| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
0.9.9 - 0.10.0
- Switch to librevenge
- Miscellaneous fixes and cleanups
|
|
|
|
|
|
|
|
|
|
librevenge is a base library for writing document import filters.
It has interfaces for text documents, vector graphics, spreadsheets
and presentations.
|
|
|
|
Overview of changes leading to 0.9.32
Thursday, July 17, 2014
=====================================
- Apply Arabic shaping features in spec order exactly.
- Another fix for Mongolian free variation selectors.
- For non-Arabic scripts in Arabic shaper apply 'rlig' and 'calt'
together.
- Minor adjustment to U+FFFD logic.
- Fix hb-coretext build.
Overview of changes leading to 0.9.31
Wednesday, July 16, 2014
=====================================
- Only accept valid UTF-8/16/32; we missed many cases before.
- Better shaping of invalid UTF-8/16/32. Falls back to
U+FFFD REPLACEMENT CHARACTER now.
- With all changes in this release, the buffer will contain fully
valid Unicode after hb_buffer_add_utf8/16/32 no matter how
broken the input is. This can be overriden though. See below.
- Fix Mongolian Variation Selectors for fonts without GDEF.
- Fix minor invalid buffer access.
- Accept zh-Hant and zh-Hans language tags. hb_ot_tag_to_language()
now uses these instead of private tags.
- Build fixes.
- New API:
* hb_buffer_add_codepoints(). This does what hb_buffer_add_utf32()
used to do, ie. no validity check on the input at all. add_utf32
now replaces invalid Unicode codepoints with the replacement
character (see below).
* hb_buffer_set_replacement_codepoint()
* hb_buffer_get_replacement_codepoint()
Previously, in hb_buffer_add_utf8 and hb_buffer_add_utf16, when
we detected broken input, we replaced that with (hb_codepoint_t)-1.
This has changed to use U+FFFD now, but can be changed using these
new API.
Overview of changes leading to 0.9.30
Wednesday, July 9, 2014
=====================================
- Update to Unicode 7.0.0:
* New scripts Manichaean and Psalter Pahlavi are shaped using
Arabic shaper.
* All the other new scripts to through the generic shaper for
now.
- Minor Indic improvements.
- Fix graphite2 backend cluster mapping [crasher!]
- API changes:
* New HB_SCRIPT_* values for Unicode 7.0 scripts.
* New function hb_ot_layout_language_get_required_feature().
- Build fixes.
|
|
|
|
Version 1.6.6 of PyEnchant has been released, with the following changes:
compatability with latest python releases
"wheels" replace "eggs" as the binary distribution format of choice
fix possible text corruption when using SpellChecker with a Chunker
|
|
|
|
1.4.22
==================================================
- refactor class-level registry on ForkedFunc child start/finish
event to become instance based (i.e. passed into the constructor)
1.4.21
==================================================
- ForkedFunc now has class-level register_on_start/on_exit()
methods to allow adding information in the boxed process.
Thanks Marc Schlaich.
- ForkedFunc in the child opens in "auto-flush" mode for
stdout/stderr so that when a subprocess dies you can see
its output even if it didn't flush itself.
- refactor traceback generation in light of pytest issue 364
(shortening tracebacks). you can now set a new traceback style
on a per-entry basis such that a caller can force entries to be
isplayed as short or long entries.
- win32: py.path.local.sysfind(name) will preferrably return files with
extensions so that if "X" and "X.bat" or "X.exe" is on the PATH,
one of the latter two will be returned.
|
|
|
|
|
|
3.8.2 (2014-07-22)
------------------
- Exclude editor backup files from egg distributions. Fixes #144.
[thet]
|
|
|
|
|
|
|
|
--0.19--
Released June 10, 2014
* Moved issue tracking to github and the main website to mpmath.org.
Several URLs and issue numbers were updated in the documentation
(Sergey B Kirpichev)
* Enabled automatic testing with Travis CI (Sergey B Kirpichev)
* Fixed many doctest issues (Sergey B Kirpichev)
* Converted line endings to LF (Ondrej Certik)
* Made polyroots() more robust (Ondrej Certik)
|
|
|
|
A Python helper module to compare two numbers generously.
|
|
|
|
|
|
|
|
-----
5.4.1
-----
* Python #7776: (ssl_support) Correct usage of host for validation when
tunneling for HTTPS.
---
5.4
---
* Issue #154: ``pkg_resources`` will now cache the zip manifests rather than
re-processing the same file from disk multiple times, but only if the
environment variable ``PKG_RESOURCES_CACHE_ZIP_MANIFESTS`` is set. Clients
that package many modules in the same zip file will see some improvement
in startup time by enabling this feature. This feature is not enabled by
default because it causes a substantial increase in memory usage.
---
5.3
---
* Issue #185: Make svn tagging work on the new style SVN metadata.
Thanks cazabon!
* Prune revision control directories (e.g .svn) from base path
as well as sub-directories.
---
5.2
---
* Added a `Developer Guide
<https://pythonhosted.org/setuptools/developer-guide.html>`_ to the official
documentation.
* Some code refactoring and cleanup was done with no intended behavioral
changes.
* During install_egg_info, the generated lines for namespace package .pth
files are now processed even during a dry run.
---
5.1
---
* Issue #202: Implemented more robust cache invalidation for the ZipImporter,
building on the work in Issue #168. Special thanks to Jurko Gospodnetic and
PJE.
-----
5.0.2
-----
* Issue #220: Restored script templates.
|
|
1.7.3
-----
- Issue #77: Fix import six on Python 3.4 with a custom loader.
- Issue #74: six.moves.xmlrpc_server should map to SimpleXMLRPCServer on Python
2 as documented not xmlrpclib.
|
|
|
|
3.8.1 (2014-07-17)
------------------
- The representation of CaselessDicts in 3.8 changed the name attribute of
Components and therefore broke the external API. This has been fixed.
[untitaker]
3.8 (2014-07-17)
----------------
- Allow dots in property names (Needed for vCard compatibility). Refs #143.
[untitaker]
- Change class representation for CaselessDict objects to always include the
class name or the class' name attribute, if available. Also show
subcomponents for Component objects.
[thet]
- Don't use data_encode for CaselessDict class representation but use dict's
__repr__ method.
[t-8ch]
- Handle parameters with multiple values, which is needed for vCard 3.0.
Refs #142.
[t-8ch]
|
|
|
|
|
|
|
|
|
|
|
|
NEXT (2.6)
-----------------------------------
- Cache exceptions from fixtures according to their scope (issue 467).
- fix issue537: Avoid importing old assertion reinterpretation code by default.
- fix issue364: shorten and enhance tracebacks representation by default.
The new "--tb=auto" option (default) will only display long tracebacks
for the first and last entry. You can get the old behaviour of printing
all entries as long entries with "--tb=long". Also short entries by
default are now printed very similarly to "--tb=native" ones.
- fix issue514: teach assertion reinterpretation about private class attributes
- change -v output to include full node IDs of tests. Users can copy
a node ID from a test run, including line number, and use it as a
positional argument in order to run only a single test.
- fix issue 475: fail early and comprehensible if calling
pytest.raises with wrong exception type.
- fix issue516: tell in getting-started about current dependencies.
- cleanup setup.py a bit and specify supported versions. Thanks Jurko
Gospodnetic for the PR.
- change XPASS colour to yellow rather then red when tests are run
with -v.
- fix issue473: work around mock putting an unbound method into a class
dict when double-patching.
- fix issue498: if a fixture finalizer fails, make sure that
the fixture is still invalidated.
- fix issue453: the result of the pytest_assertrepr_compare hook now gets
it's newlines escaped so that format_exception does not blow up.
- internal new warning system: pytest will now produce warnings when
it detects oddities in your test collection or execution.
Warnings are ultimately sent to a new pytest_logwarning hook which is
currently only implemented by the terminal plugin which displays
warnings in the summary line and shows more details when -rw (report on
warnings) is specified.
- change skips into warnings for test classes with an __init__ and
callables in test modules which look like a test but are not functions.
- fix issue436: improved finding of initial conftest files from command
line arguments by using the result of parse_known_args rather than
the previous flaky heuristics. Thanks Marc Abramowitz for tests
and initial fixing approaches in this area.
- fix issue #479: properly handle nose/unittest(2) SkipTest exceptions
during collection/loading of test modules. Thanks to Marc Schlaich
for the complete PR.
- fix issue490: include pytest_load_initial_conftests in documentation
and improve docstring.
- fix issue472: clarify that ``pytest.config.getvalue()`` cannot work
if it's triggered ahead of command line parsing.
- merge PR123: improved integration with mock.patch decorator on tests.
- fix issue412: messing with stdout/stderr FD-level streams is now
captured without crashes.
- fix issue483: trial/py33 works now properly. Thanks Daniel Grana for PR.
- improve example for pytest integration with "python setup.py test"
which now has a generic "-a" or "--pytest-args" option where you
can pass additional options as a quoted string. Thanks Trevor Bekolay.
- simplified internal capturing mechanism and made it more robust
against tests or setups changing FD1/FD2, also better integrated
now with pytest.pdb() in single tests.
- improvements to pytest's own test-suite leakage detection, courtesy of PRs
from Marc Abramowitz
- fix issue492: avoid leak in test_writeorg. Thanks Marc Abramowitz.
- fix issue493: don't run tests in doc directory with ``python setup.py test``
(use tox -e doctesting for that)
- fix issue486: better reporting and handling of early conftest loading failures
- some cleanup and simplification of internal conftest handling.
- work a bit harder to break reference cycles when catching exceptions.
Thanks Jurko Gospodnetic.
- fix issue443: fix skip examples to use proper comparison. Thanks Alex
Groenholm.
- support nose-style ``__test__`` attribute on modules, classes and
functions, including unittest-style Classes. If set to False, the
test will not be collected.
- fix issue512: show "<notset>" for arguments which might not be set
in monkeypatch plugin. Improves output in documentation.
|
|
|
|
NEWS for rsync 3.1.1 (22 Jun 2014)
Protocol: 31 (unchanged)
Changes since 3.1.0:
BUG FIXES:
- If the receiver gets bogus filenames from the sender (an unexpected
leading slash or a ".." infix dir), exit with an error. This prevents a
malicious sender from trying to inject filenames that would affect an
area outside the destination directories.
- Fixed a failure to remove the partial-transfer temp file when interrupted
(and rsync is not saving the partial files).
- Changed the chown/group/xattr-set order to avoid losing some security-
related xattr info (that would get cleared by a chown).
- Fixed a bug in the xattr-finding code that could make a non-root-run
receiver not able to find some xattr numbers.
- Fixed a bug in the early daemon protocol where a timeout failed to be
honored (e.g. if the remote side fails to send us the initial protocol
greeting).
- Fixed unintended inclusion of commas in file numbers in the daemon log.
- We once again send the 'f' sub-flag (of -e) to the server side so it
knows that we can handle incremental-recursion directory errors properly
in older protocols.
- Fixed an issue with too-aggressive keep-alive messages causing a problem
for older rsync versions early in the transfer.
- Fixed an incorrect message about backup-directory-creation when using
--dry-run and the backup dir is not an absolute path.
- Fixed a bug where a failed deletion and/or a failed sender-side removal
would not affect the exit code.
- Fixed a bug that caused a failure when combining --delete-missing-args
with --xattrs and/or --acls.
- Fixed a strange dir_depth assertion error that was caused by empty-dir
removals and/or duplicate files in the transfer.
- Fixed a problem with --info=progress2's output stats where rsync would
only update the stats at the end of each file's transfer. It now uses
the data that is flowing for the current file, making the stats more
accurate and less jumpy.
- Fixed an itemize bug that affected the combo of --link-dest, -X, and -n.
- Fixed a problem with delete messages not appearing in the log file when
the user didn't use --verbose.
- Improve chunked xattr reading for OS X.
- Removed an attempted hard-link xattr optimization that was causing a
transfer failure. This removal is flagged in the compatibility code, so
if a better fix can be discovered, we have a way to flip it on again.
- Fixed a bug when the receiver is not configured to be able to hard link
symlimks/devices/special-file items but the sender sent some of these
items flagged as hard-linked.
- We now generate a better error if the buffer overflows in do_mknod().
- Fixed a problem reading more than 16 ACLs on some OSes.
- Fixed the reading of the secrets file to avoid an infinite wait when
the username is missing.
- Fixed a parsing problem in the --usermap/--groupmap options when using
MIN-MAX numbers.
- Switched Cygwin back to using socketpair "pipes" to try to speed it up.
- Added knowledge of a few new options to rrsync.
ENHANCEMENTS:
- Tweaked the temp-file naming when --temp-dir=DIR is used: the temp-file
names will not get a '.' prepended.
- Added support for a new-compression idiom that does not compress all the
matching data in a transfer. This can help rsync to use less cpu when a
transfer has a lot of matching data, and also makes rsync compatible with
a non-bundled zlib. See the --new-compress and --old-compress options in
the manpage.
- Added the support/rsync-no-vanished wrapper script.
- Made configure more prominently mention when we failed to find yodl (in
case the user wants to be able to generate manpages from *.yo files).
- Have manpage mention how a daemon's max-verbosity setting affects info
and debug options. Also added more clarification on backslash removals
for excludes that contain wildcards.
- Have configure check if for the attr lib (for getxattr) for those systems
that need to link against it explicitly.
- Change the early dir-creation logic to only use that idiom in an
inc-recursive copy that is preserving directory times. e.g. using
--omit-dir-times will avoid these early directories being created.
- Fix a bug in cmp_time() that would return a wrong result if the 2 times
differed by an amount greater than what a time_t can hold.
DEVELOPER RELATED:
- We now include an example systemd file (in packaging/systemd).
- Tweaked configure to make sure that any intended use of the included popt
and/or zlib code is put early in the CFLAGS.
|
|
|
|
|
|
|
|
Patch #309 - 2014/07/13
* modify configure script to work around debris left by XQuartz
upgrades.
* modify test-package scripts to enable ReGIS graphics.
* improvements/additions to ReGIS graphics, e.g., drawing text (Ross
Comb).
* add --with-Xaw3dxft, to link with Xaw 3d xft library (Stephen P
Wall).
* amend reset for cursor-shape to take into account cursorUnderline
resource (report by Christian Weisgerber).
* update config.guess, config.sub
|
|
Changes in 2.0.28
Fix media syncing state not being properly reset when deauthorizing.
Changes in 2.0.27
Media syncing has been rewritten to be more reliable:
It no longer locks the account, so multiple clients can
sync media at the same time (but please do not attempt this
with the mobile clients yet, as they have not yet been
updated). This is also paving the way towards background
syncing of media in the future.
The initial sync on a new device can now avoid both sending
and receiving files that already exist on both ends.
A large number of deleted files could cause timeouts in
the past; this is no longer the case.
When resuming from a connection error, less data needs to
be sent than before.
When deauthorizing sync, the media syncing state is reset.
On Linux, the Anki folder is now Documents/Anki, which matches
the other platforms. Anki will use a previous Anki folder if
it exists, so please rename it manually if you wish to use the
new location.
Searches in the Browse screen can now be saved into the sidebar,
and the tab key order has been improved (thanks to Houssam).
These changes conflict with previous versions of the Advanced
Browser add-on, so please ensure you download the latest version
if you use this add-on.
Fixed a bug where a filtered deck with or in it could mistakenly
include suspended or buried cards.
Fixed crashes in the alternate Mac build.
The Preview window now has a "replay audio" button, with the
shortcut R like in the review screen.
The post-import log now lists the number of notes that were
unchanged by a text import. "Unchanged" includes notes that
were present in your collection as well as the imported file
but were not updated in your collection, whether because the
update option was off or they were not changed.
The options menu next to each deck in the deck list now has an
"export" button, which opens the export dialog with that deck
already selected.
Donât display the "added duplicate" message multiple times when
a note is already duplicated in the collection but only one
new duplicate was added.
Fix a bug where Anki was inserting <br>s into the text of
imported cards if using quoted multi-line notes, with HTML
disabled.
Provide more helpful errors when importing corrupt collection
packages or .anki2 files.
Anki now remembers the previous position and size of various
dialogs it didnât previously.
The "suspended" label in the card types graph has been changed
to "Suspended+Buried" since that category also counts buried
cards.
Anki now makes sure the Get Addons dialog opens at an appropriate
size when the system font size is unusual.
Check that saving a statistics image was successful before
displaying a success message, and give a warning if Anki doesnât
have permission to write to the desktop.
Give a less confusing error when Anki canât create a profile
folder.
Donât throw an error when attempting to delete unused media
that has already been removed.
Fix for exporting to csv with quotes in a field (thanks to
Rubyu)
Tweaks to latex handling which will allow add-ons to switch to
pdftex (thanks to Joel)
Round cards/minute stat instead of truncating it (thanks to
Markus)
Fix errors when importing Mnemosyneâs latest db format.
Support pasting of m4a urls.
Add a fix for an AnkiDroid bug to the Check DB function (thanks
to Houssam)
Fix tooltip colour for inverted colour schemes. (Thanks to Pat)
Donât trigger help when enter pressed in preferences screen.
Fix a bug on Windows where the main window was left open when
a corrupt collection was detected.
Fix error when Anki is running for multiple users on a multi-user
machine.
Donât try to authenticate username and password if user cancels
login dialog.
Allow changing case of profile name on Windows.
Update all duplicates when importing in update mode.
Remove initial search text when ctrl-clicking a left panel
item.
Fix rounding of cards/minute in stats; use 1 decimal place.
Fix error when exporting w/ non-latin text in profile name.
Prevent profiles from getting out of sync when rename fails.
Update note type on apkg import when only templates have changed
Search terms in the browser and filtered decks can now be
separated by a full width space (thanks to Paul)
Catch some rare startup errors on Windows and show them in a
dialog box instead of exiting silently.
A special thanks to Soren, who contributed many of the bugfixes in
this release.
|
|
mailservers, where users may not (nescessarily) have systemaccoounts,
but accountinformation is stored within an LDAP database.
|
|
|
|
|
|
|
|
|
|
|
|
those id's on the standard output. This Linux pidof port is intended to
help porting scripts.
|
|
Refer to the attr(5) manual page for general extended attribute
(EA) information and references to other EA manual pages.
|
|
of ELF executables. It is used by the Nix Packages collection
(nixpkgs).
|
