Annotation of src/usr.bin/window/README, Revision 1.3
1.2 deraadt 1: /* $OpenBSD$ */
2:
1.1 deraadt 3: /*-
4: * Copyright (c) 1990, 1993
5: * The Regents of the University of California. All rights reserved.
6: *
7: * This code is derived from software contributed to Berkeley by
8: * Edward Wang at The University of California, Berkeley.
9: *
10: * Redistribution and use in source and binary forms, with or without
11: * modification, are permitted provided that the following conditions
12: * are met:
13: * 1. Redistributions of source code must retain the above copyright
14: * notice, this list of conditions and the following disclaimer.
15: * 2. Redistributions in binary form must reproduce the above copyright
16: * notice, this list of conditions and the following disclaimer in the
17: * documentation and/or other materials provided with the distribution.
18: * 3. All advertising materials mentioning features or use of this software
19: * must display the following acknowledgement:
20: * This product includes software developed by the University of
21: * California, Berkeley and its contributors.
22: * 4. Neither the name of the University nor the names of its contributors
23: * may be used to endorse or promote products derived from this software
24: * without specific prior written permission.
25: *
26: * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
27: * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
28: * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
29: * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
30: * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
31: * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
32: * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
33: * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
34: * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
35: * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
36: * SUCH DAMAGE.
37: *
38: * @(#)README 8.1 (Berkeley) 6/6/93
39: */
40:
41: Compilation notes:
42:
43: Compiler options:
44:
45: BYTE_ORDER (used only in ww.h)
46: It should already be defined in machine/endian.h.
47: The code knows about BIG_ENDIAN, LITTLE_ENDIAN, and PDP_ENDIAN.
48: It only cares about byte order in words, so PDP_ENDIAN
49: is the same as LITTLE_ENDIAN.
50: OLD_TTY
51: If you don't have Posix termios, then define this.
52: VMIN_BUG
53: Even if you have Posix termios, define this if the MIN and TIME
54: feature in noncanonical mode doesn't work correctly.
55:
56: Ok, there's another one, STR_DEBUG. It turns on consistency checks
57: in the string allocator. It's been left on since performace doesn't
58: seem to suffer. There's an abort() somewhere when an inconsistency
59: is found. It hasn't happened in years.
60:
61: The file local.h contains locally tunable constants.
62:
63: The makefile used to be updated with mkmf; it has been changed
64: at various times to use cpp -M and, currently, mkdep. The only library
65: it needs is termcap.
66:
67: Window, as is, only runs on 4.3 (or later) machines.
68:
69: On 4.2 machines, at least these modifications must be done:
70:
71: delete uses of window size ioctls: TIOCGWINSZ, TIOCSWINSZ,
72: struct winsize
73: add to ww.h
74: typedef int fd_set;
75: #define FD_ZERO(s) (*(s) = 0)
76: #define FD_SET(b, s) (*(s) |= 1 << (b))
77: #define FD_ISSET(b, s) (*(s) & 1 << (b))
78: add to ww.h
79: #define sigmask(s) (1 << (s) - 1)
80:
81:
82: A few notes about the internals:
83:
84: The window package. Windows are opened by calling wwopen().
85: Wwwrite() is the primitive for writing to windows. Wwputc(), wwputs(),
86: and wwprintf() are also supported. Some of the outputs to windows are
87: delayed. Wwupdate() updates the terminal to match the internal screen
88: buffer. Wwspawn() spawns a child process on the other end of a window,
89: with its environment tailored to the window. Visible windows are
90: doubly linked in the order of their overlap. Wwadd() inserts a window
91: into the list at a given place. Wwdelete() deletes it. Windows not in
92: the list are not visible, though wwwrite() still works. Window was
93: written before the days of X and Sunview, so some of the terminology
94: is not standard.
95:
96: Most functions return -1 on error. Wwopen() returns the null
97: pointer. An error number is saved in wwerrno. Wwerror() returns an
98: error string based on wwerrno suitable for printing.
99:
100: The terminal drivers perform all output to the physical terminal,
101: including special functions like character and line insertion and
102: deletion. The window package keeps a list of known terminals. At
103: initialization time, the terminal type is matched against the list to
104: find the right terminal driver to use. The last driver, the generic
105: driver, matches all terminals and uses the termcap database. The
106: interface between the window package the terminal driver is the `tt'
107: structure. It contains pointers to functions to perform special
108: functions and terminal output, as well as flags about the
109: characteristics of the terminal. Most of these ideas are borrowed
110: from the Maryland window package, which in turn is based on Goslin's
111: Emacs.
112:
113: The IO system is semi-synchronous. Terminal input is signal
114: driven, and everything else is done synchronously with a single
115: select(). It is roughly event-driven, though not in a clean way.
116:
117: Normally, in both conversation mode and command mode, window
118: sleeps in a select() in wwiomux() waiting for data from the
119: pseudo-terminals. At the same time, terminal input causes SIGIO which
120: is caught by wwrint(). The select() returns when at least one of the
121: pseudo-terminals becomes ready for reading.
122:
123: Wwrint() is the interrupt handler for tty input. It reads input
124: into a linear buffer accessed through four pointers:
125:
126: +-------+--------------+----------------+
127: | empty | data | empty |
128: +-------+--------------+----------------+
129: ^ ^ ^ ^
130: | | | |
131: wwib wwibp wwibq wwibe
132:
133: Wwrint() appends characters at the end and increments wwibq (*wwibq++
134: = c), and characters are taken off the buffer at wwibp using the
135: wwgetc() and wwpeekc() macros. As is the convention in C, wwibq
136: and wwibe point to one position beyond the end. In addition,
137: wwrint() will do a longjmp(wwjmpbuf) if wwsetjmp is true. This is
138: used by wwiomux() to interrupt the select() which would otherwise
139: resume after the interrupt. (Actually, I hear this is not true,
140: but the longjmp feature is used to avoid a race condition as well.
141: Anyway, it means I didn't have to depend on a feature in a
142: daily-changing kernel, but that's another story.) The macro
143: wwinterrupt() returns true if the input buffer is non-empty.
144: Wwupdate(), wwwrite(), and wwiomux() check this condition and will
145: return at the first convenient opportunity when it becomes true.
146: In the case of wwwrite(), the flag ww_nointr in the window structure
147: overrides this. This feature allows the user to interrupt lengthy
148: outputs safely. The structure of the input buffer is designed to
149: avoid race conditions without blocking interrupts.
150:
151: Actually, wwsetjmp and wwinterrupt() are part of a software
152: interrupt scheme used by the two interrupt catchers wwrint() and
153: wwchild(). Asserting the interrupt lets the synchronous parts of
154: the program know that there's an interesting asynchronous condition
155: (i.e., got a keyboard character, or a child process died) that they
156: might want to process before anything else. The synchronous routines
157: can check for this condition with wwinterrupt() or by arranging
158: that a longjmp() be done.
159:
160: Wwiomux() copies pseudo-terminal output into their corresponding
161: windows. Without anything to do, it blocks in a select(), waiting for
162: read ready on pseudo-terminals. Reads are done into per-window buffers
163: in the window structures. When there is at least one buffer non-empty,
164: wwiomux() finds the top most of these windows and writes it using
165: wwwrite(). Then the process is repeated. A non-blocking select() is
166: done after a wwwrite() to pick up any output that may have come in
167: during the write, which may take a long time. Specifically, we use
168: this to stop output or flush buffer when a pseudo-terminal tells us to
169: (we use pty packet mode). The select() blocks only when all of the
170: windows' buffers are empty. A wwupdate() is done prior to this, which
171: is the only time the screen is guaranteed to be completely up to date.
172: Wwiomux() loops until wwinterrupt() becomes true.
173:
174: The top level routine for all this is mloop(). In conversation
175: mode, it simply calls wwiomux(), which only returns when input is
176: available. The input buffer is then written to the pseudo-terminal of
177: the current window. If the escape character is found in the input,
178: command mode is entered. Otherwise, the process is repeated. In
179: command mode, control is transferred to docmd() which returns only when
180: conversation mode is reentered. Docmd() and other command processing
181: routines typically wait for input in a loop:
182:
183: while (wwpeekc() < 0)
184: wwiomux();
185:
186: When the loop terminates, wwgetc() is used to read the input buffer.
187:
188: Output to the physical terminal is handled by the lowest level
189: routines of the window package, in the files ttoutput.c and tt.h. The
190: standard IO package is not used, to get better control over buffering
191: and to use non-blocking reads in wwrint(). The buffer size is set to
192: approximately one second of output time, based on the baudrate.
193:
194: The result of all this complexity is faster response time,
195: especially in output stopping and flushing. Wwwrite() checks
196: wwinterrupt() after every line. It also calls wwupdate() for each line
197: it writes. The output buffer is limited to one second of output time.
198: Thus, there is usually only a delay of one to two lines plus one second
199: after a ^C or ^S. Also, commands that produce lengthy output can be
200: aborted without actually showing all of it on the terminal. (Try the
201: '?' command followed by escape immediately.)