client role in reverse connections. Improvements to logging in
ultravnc_repeater, ULTRAVNC_REPEATER_NO_RFB option. Increase
SSL timeout and print message if 'repeater' mode is detected for
reverse SSL connection. Fix RECORD scroll XCopyArea detection
with recent gtk/gdk library; set X11VNC_SCROLL_MUST_EQUAL
to disable. Limit logging of RECORD error messages.
serverCert param, debugging printout, user dialogs, catch
socket exceptions, autodetect x11vnc for GET=1.
x11vnc: misc/scripts: desktop.cgi, inet6to4, panner.pl.
X11VNC_HTTPS_DOWNLOAD_WAIT_TIME, -unixpw %xxx documented, and
can run user cmd in UNIXPW_CMD. FD_XDMCP_IF for create script,
autodetect dm on udp6 only. Queries: pointer_x, pointer_y,
pointer_same, pointer_root. Switch on -xkd if keysyms per key >
4 in all cases. daemon mode improvements for connect_switch,
inet6to4, ultravnc_repeater.pl. Dynamic change of -clip do
not create new fb if WxH is unchanged.
DIRECT: with an extra XFlush and other safety measures.
fflush(stderr) much in su_verify. Make the -unixpw env. vars
UNIXPW_DISABLE_SSL and UNIXPW_DISABLE_LOCALHOST work correctly.
Make -loopbg actually imply -bg.
entire desktop. map port + 5500 in reverse connect. Add id_cmd remote
control functions for id (and other) windows. Allow zero port in SSL
reverse connections. Adjust delays between multiple reverse connections;
X11VNC_REVERSE_SLEEP_MAX env var. Add some missing mutex locks; add
INPUT_LOCK and threads_drop_input. More safety in -threads mode for
new framebuffer change. Fix some stderr leaking in -inetd mode.
-reflect. Workaround that checks valid window of selection
requestor. Wait on some ssl helper pids earlier. Workaround
XAUTHLOCALHOSTNAME for some new usage modes. Set fake fb to
requested bpp with correct masks. -padgeom once:... mode.
Set LIBXCB_ALLOW_SLOPPY_LOCK by default. rfbRandomBytes earlier.
classes/ssl: Update jars. Add "TOP_" dropdown customization to
ultravnc java viewer applet FTP panel.
The flag handling (both compiler options and include paths) are a mess at
the moment. There is no point in forcing "-O2 -g" when these are already
the defaults, and if someone changes the defaults, chances are good they
don't want you clobbering their choices.
The -Wall flag should be handled in configure and thrown into CFLAGS once
rather than every Makefile.am. Plus, this way we can control which
compilers the flag actually gets used with.
Finally, the INCLUDES variable is for -I paths, not AM_CFLAGS. Nor should
it contain -I. as this is already in the default includes setup.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>