Changes:
* Improve stability.
* HEADS UP: Compression
0.22.0 has a serious compression bug. If you use old versions,
please upgrade 0.23.0 or higher. If 0.23.0 detects the remote host
is 0.22.0, fallback to no compression mode forcibly.
* Fix many bugs.
- The option to create new displays `left-to-right' is restored upon startup.
- Fortran struct members are accessed as `A%B' instead of `A.B'.
- Structure changes in plotted displays no longer crash DDD.
- You can specify the top-level geometry via `-geometry'.
- The `pydb' debugger has been fixed to work with recent Python releases.
- Ornamented Perl debugger prompts are handled correctly.
- DDD honors the TMPDIR environment variable for creating temporary files.
- `config.guess' and `config.sub' scripts have been updated.
- Minor compilation problems with GCC 2.7.2 have been fixed.
- Minor compilation problems with GCC 3.0 (prerelease) have been fixed.
- Minor compilation problems with Sun CC have been fixed.
- Minor potential problems with 64 bit pointers have been fixed.
Patch provided by Marc Recht in PR pkg/21770.
Convert -march=athlon for older gccs to -march=pentiumpro.
Addresses PR pkg/21766 by Marc Recht.
XXX: the conversion mechanism is suboptimal, it can't handle recursive
conversions.
* Fix toolbars to be consistent (Rodney Dawes)
* Fix gnome-segv to not loop when it crashes itself (Malcolm Tredinnick,
Damon Chaplin)
* Make fonts in icon labels respect font size changes (Pasupathi Duraisamy)
* Skip hidden pages in the druid (Ricardo Fernández Pascual)
* Fix referencing problem with MDI (Jaka Mocnic)
* Fix leaks in session management and score handling (Kjartan Maraas,
Arjan Van de Ven)
* Bugfix for gnome-icon-theme (Gustavo Giraldez)
* Don't crash when double clicking the OK button in the fileselector
(Narayana Pattipati)
* API documentation fixes (Sebastian Rittau)
* Don't return FALSE in pointer functions (Morten Welinder)
* Fix compiler warnings and uninitialised vars (Malcolm Tredinnick)
* Ask gdk-pixbuf for supported mime types in gnome-icon-lookup.c
(Anders Carlsson, Christian Schaller)
* Fix leak in the font picker (Kjartan Maraas)
* Fix top-left text for icons in nautilus (Alex Larsson)
* Fix the file entry to use utf8 when setting the selected file name
(Hidetoshi Tajima)
* Fix a crash in the icon entry (Balamurali Viswanathan)
* Don't open an infinite amount of preview widgets in the icon entry
* Enable color palette in the color picker (Gustavo Carneiro)
* Clear entry items properly in gnome-entry.c (Marco Pesenti Gritti)
* Fix problems with python-based apps and non-initialized memory
(Marcus at freebsd org)
Changes:
- If an odd number of parameters was given to validate() in the array
as its first argument, the error given would be different depending on
whether or not the pure Perl or XS version of the code was being used.
- Fixed incredibly odd bug that occurred in XS code when tainting was
on with Perl 5.00503. Unfortunately, the only test case that reliably
reproduces this is one of Mason's tests, so no test was added.
- Fix some compiler warnings from MS Visual Studio 6. Reported by Ron
Hill.
- When called from the main body of a script, the validation routines
would cause an undefined value warning if validation failed. Reported
by Britton Kerin and fixed by Ilya Martynov.
- The XS code that handled callbacks expects all callbacks to return
an integer, instead of simply testing the returned value for truth.
This could cause strange warnings like 'Argument "InMemory" isn't
numeric in subroutine entry...'. Based on a bug report from Robert
Dick for another module I wrote (Lingua::ZH::CCDICT).
- Fixed Params::Validate for Perl configurations where sizeof(IV) !=
sizeof(int). Reported by Alain Barbet.
- Something around test 5 in 06-options.t appear to cause Perl to
die/segfault/something under 5.6.0. These tests will be skipped with
5.6.0 in the future. Reported by Christian Schaffner.
- When testing for a compiler, use nmake on Win32. Implemented by
Ronald Hill.
- Added regex validation option.
- Added copyright info to each source file. Added LICENSE file to
distro.
- Added an XS implementation, written by Ilya Martynov. The pure Perl
implementation is still included as a fallback for users who do not
have a compiler handy. The XS implementation gives a speed boost of
150% to 300%, depending on the complexity of the validation being
done.
This XS code is known to work Perl 5.00503+, but it may not work with
earlier versions of Perl.
- Fix Attribute::Params::Validate docs to show that attribute
declaration must have open paren immediately after "Validate" or
"ValidatePos". Reported by Britton Kerin.
- Fix bug with Perl 5.8.0 when in PERL_NO_VALIDATION mode and calling
validate() expecting a hash ref to be returned. Reported by Randal
Schwartz.
- Argh. In 0.22 some of the tests printed their test counts (1..x)
after doing the tests. This works with newer versions of
Test::Harness so I didn't notice it. Reported by Peter Asemann.
- The last version's tarball was a mess, and had blib and other junk
in it. This is a nice clean one.
- The various validation functions are now context sensitive, and will
return a reference in scalar context. This may be a helpful speedup,
especially for large parameter lists.
- More optimization work.
-- Improved the speed of both validate() and validate_pos() by about
20-25%.
-- Improved the speed of the 'no validation' mode for validate() by
over 30%, for validation_pos() by 15%, and for validate_with() by
125%.
-- The speed of validate_with() has been improved by over 100%, but
this was at the expense of not validating its own incoming arguments,
thus making it a little more fragile.
- The PERL_NO_VALIDATION env var is now _only_ checked when the module
is loaded. This means you cannot change it mid-program. This is a
backwards incompatibility.
- Added a new function, validate_with(), that allow you to set various
options on a per-invocation basis, rather than on a per-package basis.
Patch by Ken Williams.
- Fix a doc nit where I had a wrong code example. Reported by Britton
Kerin.
- Added a new validation option parameter, stack_skip, which allows
you to change how errors are reported by Params::Validate.
- Improve skipping of tests when run with 5.6.0. Patch by Ken
Williams.
- Silence an uninit value warning
- [#17940] Includes :unique and :shared in the builtin types
- From perl 5.8 { __CALLER__::foo => __PACKAGE } is missparsed,
the proper approach is to use { '__CALLER__::foo' => __PACKAGE }.
The documentation is updated to reflect this.
Reported by Dave Cross
Changes :
- Fixed circular depency in the test suite. Thanks, Rob Brown.
- Added test for $Test::Harness::Switches patch below.
- Fixed straps not respecting $Test::Harness::Switches. Thanks
to Miyagawa for the patch.
- Added t/pod.t to test POD validity.
- Handed over to Andy Lester for further maintenance.
- Stas Bekman noticed that tests with no output at all were
interpreted as passing
- Abigail and Nick Clark both hit the 100000 "huge test that will
suck up all your memory" limit with legit tests. Made the check
smarter to allow large, planned tests to work.
- Partial fix of stats display when a test fails only because there's
too many tests.
- Made wait.ph and WCOREDUMP anti-vommit protection more robust in
cases where wait.ph loads but WCOREDUMP() pukes when run.
- Added a LICENSE.
- Ilya noticed the per test skip reason was accumlating between tests.
Package changes:
Set test target (must not be run as root).
Get ride of TEXINFO_OVERRIDE (yeah one down!).
New in CSSC-0.15alpha-pl0
* Reduced the number of lines of code not tested by the test suite to
2083.
* Support for ignored deltas
* Added section in the CSSC manual on included, excluded and
ignored deltas. Various other minor improvements to manual,
including a clearer description of CSSC's interoperability
support, and how it treats long lines in its input.
* sccs get - all g-files processed after a failing file are deleted.
* Fixed compilation bug - we were using idigit() in reveral places but
not including <ctype.h>. (SourceForge bug 623164).
* The test suite now passes when run against a genuine version of
SCCS which lacks "prt".
* Support for the "-a" option of "get" and the "-p" option of "delta".
* We detect the situation if an SCCS file has a link count
greater than 1, and we refuse to open the affected file.
* Fix cssc.spec file so that it works with RPM v4 (i.e. when all
manpages are automatically compressed).
* Correction to support for MR validation on systems lacking
"fork()".
* Upgraded to autoconf version 2.53.
New in CSSC-0.14alpha-pl0
* Added Richard Kettlewell's sccs.cgi CGI script.
* Enhancements to the manual: added a Glossary, documented "val".
* Include a section in the manual explaining in more detail some of
the differences in between other SCCS versions (and which
funcitonality CSSC follows in each case).
* More test cases (we are now tracking test suite
coverage with "gcov" and devising test cases to
test the parts of the code that the test suite does not
currently reach).
* Bug-fixes for "val" (mainly corrections to the precise
value returned when validation fails)
* delta now checks that the caller is actually allowed to make
deltas to he SCCS history file (previously this check was made only
by "get -e")
* The "!44" syntax for the authorised user list is not compatible
with SCCS and hence has been disabled. Any complaints via the
usual bug reporting route, please.
* The options -Y and -M for "delta" have been removed (since they
aren't in normal SCCS and can be replaced by -y"" and -m"").
* "sccs help" and "sccs comb" now issue more helpful error
messages explaining why the relevant program is missing.
* Corrections to "what" - exit with status 1 if no matching
string was found or an unknown option letter was passed.
* Fixes for set-user-id operation (which is still not
recommended - see the CSSC manual for an explanation).
* Added warning when we see an excluded delta, explaining that
this feature is not fully tested. The warning includes a pointer
to a document explaining what they should do about this.
* Fixed bugs from SourceForge:
479916 (get -t problem),
480136, 481519, 481707, 484278 (all related to setuid
operation, and in fact all duplicates of each other),
489005 (removal of g-file when running setuid),
513800 (cope with Data General "bugfix" 4-digit years),
489311 (Locale settings cause test failure),
530969 (FILE* leak in filediff.cc),
531884 (compiler warning on 64-bit systems)
* Moved old ChangeLog to ChangLog.2.
ChangeLog:
* New keyword: compress (client only)
* Options (client)
-Z : Disables a compression.
-z : Enables a compression.
* Option (server)
-z: Sets the compression level.
0: no compression.
1: the fastest, but less compression.
9: the slowest, but best compression.
libstdc++ in gcc3.
when defined, USE_GCC_SHLIB ensures that the correct rpath is passed
to the linker, and a full dependency on the compiler package is
registered.
packages which define USE_GCC_SHLIB should not include
mk/gcc.buildlink2.mk (or gcc{,3}/buildlink2.mk) as it is handled
automatically.
- Add some additional x86 flags for gcc 3.x
- Allow the path to cc to be specified as an argument to cpuflags
- Use both the previous to handle the gcc3 package
- Add an 'unsupported' optimize_gcc.mk which... to quote its comments:
"This file is 'experimental' - which is doublespeak for unspeakably
ugly, and probably quite broken by design.
The intention is to pass additional flags to gcc to further optimise
generated code. It _will_ make it impossible to debug, may fail to
compile some code, and even generate curdled binaries. It is completely
unsupported. Any questions should be directed to <abs@netbsd.org>.
This package contains the Python bindings to the Subversion libraries
and some programs that use them. This includes cvs2svn, which
converts CVS repositories to Subversion repositories.
This package contains all the Subversion programs and libraries except
the Apache module mod_dav_svn (for which see the ap2-subversion
package) and the Python bindings and some programs which use them (for
which see the subversion-python package). Installing this package
alone will satisfy most people's needs, providing all three repository
access layers (local, svn, and dav) and the server for the svn access
layer.
This package contains apr and apr-util from the latest Apache2
distribution (currently httpd-2.0.45). It is currently (er, currently
in a few minutes) shared between the Apache2 and Subversion packages.
This package contains only the swig program itself and the base
definitions; it is useless alone. You only need this package if you
are building software that uses SWIG. You will also need a swig-LANG
package, where LANG is the language you are working with
(e.g. swig-perl, swig-python, etc.).
Still todo: import other swig runtimes, move existing swig package to
swig11 and update dependencies, and add the new swig packages to
devel/Makefile.
Library:
* Init gettext earlier to catch some bugs (Hidetoshi Tajima)
Bug #99224
* Fix search path for locale.alias. (Julio Merino) Bug #106180.
* Fix url substitution with custom web browser command
(Pasupathi Duraisamy) Bug #108909
* Plug a leak in the help routines. (Kjartan)
Help converters:
* Fix manual pages that have non-numeric filetypes such as
open.3pm. Patch from Roderich Schupp. Bug #101440.
* Fix crashes when feeding directories to the help converters.
Patch from John Fleck. Bug #104056.
* Generate help-methods.conf based on libexecdir value.
Patch from Frédéric Crozat. Bug #91169.
Build:
* Honor --disable-schemas-install. (Julio Merino)
* Require newer intltool (Kjartan)
Translations:
az (Metin Amiroff), be (Dmitry Mastrukov), bn (Taneem Ahmed),
cs (Miloslav Trmac), de (Christian Neumair), el (Kostas
Papadimas),
es (Carlos Perello Martin), id (Mohammad DAMT), ja (Takeshi
AIHANA),
ml (Guntupalli Karunkar), mn (Badral Saulig), ms (Hasbullah Bin
Pit),
ko (Changwoo Ryu), nl (Vincent Van Adrighem), no (Kjartan
Maraas),
pl (Zbigniew Chyla), pt (Duarte Loreto), ru (Dmitry Mastrukov)
- GC_invoke_finalizers could, under rare conditions, set
GC_finalizer_mem_freed to an essentially random value. This could
possibly cause unbounded heap growth for long-running applications
under some conditions. (The bug was introduced in 6.1alpha5, and
is not in gcc3.3. Thanks to Ben Hutchings for finding it.)
- Attempted to sanitize the various DLL macros. GC_USE_DLL disappeared.
GC_DLL is used instead. All internal tests are now on GC_DLL.
README.macros is now more precise about the intended meaning.
- Include DllMain in the multithreaded win32 version only if the
collector is actually built as a dll. (Thanks to Mohan Embar for
a version of the patch.)
- Hide the cygwin threadAttach/Detach functions. They were violating our
namespace rules.
- Fixed an assertion in GC_check_heap_proc. Added GC_STATIC_ASSERT.
(Thanks again to Ben Hutchings.)
- Removed some obsolete definitions for Linux/PowerPC in gcconfig.h.
- CORD_cat was not rebalancing unbalanced trees in some cases, violating
a CORD invariant. Also tweaked the rebalancing rule for
CORD_cat_char_star. (Thanks to Alexandr Petrosian for the bug report
and patch.)
- Added hand-coded structured exception handling support to mark.c.
This should enable support of dynamic libraries under win32 with
gcc-compiled code. (Thanks to Ranjit Mathew for the patch.)
Turned on dynamic library scanning for win32/gcc.
- Removed some remnants of read wrapping. (Thanks to Kenneth Schalk.)
GC_USE_LD_WRAP ws probably broken in recent versions.
- The build could fail on some platforms since gcconfig.h could include
declarations mentioning ptr_t, which was not defined, e.g. when if_mach
was built. (Thanks to Yann Dirson for pointing this out.) Also
cleaned up tests for GC_PRIVATE_H in gcconfig.h a bit.
- The GC_LOOP_ON_ABORT environment variable interfered with incremental
collection, since the write fault handler was erroneously overridden.
Handlers are now set up in the correct order.
- It used to be possible to call GC_mark_thread_local_free_lists() while
the world was not stopped during an incremental GC. This was not safe.
Fortunately, it was also unnecessary. Added GC_world_stopped flag
to avoid it. (This caused occasional crashes in GC_set_fl_marks
with thread local allocation and incremental GC. This probably happened
primarily on old, slow multiprocessors.)
- Allowed overriding of MAX_THREADS in win32_threads.c from the build
command line. (Patch from Yannis Bres.)
- Taught the IA64/linux code to determine the register backing store base from
/proc/self/maps after checking the __libc symbol, but before guessing.
(__libc symbols are on the endangered list, and the guess is likely to not
always be right for 2.6 kernels.) Restructured the code to read and parse
/proc/self/maps so it only exists in one place (all platforms).
- The -DUSE_PROC_FOR_LIBRARIES code was broken on Linux. It claimed that it
also registered the main data segment, but didn't actually do so. (I don't
think anyone actually uses this configuration, but ...)
- Made another attempt to get --enablecplusplus to do the right thing.
Since there are unavoidable problems with C programs linking against a
dynamic library that includes C++ code, I separated out the c++ code into
libgccpp.
Based on patch provided in PR pkg/21569 by Marc Recht.
