Age | Commit message (Collapse) | Author | Files | Lines |
|
|
|
|
|
|
|
Version 0.9.4
http://svn.edgewall.org/repos/babel/tags/0.9.4/
(Aug 25 2008, from branches/stable/0.9.x)
* Currency symbol definitions that is defined with choice patterns in the
CLDR data are no longer imported, so the symbol code will be used instead.
* Fixed quarter support in date formatting.
* Fixed a serious memory leak that was introduces by the support for CLDR
aliases in 0.9.3 (ticket #128).
* Locale modifiers such as "@euro" are now stripped from locale identifiers
when parsing (ticket #136).
* The system locales "C" and "POSIX" are now treated as aliases for
"en_US_POSIX", for which the CLDR provides the appropriate data. Thanks to
Manlio Perillo for the suggestion.
* Fixed JavaScript extraction for regular expression literals (ticket #138)
and concatenated strings.
* The `Translation` class in `babel.support` can now manage catalogs with
different message domains, and exposes the family of `d*gettext` functions
(ticket #137).
|
|
|
|
Changes:
2008-10-20 Mikio Hirabayashi
* tcqdb.c (tcidsetnew, tcidbsetdel): performance was improved by using mmap.
- Release: 0.9.8
|
|
|
|
Changes:
2008-10-19 Mikio Hirabayashi
* myconf.h: missing features of PATH_MAX and nan is now emulated.
* tcutil.c (tczeromap, tczerounmap): new functions.
- Release: 1.3.14
|
|
|
|
|
|
* A huge set of patches for modules and plugins, many thanks to all
the contributors (hope to list them all): cdmack, Viruzzo,
Lance, maxadamo, nvhs, russell822, Andre Martin, Daniel Black,
Picov, RISKO Gergely, blackmoon and Spil
* A new zenity-based (GNOME eye-candy) updater interface by maxadamo
* Internationalization support (and Italian localization) for the
regular fltk updater by blackmoon
* Updated versions of the cURL and gnutls libraries for Windows
* The luasocket library is now part of freepops and can be used to
write modules
We still avoid installing any of the self-updaters. Might be nice
to add in a future PKGREVISION.
|
|
* aor/: aor.c, aor.h, ar2700.c, ar5000.c, ar8000.c, ar8200.c,
ar8600.c: implemented scanning
* include/hamlib/riglist.h: added FT950
* set_ant/get_ant support for ts870s and ts2000
* kenwood/ts2000.c: S meter calibration
|
|
* aor/: aor.c, aor.h, ar2700.c, ar5000.c, ar8000.c, ar8200.c,
ar8600.c: implemented scanning
* include/hamlib/riglist.h: added FT950
* set_ant/get_ant support for ts870s and ts2000
* kenwood/ts2000.c: S meter calibration
|
|
|
|
|
|
|
|
Changes since old version 2.4:
- Fix -da option.
- Removed some debug printf which were accidentally left in.
- Avoid duplicating exif header on some commands.
- Fix up return codes.
- Preserve file permissions and time when files are modified.
- Read ISO euqivalnt and white balance from canon makernote.
- Added -rt (replace thumbnail) feature, and rotate the thumbnail also
when using the -autorot feature.
- Added -rgt (regenerate thumbnail) feature.
- Added -orp and -orl options.
- Fix typecast issue run itno with GCC 4.
- Fix shutter speed display in '-c' mode for very long shutter speeds.
- Fix some nitpicks from Debian folks.
- Fix a bug in autorot when rotating filenames with spaces in them.
- Improved handling of corrupt exif linkages in exif header.
- Added -a (rename associated files) options.
- Remove maximum jpeg sections limit.
- Added -ds option.
- On clearing rotation, clear the image and the optinoal thumbnail rotation
tags (some viewers use the wrong tag).
- Add -mkexif option to make a new exif header.
- Added IPTC handling.
- Added -q option.
- Fix handling of corrupted GPS directory.
- Extract focus distance from canon makernote.
- Extract subject range (pentax and fuji cameras).
- Fix it so it no longer deletex XMP sections.
- Improve IPTC handling a little.
- Change how date is encoded with -mkexif section to make it more compatible.
- Make jhead aware of XMP data and not delete it.
- Decode more exif tags for '-v' mode.
- Fix a bunch of potential string overflows.
- Fix bug where IPTC sction was not deleted by -purejpg.
- Fix GPS altitude decode bug.
|
|
|
|
and CVE-2008-3641. Also, it fixes a ton of bugs and has portability
enhancements. Full list of changes:
- SECURITY: The HP-GL/2 filter did not range check pen numbers
(STR #2911)
- SECURITY: The SGI image file reader did not range check
16-bit run lengths (STR #2918)
- SECURITY: The text filter did not range check cpi, lpi, or
column values (STR #2919)
- Documentation updates (STR #2904, STR #2944)
- The French web admin page was never updated (STR #2963)
- The IPP backend did not retry print jobs when the printer
reported itself as busy or unavailable (STR #2951)
- The "Set Allowed Users" web interface did not handle trailing
whitespace correctly (STR #2956)
- The PostScript filter did not work with Adobe applications
using custom page sizes (STR #2968)
- The Mac OS X USB backend did not work with some printers
that reported a bad 1284 device ID.
- The scheduler incorrectly resolved the client connection
address when HostNameLookups was set to Off (STR #2946)
- The IPP backend incorrectly stopped the local queue if
the remote server reported the "paused" state.
- The cupsGetDests() function did not catch all types of
request errors.
- The scheduler did not always log "job queued" messages
(STR #2943)
- The scheduler did not support destination filtering using
the printer-location attribute properly (STR #2945)
- The scheduler did not send the server-started,
server-restarted, or server-stopped events (STR #2927)
- The scheduler no longer enforces configuration file
permissions on symlinked files (STR #2937)
- CUPS now reinitializes the DNS resolver on failures
(STR #2920)
- The CUPS desktop menu item was broken (STR #2924)
- The PPD parser was too strict about missing keyword
values in "relaxed" mode.
- The PostScript filter incorrectly mirrored landscape
documents.
- The scheduler did not correctly update the
auth-info-required value(s) if the AuthType was Default.
- The scheduler required Kerberos authentication for
all operations on remote Kerberized printers instead
of just for the operations that needed it.
- The socket backend could wait indefinitely for back-
channel data with some devices.
- PJL panel messages were not reset correctly on older
printers (STR #2909)
- cupsfilter used the wrong default path (STR #2908)
- Fixed address matching for "BrowseAddress @IF(name)"
(STR #2910)
- Fixed compiles on AIX.
- Firefox 3 did not work with the CUPS web interface in SSL
mode (STR #2892)
- Custom options with multiple parameters were not emitted
correctly.
- Refined the cupstestppd utility.
- ppdEmit*() did not support custom JCL options (STR #2889)
- The cupstestppd utility incorrectly reported missing
"en" base translations (STR #2887)
- Documentation updates (STR #2785, STR #2861, STR #2862)
- The scheduler did not add the ending job sheet when the
job was released.
- The IPP backend did not relay marker-* attributes.
- The CUPS GNOME/KDE menu item was not localized for
Chinese (STR #2880)
- The CUPS GNOME/KDE menu item was not localized for
Japanese (STR #2876)
- The cupstestppd utility reported mixed line endings for
Mac OS and Windows PPD files (STR #2874)
- The pdftops filter did not print landscape orientation PDF
pages correctly on all printers (STR #2850)
- The scheduler did not handle expiring of implicit classes
or their members properly, leading to a configuration where
one of the members would have a short name (STR #2766)
- The scheduler and cupstestppd utilities did not support
cupsFilter and cupsPreFilter programs with spaces in their
names (STR #2866)
- Removed unused variables and assignments found by the
LLVM "clang" tool.
- Added NULL checks recommended by the LLVM "clang" tool.
- The scheduler would crash if you started a printer that
pointed to a backend that did not exist (STR #2865)
- The ppdLocalize functions incorrectly mapped all generic
locales to country-specific locales.
- The cups-driverd program did not support Simplified Chinese
or Traditional Chinese language version strings (STR #2851)
- Added an Indonesian translation (STR #2792)
- Fixed a timing issue in the backends that could cause data
corruption with the CUPS_SC_CMD_DRAIN_OUTPUT side-channel
command (STR #2858)
- The scheduler did not support "HostNameLookups" with all of
the boolean names (STR #2861)
- Fixed a compile problem with glibc 2.8 (STR #2860)
- The PostScript filter did not support %%IncludeFeature lines
in the page setup section of each page (STR #2831)
- The scheduler did not generate printer-state events when the
default printer was changed (STR #2764)
- cupstestppd incorrectly reported a warning about the PPD format
version in some locales (STR #2854)
- cupsGetPPD() and friends incorrectly returned a PPD file for
a class with no printers.
- The member-uris values for local printers in a class returned
by the scheduler did not reflect the connected hostname or
port.
- The CUPS PHP extension was not thread-safe (STR #2828)
- The scheduler incorrectly added the document-format-default
attribute to the list of "common" printer attributes, which
over time would slow down the printing system (STR #2755,
STR #2836)
- The cups-deviced and cups-driverd helper programs did not set
the CFProcessPath environment variable on Mac OS X (STR #2837)
- "lpstat -p" could report the wrong job as printing (STR #2845)
- The scheduler would crash when some cupsd.conf directives
were missing values (STR #2849)
- The web interface "move jobs" operation redirected users to
the wrong URL (STR #2815)
- The Polish web interface translation contained errors
(STR #2815)
- The scheduler did not report PostScript printer PPDs with
filters as PostScript devices.
- The scheduler did not set the job document-format attribute
for jobs submitted using Create-Job and Send-Document.
- cupsFileTell() did not work for log files opened in append
mode (STR #2810)
- The scheduler did not set QUERY_STRING all of the time
for CGI scripts (STR #2781, STR #2816)
- The scheduler now returns an error for bad job-sheets
values (STR #2775)
- Authenticated remote printing did not work over domain
sockets (STR #2750)
- The scheduler incorrectly logged errors for print filters
when a job was canceled (STR #2806, #2808)
- The scheduler no longer allows multiple RSS subscriptions
with the same URI (STR #2789)
- The scheduler now supports Kerberized printing with
multiple server names (STR #2783)
- "Satisfy any" did not work in IPP policies (STR #2782)
- The CUPS imaging library would crash with very large
images - more than 16Mx16M pixels (STR #2805)
- The PNG image loading code would crash with large images
(STR #2790)
- The scheduler did not limit the total number of filters.
- The scheduler now ensures that the RSS directory has
the correct permissions.
- The RSS notifier did not quote the feed URL in the RSS
file it created (STR #2801)
- The web interface allowed the creation and cancellation
of RSS subscriptions without a username (STR #2774)
- Increased the default MaxCopies value on Mac OS X to
9999 to match the limit imposed by the print dialog.
- The scheduler did not reject requests with an empty
Content-Length field (STR #2787)
- The scheduler did not log the current date and time and
did not escape special characters in request URIs when
logging bad requests to the access_log file (STR #2788)
|
|
|
|
* Dovecot prints an informational message about authentication problems
at startup. The message goes away after the first successful
authentication. This hopefully reduces the number of "Why doesn't
my authentication work?" questions.
+ Maildir/dbox: Try harder to assign unique UIDVALIDITY values to
mailboxes to avoid potential problems when recreating or renaming
mailboxes. The UIDVALIDITY is tracked using dovecot-uidvalidity*
files in the mail root directory.
+ Many logging improvements
- In some conditions Dovecot could have stopped using existing cache
file and never used it again until it was deleted.
- pop3 + Maildir: Make sure virtual sizes are always written to
dovecot-uidlist. This way if the indexes are lost Dovecot will never
do a huge amount of work to recalculate them.
- mbox: Fixed listing mailboxes in namespaces with prefix beginning
with '~' or '/' (i.e. UW-IMAP compatibility namespaces didn't work).
- dict quota: Don't crash when recalculating quota (when quota warnings
enabled).
- Fixes to handling "out of disk space/quota" failures.
- Blocking passdbs/userdbs (e.g. PAM, MySQL) could have failed lookups
sometimes when auth_worker_max_request_count was non-zero.
- Fixed compiling with OpenBSD
|
|
|
|
py-gobject-shared>=${PKGVERSION}, not just any old py-gobject-shared.
This should fix a problem in which py25-gobject-2.15.4 was happy to
depend on the much older py-gobject-shared-2.14.2, but other packages
that depended on py25-gobject-2.15.4 thought that it was incomplete.
Bump PKGREVISION.
|
|
version 1.25.3 or later, so adjust BUILDLINK_API_DEPENDS.pygobject
accordingly.
Not bumping PKGREVISION, because the py-gtk2 didn't build against older
versions of gobject.
|
|
into The NetBSD Packages Collection.
Releasing a new version of software (Perl module) takes a lot of
steps... finding the next version number (and making sure you didn't
already use that version number before), making sure your changelog
is updated, making sure your "make dist" results in a tarball that
builds, commiting changes (with updated version number), tagging,
and uploading the tarball somewhere. Or maybe more steps. Or not
some of the above. Maybe you forgot something! And maybe you manage
multiple projects, and each project has a different release process.
You want to be hacking, not jumping through hoops. Your contributors
want to see their patches actually make it into a release, which
won't happen if you're afraid of releases. shipit automates all
the hell. It makes life beautiful.
|
|
|
|
Collection.
Releasing a new version of software (Perl module) takes a lot of
steps... finding the next version number (and making sure you didn't
already use that version number before), making sure your changelog
is updated, making sure your "make dist" results in a tarball that
builds, commiting changes (with updated version number), tagging,
and uploading the tarball somewhere. Or maybe more steps. Or not
some of the above. Maybe you forgot something! And maybe you manage
multiple projects, and each project has a different release process.
You want to be hacking, not jumping through hoops. Your contributors
want to see their patches actually make it into a release, which
won't happen if you're afraid of releases. shipit automates all
the hell. It makes life beautiful.
|
|
|
|
|
|
ASN1-encoded files (DER, BER, PER, etc.) in Python. ASN1 is the Abstract
Syntax Notation version 1, as defined by the International Telecommunication
Union (ITU).
|
|
|
|
|
|
fix a bug that init function is deleted.
http://ruby-gnome2.svn.sourceforge.net/viewvc/ruby-gnome2/ruby-gnome2/trunk/gtk/src/makeinits.rb?r1=3336&r2=3339
Bump PKGREVISION.
|
|
|
|
|
|
Ruby library for implementing IRC server and client
|
|
|
|
Bump PKGREVISION.
|
|
|
|
deskbar-applet-2.24.1, eel-2.24.1, ekiga-3.0.1, eog-2.24.1,
epiphany-2.24.1, epiphany-extensions-2.24.1, evince-2.24.1,
evolution-2.24.1, evolution-data-server-2.24.1,
evolution-exchange-2.24.1, file-roller-2.24.1, gcalctool-5.24.1,
gconf-editor-2.24.1, geany-0.15, glibmm-2.18.1, gnome-applets-2.24.1,
gnome-games-2.24.1, gnome-system-monitor-2.24.1, gnome-themes-2.24.1,
gnome-utils-2.24.1, gsasl-0.2.29, gtk-engines-2.16.1, gtkhtml314-3.24.1,
gvfs-1.0.2, libsoup-2.24.1, nautilus-2.24.1, p5-glib2-1.200,
p5-gtk2-1.201, pango-1.22.1, ptlib-2.4.2, rdiff-backup-1.2.2,
vino-2.24.1.
|
|
|
|
== Ruby-GNOME2 0.18.0: 2008-09-30
This release fixes some memory leak bugs and has new
experimental packages.
=== Changes
Ruby/GLib:
* mkmf-gnome2.rb: add check_cairo. [Kouhei Sutou]
* fixed memory leaks. [Kouhei Sutou]
* [#2110165] Gdk::Pixbuf memory leak
* install glib-enum-types.h. [Andy Spencer]
* [#2106938] Install glib-enum-types.h
Ruby/GTK2:
* add GTK+ 2.12 related bindings. [Guillaume Cottenceau]
* fixed initialize order. [Kouhei Sutou]
* [#2113491] Incorrect initialize order (compiling with mingw)
Ruby/GooCanvas [NEW][EXPERIMENTAL]:
* added. [Kouhei Sutou]
Ruby/Pango:
* fixed memory leaks. [Kouhei Sutou]
Ruby/GtkSourceView2 [NEW][EXPERIMENTAL]:
* add. [yselkowitz]
|
|
|
|
Various bugfixes including fix building with perl 5.10.
|
|
|
|
|
|
GooCanvas is similar in many ways to GnomeCanvas and FooCanvas.
But it uses cairo for rendering, has an optional model/view split,
and uses interfaces for items & models (so you can easily turn any
application object into a canvas item or model).
|
|
|
|
- Security-related bugs in the Bluetooth ACL, Bluetooth RFCOMM, PRP, Q.931,
MATE, and USB dissectors, as well as the Tammos CommView file parser have
been fixed.
- Many other bugs have been fixed.
This update addresses the security vulnerabilies reported
in wnpa-sec-2008-06.
|
|
|