1 D-Bus 1.4.8 (2011-04-08)
4 The "It's like the beginning of a lobster" release.
6 • Rename configure.in to configure.ac, and update it to modern conventions
7 (fd.o #32245; Javier Jardón, Simon McVittie)
8 • Correctly give XDG_DATA_HOME priority over XDG_DATA_DIRS (fd.o #34496,
10 • Prevent X11 autolaunching if $DISPLAY is unset or empty, and add
11 --disable-x11-autolaunch configure option to prevent it altogether
12 in embedded environments (fd.o #19997, NB#219964; Simon McVittie)
13 • Install the documentation, and an index for Devhelp (fd.o #13495,
14 Debian #454142; Simon McVittie, Matthias Clasen)
15 • If checks are not disabled, check validity of string-like types and
16 booleans when sending them (fd.o #16338, NB#223152; Simon McVittie)
17 • Add UnknownObject, UnknownInterface, UnknownProperty and PropertyReadOnly
18 errors to dbus-shared.h (fd.o #34527, Lennart Poettering)
19 • Break up a huge conditional in config-parser so gcov can produce coverage
20 data (fd.o #10887, Simon McVittie)
21 • List which parts of the Desktop Entry specification are applicable to
22 .service files (fd.o #19159, Sven Herzberg)
23 • Don't suppress service activation if two services have the same Exec=
24 (fd.o #35750, Colin Walters)
26 · Avoid the name ELEMENT_TYPE due to namespace-pollution from winioctl.h
28 · Include _dbus_path_is_absolute in libdbus on Windows, fixing compilation
29 (fd.o #32805, Mark Brand)
31 D-Bus 1.4.6 (2010-02-17)
34 The "1, 2, miss a few, 99, 100" release.
36 • Remove unfinished changes intended to support GTest-based tests,
37 which were mistakenly included in 1.4.4
39 D-Bus 1.4.4 (2010-02-17)
42 • Switch back to using even micro versions for stable releases; 1.4.1
43 should have been called 1.4.2, so skip that version number
44 • Don't leave bad file descriptors being watched when spawning processes,
45 which could result in a busy-loop (fd.o #32992, NB#200248; possibly
46 also LP#656134, LP#680444, LP#713157)
47 • Check for MSG_NOSIGNAL correctly
48 • Fix failure to detect abstract socket support (fd.o #29895)
49 • Make _dbus_system_logv actually exit with DBUS_SYSTEM_LOG_FATAL
50 (fd.o #32262, NB#180486)
51 • Improve some error code paths (fd.o #29981, fd.o #32264, fd.o #32262,
52 fd.o #33128, fd.o #33277, fd.o #33126, NB#180486)
53 • Avoid possible symlink attacks in /tmp during compilation (fd.o #32854)
54 • Tidy up dead code (fd.o #25306, fd.o #33128, fd.o #34292, NB#180486)
55 • Improve gcc malloc annotations (fd.o #32710)
56 • If the system bus is launched via systemd, protect it from the OOM killer
57 • Documentation improvements (fd.o #11190)
58 • Avoid readdir_r, which is difficult to use correctly (fd.o #8284,
59 fd.o #15922, LP#241619)
60 • Cope with invalid files in session.d, system.d (fd.o #19186,
62 • Don't distribute generated files that embed our builddir (fd.o #30285,
64 • Raise the system bus's fd limit to be sufficient for its configuration
65 (fd.o #33474, LP#381063)
66 • Fix syslog string processing
68 • Remove broken gcov parsing code and --enable-gcov, and replace them
69 with lcov HTML reports and --enable-compiler-coverage (fd.o #10887)
71 · avoid live-lock in Windows CE due to unfair condition variables
73 · support credentials-passing (fd.o #32542)
75 · opt-in to thread safety (fd.o #33464)
77 D-Bus 1.4.1 (20 December 2010)
80 • Fix for CVE-2010-4352: sending messages with excessively-nested variants can
81 crash the bus. The existing restriction to 64-levels of nesting previously
82 only applied to the static type signature; now it also applies to dynamic
83 nesting using variants. Thanks to Rémi Denis-Courmont for discoving this
85 • OS X portability fixes, including launchd support.
86 • Windows autolaunch improvements.
89 D-Bus 1.4.0 (6 Sep 2010)
93 D-Bus 1.3.1 (23 June 2010)
95 - New standardized PropertiesChanged signal in the properties interface
96 - Various portability fixes, in particular to Windows platforms
97 - Support forking bus services, for compatibility
99 D-Bus 1.3.0 (29 July 2009)
101 - ability for dbus-send to send to any bus (--address)
102 - file descriptor passing on Unix socket transports
103 - use of GCC atomic intrinsics for better processor support
104 (requires -march=i486 or above for x86 compilation)
105 - thread-safe FD_CLOEXEC setting on recent Linux kernels (2.6.24-27 and up)
106 and glibc (2.9 for pipe2 and 2.10 for accept4)
107 - feature negotiation in the bus daemon