Here's a release of Xtrans that includes a bunch of const fixes and a
new Listen function which the X server wants to allow -nolisten tcp by default.
Alan Coopersmith (7):
Update docs to note that TransName is now const char *
Add const qualifiers to TRANS(Open...) address args
Add const qualifiers to TRANS(Reopen...) port args
Add const qualifiers to TRANS(CreateListener) port args
Add const qualifiers to TRANS(Connect) args
Add const qualifiers to TRANS(MakeAll*ServerListeners) port args
Add const qualifiers to TRANS(OpenC{L,O}TS{Server,Client}) args
Keith Packard (2):
Add TRANS(Listen) function to re-enable specific listen sockets
xtrans 1.3.5
I'm happy to announce xtrans-1.3.4
Hans de Goede (2):
configure: Also add -D_DEFAULT_SOURCE to .pc cflags to shut up glibc warnings
xtrans 1.3.4
Mark Kettenis (1):
Increase UNIX socket buffer size
Thomas Klausner (2):
Cast ctype(3) function arguments to unsigned char.
Add missing headers for free() and strlen().
Another round of fixes for the fd sending API's: improves portability,
fixes build failures on BSD & Solaris when building with the new API
disabled, and enables the new API for Solaris.
1.3.1:
Here's a new 1.3.1 release of libXtrans. The only interesting change
is that it now requires applications to specify that they want FD
passing support by defining XTRANS_SEND_FDS; the X server will do that
only on Linux. When non-Linux FD passing support is integrated, we can
change what the X server does
Keith Packard (3):
Revert "Remove 'Readv' interface"
Add SEND_FDS version of Readv
Update to version 1.3.1
Mark Kettenis (1):
Don't include file descriptor passing code by default
1.3.0:
Adam Jackson (1):
configure: Remove AM_MAINTAINER_MODE
Alan Coopersmith (3):
Remove unused TLI ("STREAMSCONN") code from xtrans
Docs: convert function synopses to docbook funcsynopsis markup
Docs: Drop pre-C89 TRANS() from docs since it's no longer in the headers
Colin Walters (1):
autogen.sh: Implement GNOME Build API
Keith Packard (3):
Remove 'Readv' interface
Add APIs to send file descriptors through the network
Update to version 1.3.0
Łukasz Stelmach (1):
Add const qualifier to unix_nolisten
1.2.7:
This release features a lot of cleanups, both from manual inspection and
from automated static code analysis, along with a couple bug fixes and
various documentation & build configuration improvments.
1.2.6:
This minor release converts the libxtrans API documentation from troff
to DocBook/XML and updates it for the current state of the code.
It also includes some minor bug fixes, and of course, the usual recent
collection of build configuration improvements and janitorial cleanups.
This release moved the xtrans.pc file from lib/pkgconfig to share/pkgconfig.
AFAICT pkgsrc deals with this properly.
Changes:
Migrate to xorg macros 1.3 & XORG_DEFAULT_OPTIONS
Move Xtrans interface docs from xorg-docs module
xtrans 1.2.5
Xtranssock: assume that we have getaddrinfo if ipv6 is enabled
Move xtrans.pc to /usr/share/pkg-config
Don't use -lws2_32 on Cygwin
Silence root ownership warnings on Cygwin
Alan Coopersmith (6):
Fix libICE build after code motion in d192bac409...
Constify string argument to is_numeric
Constify path argument to trans_mkdir()
xfs segfaults in _FontTransSocketReopen when cloning itself
Fix ifdef checks for SVR4 to do the right thing on Solaris
Version 1.2.4
Benjamin Close (2):
Revert "Fix "XXX declared 'static' but not defined""
Update is_numeric base on the declaration change in
389e01fb51ba2d708015e27d8fc17c88a0e55802
Paulo Cesar Pereira de Andrade (1):
Janitor: Correct make distcheck and compiler warnings.
Tomas Carnecky (2):
Fix "warning: format not a string literal and no format arguments"
Fix "XXX declared 'static' but not defined"
This changes the buildlink3.mk files to use an include guard for the
recursive include. The use of BUILDLINK_DEPTH, BUILDLINK_DEPENDS,
BUILDLINK_PACKAGES and BUILDLINK_ORDER is handled by a single new
variable BUILDLINK_TREE. Each buildlink3.mk file adds a pair of
enter/exit marker, which can be used to reconstruct the tree and
to determine first level includes. Avoiding := for large variables
(BUILDLINK_ORDER) speeds up parse time as += has linear complexity.
The include guard reduces system time by avoiding reading files over and
over again. For complex packages this reduces both %user and %sys time to
half of the former time.
Alan Coopersmith (4):
18748: xtrans.m4 causes configure --help to list --enable-ipv6
in wrong case.
Switch ChangeLog generation to use XORG_CHANGELOG from xorg-macros 1.2
Add bugzilla, mailing list & git repo pointers to README
Version 1.2.3
Alan Hourihane (1):
add winsock check for windows builds
Adam Jackson (5):
Massive ifdef cleanup, dropping a ton of unsupported platform code.
Drastically simplify TRANS_OPEN_MAX.
Remove DECNET support.
Finish removing OS/2 support
xtrans 1.2.2
Jeremy Huddleston (2):
Apple: Allow Xquartz to provide an additional fd to xtrans since
the dynamic addition code in the server isn't bulletproof yet.
Added a flag to enable "The OS already took care of securing this,
please skip checking xauth" for use with Apple launchd sockets.
X.Org found in NetBSD-current.
Thanks a lot to all who helped, especially Matthias Scheler who did
repeated tests on Mac OS X and older versions of NetBSD to make sure the
support for those platforms wouldn't be broken (or at least, not fatally,
as I would still expect a few hiccups here and there, because there is
only so much one can test in such limited time).
On the infrastructure side, this branch brings pkgconfig-builtin.mk, in
order to write very easily new builtin.mk files. It can actually handle
more than just pkgconfig files, but it will provide a version if it finds
such a file. x11.builtin.mk has also been made more useful and now all
existing (and future!) native-X11-related builtin.mk files should include
it.
of patch-aa).
Abstract socket support was previously patched disabled because it was broken
( see http://bugs.freedesktop.org/show_bug.cgi?id=15884 ). The fix is in two
parts.
The first part is corrected in xtrans 1.2.1: Connection failure for abstract
socket is ECONNREFUSED, not ENOENT.
The second part is in an upcoming release of libX11, but the upstream changes
have been pulled into Pkgsrc as the patches libX11/patches/patch-a[ik].
Tested on the same Linux system that prompted Bug#15884.
Bump PKGREVISION.
Adam Jackson (2):
Connection failure for abstract socket is ECONNREFUSED, not ENOENT.
xtrans 1.2.1
Alan Coopersmith (2):
LocalClose() takes a ConnInfoPtr, not an fd
Clear some pointer type mismatch warnings
Bill Nottingham (1):
Ignore mkdir() errors when creating the abstract socket.
James Cloos (1):
Remove extraneous execute bit from .c file
Jeremy Huddleston (1):
Launchd: This functionality has been moved into xorg-server
socket implementation, see:
http://bugs.freedesktop.org/show_bug.cgi?id=15884
This change only affects Linux as other platforms don't use the
abstract sockets code to start with. But anyway, bump PKGREVISION.
Alan Coopersmith (1):
Sun bug #6688467: _X11TransConvertAddress: Unknown family
type on 64-bit SPARC
Alan Hourihane (2):
fix build for MAKEWORD
disable UNIXCONN on MINGW
Colin Harrison (2):
Only call WSAGetLastError() if there has been an
Update to winsock2
James Cloos (1):
Fix length calculation for the path for abstract unix domain sockets
Julien Cristau (2):
BSD44SOCKETS is the wrong check for SOCK_MAXADDRLEN
Bump to 1.2
Loïc Minier (1):
Bug #10489: Don't retry unix socket connect()s on ENOENT.
Adam Jackson (2):
Add support for the abstract socket namespace under Linux.
xtrans 1.1
Ben Byer (5):
suppress warning about socket directory ownership on OS X
changes to support launchd on OS X
removed cvs tags
Fix for incorrect processing of recycled launchd socket on OS X
make launchd error messages less scary
Dodji Seketeli (2):
libxtrans: fix linux compilation breakage
libxtrans: fixed a typo in my last commit
Jeremy Huddleston (1):
Fixed #ifdef checks that were using i386 to use __i386__
