Commit Graph

119 Commits

Author SHA1 Message Date
Ryan Prichard
a7bc8fcf9d Rename Shared to shared and Misc to misc. 2012-04-01 02:13:21 -07:00
Ryan Prichard
17eaaadcec Use 4-space indentation in Misc/UnixEcho.cc. 2012-04-01 02:03:12 -07:00
Ryan Prichard
8eb2da5850 Fix a typo. 2012-04-01 02:00:37 -07:00
Ryan Prichard
9f26d1fc1e Remove a TODO comment for a completed task. 2012-04-01 01:42:50 -07:00
Ryan Prichard
f35e5bde74 Use 4-space indentation consistently. 2012-04-01 01:40:30 -07:00
Ryan Prichard
023d6c71f9 Realign indentation in a couple places. 2012-04-01 01:40:15 -07:00
Ryan Prichard
24196c84ea Fix an inconsistency in the README and reword a couple of sentences. 2012-04-01 01:22:04 -07:00
Ryan Prichard
3f09acdf06 Reword the readme's introduction. 2012-04-01 01:01:18 -07:00
Ryan Prichard
80d005e33e Start the agent's child process on the libwinpty client's desktop.
If I run a GUI program inside mintty using winpty, the GUI's window should
appear on the same desktop as mintty, not on the hidden desktop containing
the winpty console.

Even though STARTUPINFO.lpDesktop's type is LPTSTR (instead of LPCTSTR),
the documentation doesn't say anything about CreateProcess modifying the
lpDesktop string, so I didn't bother making a copy of desktop.c_str().
2012-03-31 00:55:49 -07:00
Ryan Prichard
68e32a2bde Use GetEnvironmentVariable instead of getenv in DebugClient.cc.
This lets libwinpty.dll read the WINPTYDBG environment variable that the
unix adapter sets using SetEnvironmentVariable.  If DebugClient.cc calls
getenv, then it seems to read a copy of the Win32 environment that the
MSVCRT initializes at startup.
2012-03-31 00:49:09 -07:00
Ryan Prichard
0d673836f2 In the DebugServer script, use sys.stdout.flush instead of "python -u".
The latter doesn't work when:
 * I'm using mintty.
 * I invoke python.exe directly, passing it the path to the script.

This came up on Windows XP testing where I installed CPython, but
python.exe wasn't in my PATH, so I couldn't run the script directly.
2012-03-30 23:49:46 -07:00
Ryan Prichard
b44bd6b655 Remove a piece of old Qt code from this Makefile. 2012-03-28 03:12:22 -07:00
Ryan Prichard
5542b37de0 Update the .gitignore file. 2012-03-28 03:12:21 -07:00
Ryan Prichard
7eae3d9e33 Send "\r" instead of "\x1b[1G" because carriage return is simpler.
There's no other motivating reason for this change.  AFAIK, both of them
work equally well and do the same thing.
2012-03-28 03:12:13 -07:00
Ryan Prichard
6a29852ace Stop sending trailing whitespace in Terminal::sendLine.
* The intent of having the length variable was to avoid sending trailing
   whitespace for each line, but the previous code never read the variable.
   Instead, every line it sent was the width of the terminal window.

 * There was a bug where running console.exe inside a TERM=cygwin terminal
   resulted in excessive line feeds.  This change mitigates that bug, but
   doesn't fix it.  The problem is that the agent assumes that after it
   writes to the last column of a line, the cursor is still on the same
   line, but this isn't true for the TERM=cygwin terminal.
2012-03-28 03:10:37 -07:00
Ryan Prichard
0427aedadd Remove unused console font code from Win32Console::setSmallFont. 2012-03-28 01:09:52 -07:00
Ryan Prichard
dbc2d51fec Select a small console font on agent startup.
This maximizes the possible console size in terms of rows * cols.
2012-03-28 01:05:26 -07:00
Ryan Prichard
6585b72375 Use STD_{INPUT,OUTPUT}_HANDLE instead of reopening CONIN$ and CONOUT$.
Reopening the files is unnecessary.
2012-03-28 01:05:14 -07:00
Ryan Prichard
67ee37396d Add a readme. 2012-03-25 19:30:47 -07:00
Ryan Prichard
46701caf39 Mark configure as executable and exit if uname -s isn't recognized. 2012-03-25 19:24:39 -07:00
Ryan Prichard
82cfc70d98 Create a configure script.
* Move the Cygwin/MSYS detection logic from config-*.mk into this new
   script.

 * Recognize either the 32-bit MinGW or the 32-bit MinGW-w64 compiler
   driver for Cygwin.
2012-03-25 15:55:03 -07:00
Ryan Prichard
71aa4a59bb Rename pconsole to winpty.
* This avoids a name conflict with an existing unrelated project.

 * I think winpty is a better name anyway.  (i) It's more obviously
   Windows-related.  (ii) I think it more accurately describes what it
   does.  A "pseudo-console" ought to mirror the console API and allow
   replacing the implementation built-in to Windows.  This project is only
   trying to provide functionality similar to using the master interface of
   a Unix pty, but for native console programs.
2012-03-25 02:29:37 -07:00
Ryan Prichard
f1842c63a1 Remove old Qt files. 2012-03-24 18:02:28 -07:00
Ryan Prichard
43fac6819b Remove the obsolete TestNet client/server code.
It won't work anymore because the agent's interface has changed.
2012-03-24 17:55:39 -07:00
Ryan Prichard
e1c4763cc4 Get the UNIX adapter working with MSYS.
* Work around UNICODE deficiencies:
    - MSYS does not have wstring.
    - mbstowcs(NULL, text, 0) returns 0 on MSYS instead of returning the
      length of the converted string.
    - printf("%ls") does not work on MSYS.
2012-03-24 17:53:43 -07:00
Ryan Prichard
35754bab98 In the UNIX adapter, remove the UNIX->Win32 environment conversion.
It doesn't work because it leaves in the Cygwin-style paths in variables
like PATH and TMP.

(Also, it doesn't build with MSYS because MSYS's old G++ toolchain lacks
std::wstring.)
2012-03-24 17:23:18 -07:00
Ryan Prichard
9ed1b5b286 Print usage in the UNIX adapter. 2012-03-24 17:20:08 -07:00
Ryan Prichard
5010f5e1a6 Print an error message when CreateProcess fails.
I suspect that the library API (pconsole_open) should not be calling exit
if it can't start the agent process, but I don't want to deal with that
right now.
2012-03-24 17:19:55 -07:00
Ryan Prichard
22f96abc03 Add comments to the pconsole.h header file. 2012-03-24 17:17:23 -07:00
Ryan Prichard
5c185c1617 In the UNIX adapter, start the child process specified on the command line.
* Also propagate the UNIX environment to Win32.  This code doesn't really
   work because variables like PATH need path-translation, but I'd like to
   save a record of it before killing it off.
2012-03-24 16:48:41 -07:00
Ryan Prichard
32627354df Work on the agent debugging code.
* Suppress debug output unless the PCONSOLEDBG enviroment variable is
   set.

 * Rename Trace to trace to be consistent with the camelCaps naming
   convention.
2012-03-24 16:43:50 -07:00
Ryan Prichard
f75800fae6 Replace assert with ASSERT. 2012-03-24 16:41:15 -07:00
Ryan Prichard
724650ac71 Add more copyright notices. 2012-03-24 16:40:18 -07:00
Ryan Prichard
0662815a3e Replace build/.placeholder with build/.gitkeep.
Apparently the convention is to name the empty directory placerholder file
.gitkeep.
2012-03-23 03:48:40 -07:00
Ryan Prichard
44554d5e9d Add an MIT copyright notice to all the source files. 2012-03-23 03:11:34 -07:00
Ryan Prichard
d3667e38a8 Review the named pipe setup code.
The goal is to setup pipes between the libpconsole process and the new
agent process in a way that's secure and doesn't create an unnecessary
burden on the library's user.  This is apparently hard.

I think this code is good enough.
2012-03-23 02:19:09 -07:00
Ryan Prichard
d568fbdeb4 Add DsrSender.h. 2012-03-21 03:31:34 -07:00
Ryan Prichard
d418679741 Include stdint.h to provide int32_t.
* This fixed a libpconsole build error that happened with mingw-w64.
2012-03-21 03:30:18 -07:00
Ryan Prichard
f66e877cba Improve error handling in libpconsole.
If we couldn't create the new pipes, then perhaps some other process has
already created them.  I think it's a security hole to start the agent if
the pipe creation failed.
2012-03-21 03:28:31 -07:00
Ryan Prichard
7e6f8195ba Turn on -Wall and fix build warnings. No bug fixes. 2012-03-21 00:14:04 -07:00
Ryan Prichard
0afc69c733 Move the three output binaries into a build directory. 2012-03-21 00:13:46 -07:00
Ryan Prichard
4cfb5e1c73 Expand a comment. 2012-03-20 23:46:12 -07:00
Ryan Prichard
f397150b6c Add some TODO items. 2012-03-20 23:45:59 -07:00
Ryan Prichard
0260b423a3 Continue working on console input.
* Remove most of the encoding table and instead generate the KeyDescriptor
   entries dynamically at startup.

 * I think I'm handling input tolerably well for TERM=xterm terminals, but
   I removed the support I had for rxvt for now.  I don't think it makes
   sense to fill out the table, but the new code I have for TERM=xterm is
   also bloated.
2012-03-20 04:15:25 -07:00
Ryan Prichard
1fdda9e1d4 Work on console input handling.
* If bytes are still queued after processing as many keypresses as
   possible, send a DSR to the console.  The console should respond with
   a DSR reply, which will flush out the input buffer without having to
   wait for a timeout.

 * Add ah hoc code for ALT-<character> rather than listing every character
   in the table.

 * When keypresses have Alt/Ctrl/Shift modifiers, generate extra
   INPUT_RECORDS to press and release each modifier.  EDIT.COM seemed to
   require these for some Ctrl-<key> keypresses I tried.
2012-03-20 02:04:46 -07:00
Ryan Prichard
d32e36032d Fix an bug in ConsoleInput::appendChar: set dwControlKeyState. 2012-03-19 22:14:34 -07:00
Ryan Prichard
98970ead5c Use ESC and CSI macros in ConsoleInput::keyDescriptorTable. 2012-03-19 22:14:06 -07:00
Ryan Prichard
ac76cbcc20 Initial checkin of improved console input handling.
* Generate Ctrl-C events by calling GenerateConsoleCtrlEvent.

   I noticed that calls to this routine don't behave exactly the same as
   a real Ctrl-C keypress.  In Python, pressing Ctrl-C immediately
   displays a new "KeyboardInterrupt" line.  Calling
   GenerateConsoleCtrlEvent has no immediate effect, but after pressing
   Enter, Python displays a stack trace where a KeyboardInterrupt was
   raised.  After some testing, I suspect the issue is that a real Ctrl-C
   keypress interrupts a blocking console read, but
   GenerateConsoleCtrlEvent does not.

   I also tried synthesizing Ctrl-C using (a) PostMessage with
   WM_{CHAR,KEYDOWN}, and (b) SendInput.  I couldn't get either to work.

 * Recognize ESC sequences.  The set of recognized sequences is ad hoc.

 * Recognize UTF-8-encoded characters and convert them to UTF-16.

 * The code currently uses a timeout to differentiate between pressing ESC
   and pressing a key that generates an ESC sequence.  I have a theory that
   I can use the "Device Status Report" ESC sequences to avoid this
   timeout.
2012-03-19 21:30:16 -07:00
Ryan Prichard
0554f9fbde In Agent::handleStartProcessPacket, call GetLastError immediately after
CreateProcess.
2012-03-19 21:17:43 -07:00
Ryan Prichard
7dbe7537c3 Display Ctrl-C events in the ShowConsoleInput test program.
Now that Ctrl-C doesn't exit ShowConsoleInput, make the program exit when
the user hits ESC three times (3 up and 3 down for a total of 6 events).
2012-03-19 21:13:10 -07:00