The CMPH Library encapsulates the newest and more efficient algorithms in an
easy-to-use, production-quality, fast API. The library was designed to work
with big entries that cannot fit in the main memory. It has been used
successfully for constructing minimal perfect hash functions for sets with
more than 100 million of keys, and we intend to expand this number to the
order of billion of keys. Although there is a lack of similar libraries, we
can point out some of the distinguishable features of the CMPH Library:
- Fast.
- Space-efficient with main memory usage carefully documented.
- The best modern algorithms are available (or at least scheduled for
implementation :-)).
- Works with in-disk key sets through of using the adapter pattern.
- Serialization of hash functions.
- Portable C code (currently works on GNU/Linux and WIN32 and is reported
to work in OpenBSD and Solaris).
- Object oriented implementation.
- Easily extensible.
- Well encapsulated API aiming binary compatibility through releases.
- Free Software.
Note that the 2-clause-bsd is not identical. (I know there are
hundreds of different BSD licenses -- we should not track them all.)
The DESCRiption follows:
Log4cplus provides a port of the log4j logging framework for C++.
Log4cplus is a featureful log facility with various filters, run-time
re-configurations, and Wide Character (UNICODE) support. It provides
log levels, hierarchal loggers, NDC (Nested Diagnostic Context),
and log rotation support. It can log to the console, files, syslog,
Windows events, or via sockets.
Firefox 4 is based on the Gecko 2.0 Web platform. This release features
JavaScript execution speeds up to six times faster than the previous
version, new capabilities for Web Developers and Add-on Developers such as
hardware accelerated graphics and HTML5 technologies, and a completely
revised user interface.
* Various MinGW portability fixes.
* Various git-p4 enhancements (in contrib).
* Various vcs-svn, git-svn and gitk enhancements and fixes.
* Various git-gui updates (0.14.0).
* Update to more modern HP-UX port.
* The codebase is getting prepared for i18n/l10n; no translated
strings nor translation mechanism in the code yet, but the strings
are being marked for l10n.
* The bash completion script can now complete symmetric difference
for "git diff" command, e.g. "git diff ...bra<TAB>".
* The default minimum length of abbreviated and unique object names
can now be configured by setting the core.abbrev configuration
variable.
* "git apply -v" reports offset lines when the patch does not apply at
the exact location recorded in the diff output.
* "git config" used to be also known as "git repo-config", but the old
name is now officially deprecated.
* "git checkout --detach <commit>" is a more user friendly synonym for
"git checkout <commit>^0".
* "git checkout" performed on detached HEAD gives a warning and
advice when the commit being left behind will become unreachable from
any branch or tag.
* "git cherry-pick" and "git revert" can be told to use a custom merge
strategy, similar to "git rebase".
* "git cherry-pick" remembers which commit failed to apply when it is
stopped by conflicts, making it unnecessary to use "commit -c $commit"
to conclude it.
* "git cvsimport" bails out immediately when the cvs server cannot be
reached, without spewing unnecessary error messages that complain about
the server response it never got.
* "git fetch" vs "git upload-pack" transfer learned 'no-done'
protocol extension to save one round-trip after the content
negotiation is done. This saves one HTTP RPC, reducing the overall
latency for a trivial fetch.
* "git fetch" can be told to recursively fetch submodules on-demand.
* "git grep -f <filename>" learned to treat "-" as "read from the
standard input stream".
* "git grep --no-index" did not honor pathspecs correctly, returning
paths outside the specified area.
* "git init" learned the --separate-git-dir option to allow the git
directory for a new repository created elsewhere and linked via the
gitdir mechanism. This is primarily to help submodule support later
to switch between a branch of superproject that has the submodule
and another that does not.
* "git log" type commands now understand globbing pathspecs. You
can say "git log -- '*.txt'" for example.
* "git log" family of commands learned --cherry and --cherry-mark
options that can be used to view two diverged branches while omitting
or highlighting equivalent changes that appear on both sides of a
symmetric difference (e.g. "log --cherry A...B").
* A lazy "git merge" that didn't say what to merge used to be an error.
When run on a branch that has an upstream defined, however, the command
now merges from the configured upstream.
* "git mergetool" learned how to drive "beyond compare 3" as well.
* "git rerere forget" without pathspec used to forget all the saved
conflicts that relate to the current merge; it now requires you to
give it pathspecs.
* "git rev-list --objects $revs -- $pathspec" now limits the objects listed
in its output properly with the pathspec, in preparation for narrow
clones.
* "git push" with no parameters gives better advice messages when
"tracking" is used as the push.default semantics or there is no remote
configured yet.
* A possible value to the "push.default" configuration variable,
'tracking', gained a synonym that more naturally describes what it
does, 'upstream'.
* "git rerere" learned a new subcommand "remaining" that is similar to
"status" and lists the paths that had conflicts which are known to
rerere, but excludes the paths that have already been marked as
resolved in the index from its output. "git mergetool" has been
updated to use this facility.
perl. I'm not sure there is a better way to fix this since mtn-cleanup is
installed into ${PREFIX}/bin/.
Also skip the interpreter check for some helper files that are not used by
default. These may be put in separate packages but we can do this later;
for now I'm just adding a comment mentioning the idea.
Bump PKGREVISION to 2.
This requires a dependency on Python. Bump PKGREVISION to 1.
This is probably not the best solution: I guess it'd be nice to have a
libevent package that only included the binary libraries and no dependency
on Python, and another libevent-dev package with the event_rpcgen.py binary
and the Python dependency. But this can be done later; I want to just fix
the breakage first.
This release, in addition to being a bugfix release, also resolves a
number of security vulnerabilities. It resolves CVE-2011-1685,
CVE-2011-1686, CVE-2011-1687, CVE-2011-1688, CVE-2011-1689, and
CVE-2011-1690.
* Cleanups identified by perlcritic.
* Clear the system attribute cache to avoid 'sticky' attributes like
the queue subject tag.
* Fix our signature escaping so we better match FCKEditor and don't
misidentify signatures during processing.
* Add the ability to create BasedOn Custom Fields from intiialdata
* Provide a callback to affect the display format in admin pages
* Fix id prefixing on Custom Fields to be RTIR compatible
* Fix#16656 - Requestors with OwnTicket could show up in the owner list
in other Queues.
* Don't attach the original multipart mail to notifications that already
contain one part of it.
* Work around CGI.pm 3.51 and 3.52 which add ; charse=ISO-8859-1 to our
utf-8 encoded javascript.
pkgsrc fixes
* Let to use system bzip2 library instead of included one,
same as p5-Compress-Raw-Zlib.
CHANGES
-------
2.033 11 Jan 2011
* Fixed typos & spelling errors.
[perl# 81782]
2.032 4 Jan 2011
* No Changes
2.031 21 September 2010
* Updated to use bzip2 1.0.6
Version 1.0.6 removes a potential security vulnerability,
CVE-2010-0405, so all users are recommended to upgrade immediately.
Changes from Ant 1.8.1 TO Ant 1.8.2
===================================
Changes that could break older environments:
-------------------------------------------
* Prior to Ant 1.8.0 the <copy> task and several other tasks would
overwrite read-only destination files. Starting with 1.8.0 they
would only do so under special circumstances. Ant 1.8.2 now
consistently won't replace a read-only file by default. The same is
true for a number of other tasks.
The <copy>, <move> and <echo> tasks now have a new force attribute
and <concat> has a new forceReadonly attribute that can be used to
make the task overwrite read-only destinations.
Bugzilla Report 49261.
* Removed ant-nodeps.jar; it is now merged into ant.jar.
* DOMElementWriter#encode used to employ special code before encoding
ampersands so that { remained { rather than being turned
into &#123;. This is no longer the case, ampersands will now
be encoded unconditionally.
Also DOMElementWriter#encodeData will treat CDATA sections containing a
literal "]]>" sequence different now - it will split the CDATA
section between the second "]" and ">" and create two sections.
This affects <echoxml> task as well as the XML logger or JUnit
formatter where ampersands will now always get encoded.
In addition DOMElementWriter will now replace the characters \t, \r
and \n in attribute values by entity references.
Bugzilla Report 49404.
* The list elements returned by ProjectHelper#getExtensionStack are
now String arrays of length 3 rather than 2 in order to support the
onMissingExtensionPoint attribute.
Bugzilla Report 49473.
* When using <property file="..." prefix="..."/> properties defined
inside the same file will only get used in expansions if the ${}
reference uses the same prefix. This is different from Ant 1.8.1
but is the same behavior Ant 1.8.0 and earlier exhibited.
A new attribute prefixValues can be used to re-enable the behavior
of Ant 1.8.1.
Bugzilla Report 49373.
* The files and directories used by Git, Mercurial and Bazaar to
store their information are now excluded by the defaultexcludes.
Bugzilla Report 49624.
* The <junit> task no longer generates TestListener events - which
have been introduced in ant 1.7.0 - by default. The task has a new
attribute enableTestListenerEvents and a new "magic" property
ant.junit.enabletestlistenerevents has been added that can be used
to reinstate the old behavior.
Fixed bugs:
-----------
* hostinfo now prefers addresses with a hostname over addresses without
a hostname, provided the addresses have the same scope.
For local lookup, no IP address will be put in NAME / DOMAIN anymore.
For remote lookup, if a host name was provided and only an IP address is
found, the IP address will no longer overwrite the host name provided to the
task.
Bugzilla Report 49513
* mmap-based file copy problems under JDK 1.4 on Linux.
Bugzilla Report 49430.
* The Sun JVM tries to mmap the entire file during a copy.
For large files this is not feasible.
We now explicitly request to copy at most 16 MiB per request.
Bugzilla Report 49326.
* DemuxInputStream.read() should return unsigned values
Bugzilla Report 49279.
* The MIME mailer ignored the port parameter when using SSL.
Bugzilla Report 49267.
* <xslt> ignored the classpath when using the default TraX processor.
Bugzilla Report 49271.
* <checksum>'s totalproperty only worked reliably if the same file
name didn't occur inside more than one directory.
Bugzilla Report 36748.
* <ftp> could fail to download files from remote subdirectories under
certain circumstances.
Bugzilla Report 49296.
* <junit> will now produce better diagnostics when it fails to delete
a temporary file.
Bugzilla Report 49419.
* Ant would often scan directories even though there were known to
only hold excluded files when evaluating filesets. This never
resulted in wrong results but degraded performance of the scan
itself.
Bugzilla Report 49420.
* <javac> failed for long command lines on OS/2.
Bugzilla Report 49425.
* <junitreport> did not handle encodings well for stdout/stderr.
Bugzilla Report 49418.
* <junit> could issue a warning about multiple versions of Ant on the
CLASSPATH if two CLASSPATH entries differed in case on a
case-insensitive file system.
Bugzilla Report 49041.
* The <restrict> resource collection was checking every resource even if
we actually just want the first one, like in the example of use of
resourcelist in the documentation (getting the first available resource
from a mirror list).
* A race condition could lead to build failures if multiple <mkdir>
tasks were trying to create the same directory.
Bugzilla Report 49572.
* the toString() method of the Resources class - and thus any
${toString:} expansion of a reference to a <resources> element -
didn't iterate over its nested elements if it hadn't done so prior
to the toString invocation already.
Bugzilla Report 49588.
* <apply> in parallel mode didn't work together with a nested
<redirector> if maxparallel was <= 0 (the default) or no source
files matched.
Bugzilla Report 49594.
* <jar filesetmanifest="merge"> didn't work for manifests added via
<zipfileset>s that used the prefix or fullpath attributes.
Bugzilla Report 49605.
* <tempfile createfile="true"> would cause an error unless the prefix
attribute has been specified.
Bugzilla Report 49755.
* If forked, after finished <java> was still reading the input stream
for a bunch of characters, then stealing them from a following <input>.
Bugzilla Report 49119.
* Ant could be leaking threads for each forked process (started by
<exec>, <apply>, <java> or similar tasks) that didn't receive input
from a resource or string explicitly.
Bugzilla Report 49587.
* Project#setDefault threw an exception when null was passed in as
argument, even though the Javadoc says null is a valid value.
Bugzilla Report 49803.
* runant.py would swallow the first argument if CLASSPATH wasn't set.
Bugzilla Report 49963.
* <taskdef> failed to load resources from jar files contained in a
directory that has a "!" in its name.
Bugzilla Report 50007.
* ant.bat exit strategy improvements and issues
make the exit codes work in environments where 4NT or MKS are installed
Bugzilla Report 41039.
* <signjar> would fail if used via its Java API and the File passed
into the setJar method was not "normalized" (i.e. contained ".."
segments).
Bugzilla Report 50081.
* <delete> ignored <fileset>'s errorOnMissingDir attribute
Bugzilla Report 50124.
* <symlink> failed to close files when reading a list of symbolic
links from a properties file.
Bugzilla Report 50136.
* <parallel> could allow tasks to start executing even if a task
scheduled to run before them timed out.
Bugzilla Report 49527.
* If a <junit> batch with multiple tests times out Ant logs a message
about a test named Batch-With-Multiple-Tests since 1.8.0 but the
logic that determined the Java package of this pseudo-test has been
wrong.
Bugzilla Report 45227.
* <propertyfile> didn't preserve the original linefeed style when
updating a file.
Bugzilla Report 50049.
* <zip>'s whenEmpty behavior never consulted the non-fileset
resources so the task could fail even though resources have been
provided using non-fileset resource collections.
Bugzilla Issue 50115.
* ftp chmod could throw a NPE.
Bugzilla report 50217.
* The project help (-p option in the command line) will now print
the dependencies of the targets in debug mode (-d on the command
line)
Other changes:
--------------
* <concat>'s force attribute has been deprecated in favor of a new
overwrite attribute that is consistent with <copy>'s attribute
names.
* You can now specify a list of methods to run in a JUnit test case.
Bugzilla Report 34748.
* properties in files read because of the -propertyfile command line
option will now get resolved against other properties that are
defined before the project starts executing (those from the same or
earlier -propertfiles or defined via the -D option).
Bugzilla Report 18732.
* <pathelement>s can now contain wildcards in order to use wildcard
CLASSPATH entries introduced with Java6.
The wildcards are not expanded or even evaluated by Ant and will be
used literally. The resulting path may be unusable as a CLASSPATH
for Java versions prior to Java6 and likely doesn't mean anything
when used in any other way than a CLASSPATH for a forked Java VM.
Bugzilla Report 46842.
* A new attribute allows targets to deal with nonexistent extension
points, i.e. they can extend an extension-point if it has been
defined or silently work as plain targets if it hasn't. This is
useful for targets that get included/imported in different
scenarios where a given extension-point may or may not exist.
Bugzilla Report 49473.
* Ant now logs a warning message if it fails to change the file
modification time in for example when using <touch> or preserving
timestamps in various tasks.
Bugzilla Report 49485.
* ProjectHelpers can now be installed dynamically via the <projecthelper>
Ant task.
* <import> is now able to switch to the proper ProjectHelper to parse
the imported resource. This means that several kinds of different build
files can import each other.
* <copy tofile=""> now also works for non-filesystem resources.
Bugzilla Report 49756.
* The <linecontainsregexp> filter now supports a casesensitive
attribute.
* The <containsregexp> selector now supports casesensitive, multiline
and singleline attributes.
Bugzilla Report 49764.
* A new <cutdirsmapper> can be used like wget's --cut-dirs option to
strip leading directories from file names.
* <javah> now supports the GNU project's gcjh compiler.
Bugzilla Report 50149.
* <checksum> supports additional views of a file's path as elements
for a custom pattern.
Bugzilla Report 50114.
* JUnit XMLResultAggregator logs the stack trace of caught IO exceptions
in verbose runs.
Bugzilla Report 48836.
* StringUtils.parseHumanSizes() should turn parse failures into
BuildExceptions.
Bugzilla Report 48835.
* New task <bindtargets> to make a list of targets bound to some
specified extension point.
* Initial support for OpenJDK7 has been added.
* Ant now uses java.net.CookieStore rather than
java.util.ServiceLocator to detect whether the environment is a
Java 1.6 system. This means releases of gcj/gij at the time of
this release of Ant are detected as Java 1.5 and not 1.6.
Bugzilla Report 50256.
* It is now possible to write a compiler adapter for <javac> that
compiles sources with extensions other than .java (but that still
compile to .class files).
Bugzilla Report 48829.
* The performance of VectorSet#add(Object) has been improved which
should also benefit any operation that scans directories in Ant.
Bugzilla Report 50200.
