries more powerful, sorts music more
intelligently, and removes a performance bottleneck. There's an experimental
new plugin for synchronizing metadata with music players.
Packagers should also note a new dependency in this version: the `Jellyfish`_
Python library makes our text comparisons (a big part of the auto-tagging
process) go much faster.
1.3.11 (April 5, 2015)
----------------------
In this release, we refactored the logging system to be more flexible and more
useful. There are more granular levels of verbosity, the output from plugins
should be more consistent, and several kinds of logging bugs should be
impossible in the future.
There are also two new plugins: one for filtering the files you import and an
evolved plugin for using album art as directory thumbnails in file managers.
There's a new source for album art, and the importer now records the source of
match data. This is a particularly huge release---there's lots more below.
There's one big change with this release: **Python 2.6 is no longer
supported**. You'll need Python 2.7. Please trust us when we say this let us
remove a surprising number of ugly hacks throughout the code.
Jellyfish is a python library for doing approximate and phonetic matching of
strings.
Included Algorithms:
- String comparison:
* Levenshtein Distance
* Damerau-Levenshtein Distance
* Jaro Distance
* Jaro-Winkler Distance
* Match Rating Approach Comparison
* Hamming Distance
- Phonetic encoding:
* American Soundex
* Metaphone
* NYSIIS (New York State Identification and Intelligence System)
* Match Rating Codex
* Fixed a security vulnerability in the page redirect field which allowed users
to insert JavaScript code.
* Fixed a security vulnerability where the ``next`` parameter for the toolbar login
was not sanitised and could point to another domain.
- Undef PySlice_GetIndicesEx, see https://bugs.python.org/issue29943
- Fix for file with DPI in EXIF but not metadata, and XResolution is an int rather than tuple
- Docs: Removed broken download counter badge
- Docs: Fixed rst syntax error
New Features
General
- Add automatically updating live preview functionality (C-c C-c l)
via the native eww browser.
- Use autoload to enable markdown-mode in auto-mode-alist for files
with .text, .markdown, and .md extensions.
- Use Travis CI for automated build testing.
Element Insertion and Editing
- ATX heading subtree promotion and demotion via M-S-LEFT, and
M-S-RIGHT.
- ATX heading subtree moving up and down via M-S-UP and
M-S-DOWN.
- Convert inline links to reference links when
markdown-insert-reference-link-dwim is used when the point is at
an inline link.
Customizable Features
- Allow linking to multiple stylesheets in markdown-css-paths
list. Use stylesheets for both preview and export. Previous
markdown-css-path (singular) is now deprecated.
- Customizable default unordered list marker via
markdown-unordered-list-item-prefix.
- Add asymmetric ATX heading adornment option
markdown-asymmetric-header.
Font Lock
- Separate highlighting for Markdown markup characters (asterisks,
underscores, backquotes, etc.) to aid in readability.
- General font lock improvements for comments, code blocks,
blockquotes, headings, horizontal rules, bold, italics
- Font lock for bold, italics, and LaTeX math work inside block
elements such as headings and blockquotes.
- Font lock for <kbd> tags.
GFM
- Support GFM-style code blocks in markdown-mode (as well as
gfm-mode).
- New function markdown-electric-backquote will prompt for a
language name for GFM code blocks. This can be disabled by
customizing the variable markdown-gfm-use-electric-backquote.
- Completion of programming language names for GFM code blocks. A
list of pre-defined languages is included, but this can be
augmented by setting markdown-gfm-additional-languages.
- Strikethrough support in gfm-mode.
- Support for GFM toggling checkboxes mouse-1 or RET. This is
controlled by a new custom variable,
markdown-make-gfm-checkboxes-buttons.
Other Extensions
- Font lock and filling for Pandoc "fancy list", which use # as
the list marker.
- Basic support for filling of definition lists.
- Support Ikiwiki-style search for wiki links that allows links
relative to parent directories. Enable this by setting
markdown-wiki-link-search-parent-directories.
Bug Fixes
- Fix bug in markdown-complete-region/buffer where level-two
Setext headings could be confused with horizontal
rules. Includes a unit test. Thanks to Gunnar Franke for the
report.
- Fix filling when a decimal number appears at column zero, which
could be confused with an ordered list item.
- Fix buffer-wide markup completion.
- Fix font-lock for GFM code blocks without language keywords.
- Improved Setext header insertion to support wide characters.
- Fix expensive paragraph-separate regular expression.
- Make comment-auto-fill-only-comments a buffer-local variable,
which allows for better default filling behavior in cases where
the global variable is non-nil.
- Fix Emacs 23 compatibility by checking for
font-lock-refresh-defaults before calling it.
- Handle reference definitions when filling paragraphs.
- Improve filling of list items with indentation.
- Properly handle footnotes when filling.
- Fix issues with markdown-footnote-kill and related functions.
- Improve font lock for fenced code blocks.
- Avoid avoid overwriting source file when exporting if source
file has .html extension.
- Fix and improve ordered list behavior to preserve digit spacing
and avoid an infinite loop in certain cases. Adjust ordered list
whitespace when marker digit count increases.
- Improve reference definition regular expression to avoid matching
multiple reference links in one line.
- Allow spaces in fenced code language identifiers.
- Improve font lock for preformatted blocks and fenced code blocks.
- Fix out-of-order HTML output.
- Add console-friendly backspace and tab bindings.
- Better treatment of files without extensions for wiki
links. When files have no extensions, don't append a lone
period.
- Call looking-back with two arguments for compatibility with
Emacs 25.1.
- Make (beginning-of-defun -1) go to next title when point is at
beginning of defun.
- Ignore headings in code blocks for font lock.
- Don't highlight wiki links in code blocks.
- Don't move to links in code blocks with C-c C-p and C-c C-n.
- Fix hanging indentation for list items and single-line
preformatted blocks.
- Better rejection of false positives for italics with respect to
other inline elements (inline code and bold).
- Predicate functions should not modify match data.
- Use correct list marker from previous list level when using C-u
M-RET to insert a dedented list item. Prevent an infinite loop
in some cases.
- Reduce lag when scrolling or inserting text into large
files.
- Avoid confusing tramp errors with malformed wiki
links.
while ago, but since then any packages that included both curses and
terminfo ended up with conflicting BUILDLINK_TRANSFORM entries, leaving it
to include ordering to determine whether the builtin or pkgsrc curses was
used. This keeps them in sync, at least by default, though ideally we'd
ensure that at an infrastructure level.
Changelog:
Fixed
Various security fixes
Make form validation errors and date picker panel visible to the user (Bug 1341190)
Changed
The non-standard showDialog argument to window.find is now ignored (Bug 1348409)
Security fixes:
#CVE-2017-5031: Use after free in ANGLE
leading and trailing spaces in server name don’t raise exception anymore - DitContentRule is properly read from the schema - added validator for Active Directory timestamp - Mock strategies raise an exception if a non-bytes value is added to the schema when no offline schema is provided (str and int are automatically converted) - added custom_validators property to Mock strategies - modifying objectClass with bytes values doesn’t raise an exception anymore (but it may fail anyway because of server constraints) - ensure that config sequence parameters are properly set - allow case insensitive attribute and class names in config parameters - added server.schema.is_valid() to check if the schema is available - empty schema properties are set to empty dict() instead of None - schema definitions with traling and leading spaces are now properly parsed and don’t raise an LDAPSchemaError exception anymore - fixed error when flaky servers (OpenLDAP) don’t return the correct response with StartTls
Unfortunately pup doesn't have any man pages. However, README.md is
a good introduction to it so install it in `share/doc/pup' so that binary
package users can easily access it.
Bump PKGREVISION.
Discussed with <fhajny>
Allowed disabling server-side cursors on PostgreSQL
Bugfixes:
Made migrations respect Index’s name argument. If you created a named index with Django 1.11, makemigrations will create a migration to recreate the index with the correct name.
Fixed a crash when using a __icontains lookup on a ArrayField.
Fixed a crash when using a two-tuple in EmailMessage’s attachments argument.
Fixed QuerySet.filter() crash when it references the name of a OneToOneField primary key.
Fixed empty POST data table appearing instead of “No POST data” in HTML debug page.
Restored BoundFields without any choices evaluating to True.
Prevented SessionBase.cycle_key() from losing session data if _session_cache isn’t populated.
Fixed layout of ReadOnlyPasswordHashWidget (used in the admin’s user change page).
Allowed prefetch calls on managers with custom ModelIterable subclasses.
Fixed change password link in the contrib.auth admin for el, es_MX, and pt translations.
Restored the output of the class attribute in the <ul> of widgets that use the multiple_input.html template. This fixes ModelAdmin.radio_fields with admin.HORIZONTAL.
Fixed crash in BaseGeometryWidget.subwidgets().
Fixed exception reraising in ORM query execution when cursor.execute() fails and the subsequent cursor.close() also fails.
Fixed a regression where CheckboxSelectMultiple, NullBooleanSelect, RadioSelect, SelectMultiple, and Select localized option values.
Corrected the stack level of unordered queryset pagination warnings.
Fixed a regression causing incorrect queries for __in subquery lookups when models use ForeignKey.to_field.
Fixed crash when overriding the template of django.views.static.directory_index().
Fixed a regression in formset min_num validation with unchanged forms that have initial data.
Prepared for cx_Oracle 6.0 support.
Updated the contrib.postgres SplitArrayWidget to use template-based widget rendering.
Fixed crash in BaseGeometryWidget.get_context() when overriding existing attrs.
Prevented AddIndex and RemoveIndex from mutating model state.
Prevented migrations from dropping database indexes from Meta.indexes when changing Field.db_index to False.
Fixed a regression in choice ordering in form fields with grouped and non-grouped options.
Fixed crash in BaseInlineFormSet._construct_form() when using save_as_new.
Fixed a regression where Model._state.db wasn’t set correctly on multi-table inheritance parent models after saving a child model.
Corrected the return type of ArrayField(CITextField()) values retrieved from the database.
Fixed QuerySet.prefetch_related() crash when fetching relations in nested Prefetch objects.
Prevented hiding GDAL errors if it’s not installed when using contrib.gis. (It’s a required dependency as of Django 1.11.)
Fixed a regression causing __in lookups on a foreign key to fail when using the foreign key’s parent model as the lookup
