New Libraries
* Icl: Interval Container Library, interval sets and maps and aggregation of
associated values, from Joachim Faulhaber.
Updated Libraries
* Array:
- Added support for cbegin/cend
- Fixed a problem with the Sun compiler
* Asio:
- Fixed a problem on older Linux kernels (where epoll is used without timerfd
support) that prevents timely delivery of deadline_timer handlers, after
the program has been running for some time
* Bind:
- make_adaptable now documented
* Concept Check:
- fixed warnings with self-assignment
* Filesystem:
- Version 3 of the library is now the default.
- IBM vacpp: Workaround for compiler bug affecting iterator_facade
- Verify, clarify, document that <boost/config/user.hpp> can be used to
specify BOOST_FILESYSTEM_VERSIO
- Replaced C-style assert with BOOST_ASSERT.
- Undeprecated unique_path(). Instead, add a note mentioning the workaround
for lack of thread safety and possible change to cwd. unique_path() is just
too convenient to deprecate!
- Cleared several GCC warnings.
- Changed V2 code to use BOOST_THROW_EXCEPTION.
- Windows: Fix status() to report non-symlink reparse point correctly.
- Add symlink_option to recursive_directory_iterator, allowing control over
recursion into directory symlinks. Note that the default is changed to not
recurse into directory symlinks.
- Reference documentation cleanup, including fixing missing and broken links,
and adding missing functions.
- Miscellaneous implementation code cleanup.
* Fusion:
- vector copy constructor now copies sequence members in the same order on different platforms
* Graph:
- Fixed Graphviz output to work on Visual C++ 7.1.
- Replaced assert with BOOST_ASSERT.
- Changed to Boost.Filesystem v3.
More...
a. Fixed a bug that caused the unconverted filename to be added to the list of
completions when the application specified filename conversion functions.
b. Fixed a bug that caused the wrong filename to be passed to opendir when the
application has specified a filename dequoting function.
c. Fixed a bug when repeating a character search in vi mode in the case where
there was no search to repeat.
d. When show-all-if-ambiguous is set, the completion routines no longer insert
a common match prefix that is shorter than the text being completed.
e. The full set of vi editing commands may now be used in callback mode.
dimensions while running in `no-echo' mode.
h. Fixed a bug that caused readline to dump core if an application called
rl_prep_terminal without setting rl_instream.
i. Fixed a bug that caused meta-prefixed characters bound to incremental
search forward or backward to not be recognized if they were typed
subsequently.
j. The incremental search code treats key sequences that map to the same
functions as (default) ^G, ^W, and ^Y as equivalent to those characters.
k. Fixed a bug in menu-complete that caused it to misbehave with large
negative argument.
l. Fixed a bug that caused vi-mode yank-last-arg to ring the bell when invoked
at the end of the line.
m. Fixed a bug that made an explicit argument of 0 to yank-last-arg behave
as if it were a negative argument.
n. Fixed a bug that caused directory names in words to be completed to not
be dequoted correctly.
New Features
a. The history library does not try to write the history filename in the
current directory if $HOME is unset. This closes a potential security
problem if the application does not specify a history filename.
b. New bindable variable `completion-display-width' to set the number of
columns used when displaying completions.
c. New bindable variable `completion-case-map' to cause case-insensitive
completion to treat `-' and `_' as identical.
d. There are new bindable vi-mode command names to avoid readline's case-
insensitive matching not allowing them to be bound separately.
e. New bindable variable `menu-complete-display-prefix' causes the menu
completion code to display the common prefix of the possible completions
before cycling through the list, instead of after.
* Fix crash in GraphVizWriter when GRAPHVIZ_TARGET_IGNORE_REGEX is used
* FindPerlLibs: Add notice of copyright
* libarchive: Define major/minor/makedev only where needed
* libarchive: Use OpenSSL only if CMAKE_USE_OPENSSL
* Fix documentation of MSVC_VERSION
* Silence the may be used uninitialized warnings: initialize stuff.
* CPack Tests the different ways of packaging components
* Avoid foreach IN LISTS syntax which is not supported by CMake 2.6
Changes include
- man page claims that you can use auth type = none, but in reality this
didn't work (insisted on a username being given and then used, plus always
advertised CHAP to the target). Make initiator work as advertised (i.e.
defaults to auth type none and so don't require a username).
- document -D option to iscsi-initiator
- Remove hostname from pathname to storage in mount point.
This means /mnt/mytarget.domain.local/target0/storage is now
/mnt/target0/storage.
- Note limitations of CHAP implementation in BUGS section.
- Skip target if TargetName is empty.
- Use relevant TargetAddress, not just first one we happen to find.
- Handle NOP-OUT CmdSN and immediate bit.
- Handle NOP-IN TransferTag=0xffffffff.
- Interim solution for dealing with Underflow bit in iSCSI response.
- iscsi-initiator now talks to istgt and other targets.
- Retry read capacity. Device may not be ready on first access, so need to
wait and re-issue.
gio-querymodules utility does not delete the file when no modules are left,
so we were left with a giomodule.cache file after removing glib2 that was
not managed by any package.
Also, and as a side effect, ensure that the files are created during
install time so that in-place updates work correctly.
Bump PKGREVISION to 2.
some packages deduce the full name to the tool instead of relying on the
PATH.
Attempt to do better by setting the GLIB_COMPILE_SCHEMAS variable to
a non-absolute name during the build, hoping that such packages do define
this Makefile variable.
This should fix the build of evince as reported in PR pkg/44589.
Citrix XenServer includes a XML-RPC based API, providing programmatic access
to the extensive set of XenServer management features and tools. The
XenServer API can be called from a remote system as well as local to the
XenServer host. Remote calls are generally made securely over HTTPS, using
port 443.
Changes since 2.2.0
-------------------
Bugfixes
Changes since 2.1.1
-------------------
Features:
- Source code windows have now a tab attached, which makes switching
source files much easier.
- Source code windows now show line numbers at the left.
- There are now "Find Next" and "Find Previous" commands with shortcuts
F3 and Shift+F3.
- Improved support of template types in the type tables (which are used
to show structure members next to a structure variable). Notably, the
number of elements in STL and Qt collection classes are shown.
- Arguments for the debugged program can be passed on KDbg's command line.
Bug fixes
Upstream changes:
2011-01-28 Joerg Billeter <j@bitron.ch>
Release 0.6.1
Fix memory leak in LinkedList.clear
Based on patch by Travis Reitter, fixes bug 639254.
2011-01-20 Maciej Piechotka <uzytkownik2@gmail.com>
Remove depending on order of iteration in read-only collections' test
2010-11-20 Maciej Piechotka <uzytkownik2@gmail.com>
Fix memory leak on freeingi LinkedList
If fixes issue described in bug #63522
While here, add pdf option to enable pdf support without GTK2+.
= NEWS --- The history of Cutter
== [release-1-1-7] 1.1.7: 2011-02-13
=== Cutter
==== Improvements
* Supported growlnotify.
==== Changes
* Droped Debian GNU/Linux lenny support.
==== Fixes
* Fixed using unknown symbol on CentOS 5.5.
[Reported by Hiroaki Nakamura]
=== Document
===== Fixes
* Fixed typos in tutorial.
[Reported by Hiroaki Nakamura]
=== Thanks
* Hiroaki Nakamura
Supposedly API backwards-compatible with libevent 1.x.
Will need a recursive revbump for shlib name changes, though.
All regression tests pass except "thread_conditions_simple". Not sure
why yet, but it's new API so shouldn't break anything existing in pkgsrc
today.
Changes in libevent 2.0 include cleaner interfaces, buffered I/O improvements,
improved multithreading and openssl integration.
* lua-mode now lives in github.
* Small patch to show how to make Emacs automatically use lua-mode for
files that have a hash-bang line containing "lua", even if they don't
have a ".lua" suffix.
* Customization of indent level and comments.
syntax which compiles into Python modules for maximum performance. Mako's
syntax and API borrows from the best ideas of many others, including Django
templates, Cheetah, Myghty, and Genshi. Conceptually, Mako is an embedded
Python (i.e. Python Server Page) language, which refines the familiar ideas of
componentized layout and inheritance to produce one of the most straightforward
and flexible models available, while also maintaining close ties to Python
calling and scoping semantics.
fix a packaging botch whereby the file of random numbers was not installed,
and a reference to the build directory was embedded in the executables.
now the regression tests run to completion:
Running ex-dep:
Running ex-ham7a:
Running ex-ham7b:
Running ex-ldpc-encode:
Running ex-ldpc36-1000a:
Running ex-ldpc36-5000a:
Running ex-ldpcvar-5000a:
Running ex-wrong-model:
155c155
< Bit error rate (on message bits only): 1.320e-02
---
> Bit error rate (on message bits only): 1.299e-02
The POCO C++ Libraries (POCO stands for POrtable COmponents) are
open source C++ class libraries that simplify and accelerate the
development of network-centric, portable applications in C++. The
libraries integrate perfectly with the C++ Standard Library and
fill many of the functional gaps left open by it. Their modular
and efficient design and implementation makes the POCO C++ Libraries
extremely well suited for embedded development, an area where the
C++ programming language is becoming increasingly popular, due to
its suitability for both low-level (device I/O, interrupt handlers,
etc.) and high-level object-oriented development. Of course, the
POCO C++ Libraries are also ready for enterprise-level challenges.
Collection.
Term::Shell makes it joyfully easy to write command-line interfaces in Perl.
All the boring details like command-line parsing and terminal handling are
done for you.
1.7.5 (2011-02-01)
This is a quick bugfix release to include some fixes accidentally
dropped from 1.7.4.
subrepo: fix update -C with svn subrepos when cwd != repo.root
subrepo: make update -C clean the working directory for svn subrepos
url: add --insecure option to bypass verification of ssl certificates
win32mbcs: Fix typo in documentation
1.7.4 (2011-02-01)
This is a scheduled bugfix release that smooths out some of the
rough edged introduced with 1.7.3's HTTPS certificate verification.
bookmarks: always write undo file
bookmarks: respect rollbacks dryrun parameter
hgrc.5: mention that web.cacerts are run through util.expandpath
opener: force copy on 'a'ppend if nlinks() returns 0 (issue1922)
subrepo: compare svn subrepo state to last committed revision
subrepo: do not report known files inside repositories as unknown
url: 'ssh known host'-like checking of fingerprints of HTTPS certificates (see CACertificates)
url: check subjectAltName when verifying ssl certificate
url: expand path for web.cacerts
url: fix UnicodeDecodeError on certificate verification error
win32: add cacert.pem file to Inno Setup installer
win32: win32console.GetStdHandle() can return None
RP pkg/44475
Changelog:
New in CSSC-1.3.0, 2010-11-07
* This release is identical to CSSC-1.3.0rc1 apart from the
version number.
New in CSSC-1.3.0rc1, 2010-05-16
* The CSSC manual is now published under version 1.3 of the
GNU Free Documentation License.
* The code and build system have been modernised and some
redundant files have been removed. This may make it harder
to build CSSC on very old systems. However, it's now easier
to maintain. If CSSC no longer builds on a system that's
important to you, please send mail to <bug-cssc@gnu.org>
explaining your problem.
* CSSC now uses the GNU portability library, gnulib.
* CSSC now requires a C++ compiler with working exceptions and
a working STL implementation.
* Adoption of some STL data structures has made some parts of
CSSC faster, notably "val".
* CSSC now includes unit tests. We also run the y2k tests for
"make check". If valgrind is installed, it will be used to
for verification during tests. You can turn this off by
givig the --without-valgrind option to configure.
* The configure script now supports a new option
"--enable-warnings" which enables many GCC warning options.
New in CSSC-1.2.0, 2009-04-11
* This stable release is almost identical to the 1.1.1 test
release; there are no functional changes.
New in CSSC-1.1.1, 2008-01-18
* Updated the NEWS file to give estimated release dates
and user-visible changes for ancient releases.
* The -l and -L options of "get" are now implemented.
* If you use "admin -dlX" to unlock release X, but all
releases are locked, an explanatory error message is issued.
Previously, the attempt was simply silently ignored.
* prs no longer has a -R option.
* When CSSC tools are passed the name of a directory on the
command line, they operate on the SCCS files in that directory.
As of this release, subdirectories are ignored (instead of
generating an error when CSSC tries to open the directory as
a file).
New in CSSC-1.1.0, 2007-12-17
* GNU CSSC is now distributed under version 3 of the GNU General
Public License. The documentation is distributed under
version 1.2 of the GNU Free Documentation License.
* We now require the C++ compiler to support exceptions in
order to compile successfully. Previously, CSSC allowed
compilation without exceptions, but this meant that graceful
recovery from failures was not possible.
* Fixed a spurious failure when writing out the SCCS file when
the Y flag is set in the SCCS file.
* "sccs unedit /tmp/SCCS/s.foo" will no longer cause the
deletion of ./s.foo, if it exists. Instead, ./foo will be
deleted, as intended.
* Support for some older SCCS files as produced by Unix System III
(some of these have leading spaces in the counts of inserted/
deleted/unchanged lines in the delta table).
* Compilation fixes for Darwin
Apache log4cxx is a logging framework for C++ patterned after Apache
log4j. Logging can be selectively enabled or disabled, and sent to
different and multiple output targets in user-chosen formats.
post-review is a command line tool for Windows, Linux and MacOS X
that simplifies both creating and updating review requests. It
can look at your source directory, generate a diff, and upload it
to a new or existing review request on an associated Review Board
server. This saves a lot of time over the web UI, and for some
types of code repositories (such as Perforce), its actually required
in order to produce a compatible diff.
mock is a Python module that provides a core Mock class. It removes
the need to create a host of stubs throughout your test suite. After
performing an action, you can make assertions about which methods /
attributes were used and arguments they were called with. You can
also specify return values and set needed attributes in the normal
way.
* The xfuncname pattern used by "git diff" and "git grep" to show the
last notable line in context were broken for python and ruby for a long
time.
* "git merge" into an unborn branch removed an untracked file "foo" from
the working tree when merged branch had "foo" (this fix was already in
1.7.3.3 but was omitted from the release notes by mistake).
* "git status -s" did not quote unprintable characters in paths as
documented.
* "git am --abort" used to always reset to the commit at the beginning of
the last "am" invocation that has stopped, losing any unrelated commits
that may have been made since then. Now it refrains from doing so and
instead issues a warning.
* "git blame" incorrectly reused bogusly cached result of textconv
filter for files from the working tree.
* "git commit" used to abort after the user edited the log message
when the committer information was not correctly set up. It now
aborts before starting the editor.
* "git commit --date=invalid" used to silently ignore the incorrectly
specified date; it is now diagnosed as an error.
* "git rebase --skip" to skip the last commit in a series used to fail
to run post-rewrite hook and to copy notes from old commits that have
successfully been rebased so far. Now it do (backmerge ef88ad2).
* "gitweb" tried to show a wrong feed logo when none was specified.
* Fix build error when cross-compiling to MinGW.
Changes 1.1:
* Update gnulib files.
* The core library is now compiled with many warning flags.
Some warnings have been fixed too.
Changes 1.0:
* Update gnulib files.
Changes 0.4.2:
* Fix endian related bug in self test.
* Add a note in README about NTLM being a insecure protocol.
* Cleaned up internal endianness handling.
* Update of gnulib files.
Changes 0.4.1:
* Update of gnulib files.
Changes 0.4.0:
* Only public API functions are exported in shared library.
Using libtool's -export-symbols-regex.
* Update of gnulib files.
* change dependency on py-cairo and gobject-introspection to build.
Those are required by configure script, but not required by
installed files. Furthermore, this package is python-version-independent
part of py-gobject, but full dependency of those packages introduce
dependency on certain python version.
* add direct full dependency on libffi, because it is already in buildlink3.mk.
(previously, it was indirect dependency from gobject-introspection, but
py-gobject is using it directly).
Bump PKGREVISION.
Big changes:
- Allows the symbolic links which point to directories, and treats
the names of struct, union, enum and enum member as definitions.
(5.7)
- Uses GPLv3 and LGPLv3 license. (5.6)
- Becomes GNU BOKINware. (5.5)
Details:
http://cvs.savannah.gnu.org/viewvc/global/global/NEWS?view=markup
Bunch of incompatible format changes. DON'T FORGET TO RE-GENERATE TAGS!
rebuild the precompiled database upon install/deinstall. This is mostly
a copy/paste of the same logic to deal with gio modules found in modules.mk.
No PKGREVISION bump because this does not affect the glib2 package in
itself.
