99e8626a01
Tidy check for exception in Network::Connection::send().
2012-04-15 23:29:05 -04:00
7103c5f48b
Using fclose() instead of close() means future fprintf() doesn't kill us on Mac.
2012-04-15 23:25:21 -04:00
f376b80018
Fix protobuf -I lines to build with non-system protobuf
2012-04-15 23:20:10 -04:00
aca69ef684
Server does not need to raise exception on failed sendto(). Fixes #183 .
2012-04-15 23:13:58 -04:00
6b6747cda7
Fix valgrind error with 1-col char then 2-col char in a 1x1 terminal
2012-04-15 17:11:40 -04:00
cd2d482484
Don't use *_t for our own typedefs
...
It's reserved by POSIX.
(closes #158 pull request)
2012-04-15 13:50:18 -04:00
937882caff
Remove Boost dep for Fedora too
2012-04-15 13:49:15 -04:00
280400152d
Replace one find_if with a loop
...
g++/libstdc++ 4.1.2 is unhappy with this use of bind2nd, possibly because the
argument is a reference. This is a problem for EL5.
2012-04-15 13:49:15 -04:00
8bcc001b52
Remove Boost entirely
2012-04-15 13:49:15 -04:00
2675f13676
Remove boost::lambda remaining instances from terminaloverlay.cc
2012-04-15 13:49:15 -04:00
06fa8eaedb
Remove boost::lambda from PredictionEngine::get_or_make_row
2012-04-15 13:49:15 -04:00
5f1bc88ed8
Remove boost::lambda with for_each in terminaloverlay.cc
2012-04-15 13:49:15 -04:00
8f099cafbb
Remove boost::lambda from ConditionalOverlayCell::get_validity
2012-04-15 13:49:15 -04:00
83694977a7
Remove boost::lambda from transportsender.cc
2012-04-15 13:49:15 -04:00
8bc9aa33bf
Remove boost::lambda from completeterminal.cc
2012-04-15 13:49:15 -04:00
4152af9c8e
Remove BOOST_AUTO from terminaloverlay.cc
2012-04-15 13:49:15 -04:00
5320fbd346
Remove BOOST_AUTO from completeterminal.cc
2012-04-15 13:49:15 -04:00
6e56fd2157
Remove BOOST_AUTO from terminaldisplay.cc
2012-04-15 13:49:15 -04:00
87a5f86794
Remove BOOST_AUTO from terminalframebuffer.cc
2012-04-15 13:49:15 -04:00
5904d4f5e6
Remove BOOST_AUTO from transportsender.cc
2012-04-15 13:49:15 -04:00
cc8b1f293a
Forcibly detach earlier in blind effort to fix #114 (needs testing)
2012-04-15 12:38:48 -04:00
8f60f7ab05
Convey locale-related environment variables as backup, and nicer error.
...
Fixes #74 .
2012-04-14 16:56:48 -04:00
597756b846
Document new -v (verbose) server switch.
2012-04-14 13:56:46 -04:00
a4221d17a4
Disable posterization and rely solely on TERM to tell remote end about colors
...
(should help #155 github issue)
2012-04-14 13:46:38 -04:00
0e9be1b471
Close terminal fd's to properly detach on old sshd (e.g. RHEL/CentOS 5.0).
...
May help some sufferers of #114 github issue.
2012-04-14 12:37:28 -04:00
1823f7e860
Provide more helpful error message on lack of -t argument.
2012-04-14 12:29:05 -04:00
5c20c8d4be
Keep using OSC 0 when possible to preserve compability with older clients.
...
Re: #137
2012-04-14 02:11:31 -04:00
d0a818d2e2
Track icon name and window title separately. Implement MOSH_TITLE_NOPREFIX
2012-04-14 01:55:28 -04:00
53f79e4261
Fix vttest "Funny scrolling regions" by ignoring some invalid scroll regions
...
Addresses #155 github issue.
2012-04-14 01:11:57 -04:00
45487f6623
Handle crazy cursor more gracefully ( closes #168 )
2012-04-13 14:03:16 -04:00
7be90ae244
Close server-related file descriptors before executing child process
...
Closes #170 github issue.
2012-04-13 13:48:56 -04:00
0eec0b60f0
Linux: Use our own signalfd wrapper, rather than libstddjb
...
selfpipe already does a fine job of interfacing to signalfd. But Debian and
Ubuntu want us to depend on the skalibs-dev package rather than build libstddjb
ourselves. That would be fine except that skalibs-dev has static libraries
only, and they aren't built with -fPIC. This interferes with building
mosh-{client,server} as position-independent executables, which is a desirable
security measure.
So we have our own wrapper, which invokes either signalfd or selfpipe. And we
build it ourselves with our own flags, because it's part of the Mosh project
proper.
(closes #108 )
2012-04-13 12:33:21 -04:00
93325d3ef0
Include -fPIE when we check for $LD -pie
...
configure's linker flag check also compiles a C++ program. g++ -pie informs
both the compiler and linker, but clang++ -Wl,-pie is only a linker flag, and
the linker will complain that the code hasn't been compiled as
position-independent.
2012-04-13 12:27:27 -04:00
8c16ca8123
Try $LD -Wl,-pie if $LD -pie fails
...
clang on OS X wants the former.
c.f. http://bugs.ruby-lang.org/issues/5697
2012-04-13 12:27:27 -04:00
d2434d1a79
Check compiler support for -fno-default-inline and -pipe
...
clang was spewing warnings about the unrecognized -fno-default-inline. (Oddly,
it warns only with -c, not when compiling directly to an executable.) For
completeness we also check -pipe, even though clang is OK with that one.
It should be fine to omit either flag. gcc -fno-default-inline drops the
implicit 'inline' annotation on functions defined inside a class scope, but
'inline' is only a hint anyway. -fno-default-inline does not change linkage.
-pipe is merely a compile speed optimization.
2012-04-13 12:27:27 -04:00
62b6137ff5
Grep for clang warnings when testing compiler flag support
...
There is no way to make clang's "argument unused" warning fatal.
Ideally the search string would also include 'clang: ' but this output might
depend on clang's argv[0].
2012-04-13 12:27:27 -04:00
b5ab0d0736
Document the binary hardening measures
2012-04-13 12:27:25 -04:00
349e2d20c8
Use the hardening flags
...
Closes #79 .
2012-04-13 12:25:09 -04:00
ffd7332f8d
Detect binary hardening flags supported by the compiler and linker
2012-04-13 12:25:09 -04:00
f7fc34e0c7
Add ax_check_{compile,link}_flag.m4
2012-04-13 12:25:09 -04:00
3d76ec8208
Skip 'new' otherwise BSD getopt will think there are no options.
...
IMO 'new' should be dropped entirely since it's not optional and 'new' is the only choice.
If it were to change in future then it should go after all the options.
(closes #111 )
2012-04-13 12:07:49 -04:00
1f34a08346
AX_CHECK_LIBRARY: Bump serial to match upstream
...
The m4_ifnblank patch was accepted upstream as
http://git.savannah.gnu.org/gitweb/?p=autoconf-archive.git;a=commitdiff;h=7ea1d1b18e5c1c3300936dd89c3e55b5b370ae72
Signed-off-by: Anders Kaseorg <andersk@mit.edu >
(closes #153 )
2012-04-13 12:06:23 -04:00
b86cf72ea4
README: Clarify that mosh-{client,server} don't need to be "installed" per se
2012-04-12 22:02:28 -07:00
88de326344
README: Mention IRC channel
2012-04-12 21:53:02 -07:00
b803c49fc9
README: Add warning about env var security
2012-04-12 21:53:02 -07:00
afa09948fe
README: Remove some information duplicated on mosh.mit.edu
2012-04-12 21:53:02 -07:00
99842c974f
Augment Ubuntu PPA installation instructions.
...
Ubuntu 10.04 does not recognise `sudo add-apt-repository ppa:blah` out of the box.
2012-04-11 11:29:31 +03:00
719e085f3e
Bump version number
2012-04-11 03:37:09 -04:00
a67ae7b0f4
Support 16-color set (per Anders Kaseorg).
...
Closes #106 github issue.
Closes #133 github issue.
Closes #131 github issue.
2012-04-11 03:09:26 -04:00
774e50dc19
Allow lowercase "utf-8" in charset for OS X iTerm.
...
Closes #104 github pull request.
2012-04-11 02:49:35 -04:00
Keith Winstein