- WARN: devel/ply/Makefile:22: Found absolute pathname: /${EGDIR}
As ${EGDIR} is already an absolute pathname, there's no need to prefix it
with a slash.
Overview of changes between 1.10.1 and 1.10.2
=============================================
* New improved Tibetan shaper module. [Pema Geyleg]
* Bug fix in Khmer shaper module. [Jens Herden]
* Respect fontconfig reassignment of font pixelsize. [Funda Wang]
* Make OpenType GPOS handling more robust/correct. [Greg Aumann]
* Various documentation fixes and improvements.
* Misc bug fixes.
* Open Pango modules with lazy-bind flags. [John Rice]
library with some exciting features such as approximate (fuzzy) matching.
At the core of TRE is a new algorithm for regular expression matching with
submatch addressing. The algorithm uses linear worst-case time in the length
of the text being searched, and quadratic worst-case time in the length of
the used regular expression. In other words, the time complexity of the
algorithm is O(M2N), where M is the length of the regular expression and N
is the length of the text. The used space is also quadratic on the length
of the regex, but does not depend on the searched string. This quadratic
behaviour occurs only on pathological cases which are probably very rare
in practice.
Major key management changes:
- Private keys are no longer stored in your database. They
are stored in ~/.monotone/keys/ (Unix, OS X) or
%APPDATA%\monotone\keys\ (Windows). 'db migrate' will
automatically move your keys out of your database and into
their proper location. Consequences:
- 'genkey' no longer requires a database. Simply run it
once when you first start using monotone, even before you
have created a database.
- Running 'genkey' once will suffice to give all databases
on one computer access to your key. No more fiddling with
'read'.
- When you want to make your key available on another
computer, simply copy over the appropriate file from your
'keys' directory to the corresponding directory on the new
computer.
- Private keys also use a more standard on-disk envelope
encoding ("PBE-PKCS5v20(SHA-1,TripleDES/CBC)") instead of
previous ARC4. More secure, and with extra crypto karma.
Netsync changes:
- Command line syntax for 'serve' changed; administrators WILL
have to adjust scripts.
monotone serve my.host.com "*"
becomes
monotone serve --bind=my.host.com "*"
or simply
monotone serve "*"
(to serve on the default port, on all interfaces).
- Speaking of which, we can now bind to all interfaces; run
'serve' without passing --bind, or with passing
--bind=:port, and monotone will listen on all interfaces.
- New option '--key-to-push' for 'push', 'sync', allows
administrator to push a new user's public key into a running
server without restarting it.
- Netsync permission hooks have new defaults that read a
description of allowed access out of a standard,
basic_io-based textfile (the same stanza-based format that
revisions use). Current hooks will continue to work, but
users may prefer to transition to this format; see manual
for details.
- Between these, it is now straightforward to change
permissions and add users without restarting your server.
- Improvements to experimental "usher" facility.
UI improvements:
- New convenience options "add --unknown", "drop --missing",
"revert --missing" do what you'd expect -- add all
non-ignored non-versioned files, drop all
deleted-but-undropped files, and restore all
deleted-but-undropped files, respectively.
- New selector "h:" to select heads of a branch. "h:" means
heads of current branch, "h:mybranch" means heads of
mybranch.
- Similarly, "b:" selector with no argument now refers to
current branch.
- Commit messages now have a blank line at the top so you can
start typing directly.
- No more obscure error messages when multiple monotone
processes attempt to access a single database at the same
time; we now fail early with a more sensible error message.
(Concurrent access has never caused database corruption;
this simply makes the corruption prevention less frustrating
for the user.)
- New handlers for SIGTERM, SIGINT to rollback database
transactions. Not visible to users (unless you're really
looking carefully). (Again, killing monotone has never been
able to cause database corruption; this simply causes the
transactions to be rolled back immediately, rather than the
next time monotone runs, which improves robustness in some
theoretical way.)
Changes in 'automate':
- New command 'automate keys' to get information on existing
keys in basic_io format.
Updated translations:
- fr
Smaller changes:
- Improved handling of multibyte characters in message
displays.
- Fixes to Botan's memory allocator, to avoid pathological
slowdowns in some rare cases.
- Fix bug in delta-storage code; we were not being as aggressive
about delta-compressing files and manifests as we should
have been.
- Minor bugs fixed, error messages improved.
- Upgrading from 0.23: You must run 'db migrate' and
provide your password, for each database.
the appropriate tool via USE_TOOLS (usually "gs:run"), and remove
ghostscript.mk. This change removes a rather out-dated file from
pkgsrc and switches packages to use the more compact implementation
of the Ghostcript-handling inside the tools framework.
libgsf 1.13.3
J.H.M. Dassen (Ray) :
* Improve documentation.
Jody :
* More work on extension interface to sax import wrapper.
--------------------------------------------------------------------------
libgsf 1.13.0
Jody:
* API changes SAX wrappers for help OpenDoc and SpreadsheetML.
Kasal:
* Improve libtool versioning.
* Relax the gconf requirement.
Luciano Wolf:
* OpenDocument meta stream parser.
Sven Herzberg:
* Fix for GsfOutputGnomeVFS
API change with shlib major change -> BUILDLINK_DEPENDS bumped.
Changes between 1.3.1 and 1.4.0 (12-Sep-2003 to 02-Sep-2005)
*) Optimize insertion of free memory chunks by merging with
the spare area if possible to decrease memory fragmentation.
*) Fix mm_realloc() function: If the memory chunk passed to mm_realloc()
can't be extended and a new chunk must be allocated, the old memory
is copied into the new chunk with a call to memcpy(3). However, the
used size is the length of the new data and will cause memcpy(3) to
access memory beyond the old data chunk's boundaries.
[Kirk Petersen]
*) Upgraded build environment to GNU Libtool 1.5.20 and GNU shtool 2.0.2
[Ralf S. Engelschall]
*) Updated all copyright messages for year 2005.
[Ralf S. Engelschall]
2.103 Wed Sep 28 2005
- Removed Test::Inline::IO::File::VCS. It is incompleted, and never
should have been distributed in the dist.
2.102 Thu Sep 22 2005
- To handle people with problems installing RT and a few other
programs, add a dependency on Pod::Tests
2.101 Tue Sep 13 2005
- Rebuilding to synchronize with new Module::Install and
ExtUtils::AutoInstall
- Tweaks the per-section test code to make more applicable value
appear in the failure summary.
- Removed all remaining use of UNIVERSAL::isa, switching to a
more accurate Params::Util approach. Should allow for improved
extensibility.
2.100 Thu Jul 14 2005
- Worked as good as I hoped, moving to production release
- Bundling a couple of build-time-only deps
2.099_01 Tue Jul 12 2005
- More work on making inline2test just DWYM
- Added Test::Inline::Content::Simple
2.099_01 Sun Jul 10 2005
- Starting to implement the 2.100 roadmap
- Moved Test::Inline::Handler::Extract to Test::Inline::Extract
- Moved Test::Inline::Handler::File to Test::Inline::IO::File
- Moved Test::Inline::Handler::File::VCS to Test::Inline::IO::File::VCS
- Convert inline2test to use a config file
- Added Test::Inline::Content
- Added Test::Inline::Content::Legacy
- Added Test::Inline::Content::Default
- Reorganised code to make it all work again
- Added 12_content.t
1.101 Mon Oct 10 2005
- It appears I overtightened the naming rules. Reversed
1.100 Mon Oct 10 2005
- Significant version bump to announce addition of major functionality
- Added Algorithm::Dependency::Weight
- Upgraded a few remaining things to Params::Util
1.04 Wed Sep 14 2005
- Converted to Module::Install
- Added Algorithm::Dependency::Source::HoA by request
- Modernised the layout a little
- Upgraded from UNIVERSAL::isa to Params::Util to support more
complex uses of objects like Adapaters and Decorators.
1.15 Thurs May 26, 2005
- added optional use of weakened parent references
and improved the old circular reference DESTROY
model to make more sense. See the documantation
for more info.
- fixed bug in the fixDepth() function
0.21
- Most of build_requires should have been in requires, which was causing
CPANPLUS to choke on installs. Fixed (thanks Jos I. Boumans)
- Test names now called test descriptions to fit in with latest TAP style
- Added link to tada list to TO DO section of documentation
- Added COMMUNITY section to POD
- Added description of how to use Test::Exception in a sub-passing non
prototype style (after feedback from Jim Keenan & Perrin)
1.17 -- Mon May 23 08:55:26 CDT 2005
Bug Fixes
* Update XS code to declare PERL_UNUSED_DECL conditionally
1.16 -- Fri May 20 10:22:49 CDT 2005
Bug Fixes
* Change to refaddr.t test to avoid false errors on some 64 bit platforms
* Fix all perl only tests to work when in the core build environment
* Fix looks like number test to work for 5.8.5 and above
1.15 -- Fri May 13 11:01:15 CDT 2005
Bug Fixes
* Fixed memory leak in first()
Enhancements
* Converted tests to use Test::More
* Improved test coverage
* Changed Makefile.PL to use Module::Install
* Refactor use of Sv..X() macros to be Sv.._set()
* Changes from Jarkko for Symbian port of Perl
* Documentation updates to weaken()
0.78 Jul 19, 2005
- If an overloaded object returned false in boolean context, then it
would always fail "can" tests. Patch by Chi-Fung Fan.
0.77 Apr 29, 2005
- Neither --xs or --pm worked with the Makefile.PL the way they were
supposed to. Reported by Doug Treder.
- Moved source to my personal SVN repo.
https://svn.urth.org/svn/Params-Validate
- Updated ppport.h, which allows XS version to work with 5.00504.
This was broken since 0.75, at least.
1.22 Sep 30, 2005
- Added Exception::Class::Base->caught() so you can write
"My::Error->caught()". Apparently this is what Damian documented in
Perl Best Practices. Thanks to JD Hedden for pointing this out.
- Require Devel::StackTrace 1.12 to fix a test failure on Win32.
Reported by Garrett Goebel.
1.21 Apr 17, 2005
- Added Exception::Class->caught() as syntactic sugar for catching
exceptions in a "safe" manner. See the docs for details. Suggested
by Damian Conway.
0.04 Sat Sep 24 12:36:56 UTC 2005
- Tony Bowden now maintainer
- Document how to value when creating data
- Complete rewrite of tests
0.03 Tue Mar 11 18:30:01 GMT 2003
- Rearranged the docs a smidge.
- Added Japanese docs from perldocjp (thanks Atsuhi Kato)
- mk_classdata() is now explicitly only a class method
- Added this change log.
0.09 Thu Jun 30 15:02:37 BST 2005
- Fix a bug reported by Damian: want doesn't work (crashes) if
it's called from within the guard of a loop. See the comment
above upcontext_plus in Want.xs.
- Runs under the debugger!
- Give an error message (rather than segfaulting) if called from
a tie handler.
"Params::Util" provides a basic set of importable functions that
makes checking parameters a hell of a lot easier.
The functions provided by "Params::Util" check in the most strictly
correct manner, and in should not be fooled by odd cases.
Test::Deep gives you very flexible ways to check that the result
you got is the result you were expecting. At its simplest it compares
two structures by going through each level, ensuring that the values
match, that arrays and hashes have the same elements and that
references are blessed into the correct class. It also handles
circular data structures without getting caught in an infinite
loop.
Where it becomes more interesting is in allowing you to do something
besides simple exact comparisons. With strings, the = operator
checks that 2 strings are exactly equal but sometimes that's not
what you want. When you don't know exactly what the string should
be but you do know some things about how it should look, = is no
good and you must use pattern matching instead. Test::Deep provides
pattern matching for complex data structures.
In general, your tests shouldn't produce warnings. This modules
causes any warnings to be captured and stored. It automatically
adds an extra test that will run when your script ends to check
that there were no warnings. If there were any warnings, the test
will give a "not ok" and diagnostics of where, when, and what the
warning was, including a stack trace of what was going on when it
occurred.
If some of your tests are supposed to produce warnings then you
should be capturing and checking them with Test::Warn, that way
Test::NoWarnings will not see them and so not complain.
1.21 Wed Sep 28 2005
- Documenting ->autouse(@classes)
- Tidying up the POD a little to refer to features
in an order much more relevant to the reader.
1.20 Tue Aug 18 2005
- Catch and rethrow all error when autoloading during a ->can.
- Debugging can now only be done at compile time, and thus is
optimised out. The :debug flag is removed and instead you need
to preset $Class::Autouse::DEBUG.
Given that this was only for my benefit in the first place, it's
not appropriate to slow down everything else because of it.
1.18 Tue Jul 26 2005
- Added :nostat module to disable the initial stat checks
of modules.
1.00 Thu Sep 8 2005
- Has proven itself stable, promoting to 1.00
- Various small POD clean ups
- Adding default notifier to auto-compile the CGI functions
- Rebuilding with a newer (better) Module::Install
1.22 - Sun Oct 23 23:45:51 CDT 2005
[ENHANCEMENTS]
* Added check for lines with only whitespace before or after a
directive. Older POD formatters can get confused by such
lines since they goof up paragraph mode. (RT #6467)
[FIXES]
* Bumped up requirements to Test::Simple 0.62.
1.20 - Wed Jun 23 00:28:35 CDT 2004
[ENHANCEMENTS]
* Looks in blib/ if there is one, otherwise looks in lib/
* Thanks to David Wheeler for nudging.
Package change: depend on p5-Test-Simple for testing instead
of p5-Test-Builder-Tester, the latter is now included in the
former.
== Changes for 0.63 - September 12, 2005
* Added some text in the POD's synopsis regarding --defaultdeps and the
environment variable.
== Changes for 0.62 - July 9, 2005
* With CPANPLUS, we no longer probe $Config{sitelib} for writability
and suggest `sudo` for the entire installation process, as CPANPLUS
can handle this with `sudo` on `make install`. Reported by Sungo.
1.13 Wed Sep 28 2005
- Fixed a minor POD bug in the synopsis
1.12 Fri Sep 9 2005
- Added a fix for classes with insanely broken ->isa methods
that cause Perl to die.
1.11 Tue Sep 6 2005
- It occured to me after I added ->find that what it _really_ is
is a way to find all the subclasses, but then include the class
itself in the returned list. This method makes much more sense
it I don't return the class itself, and rename it ->subclasses
- Fixed broken Makefile.PL
1.10 Mon Sep 5 2005
- Added the ->find method
- Cleaned up and reorganised the POD
- Made sure all return conditions are documented properly
- Converted to Module::Install
1.09 skipped
