Annotation of src/usr.bin/window/README, Revision 1.1.1.1
1.1 deraadt 1: /*-
2: * Copyright (c) 1990, 1993
3: * The Regents of the University of California. All rights reserved.
4: *
5: * This code is derived from software contributed to Berkeley by
6: * Edward Wang at The University of California, Berkeley.
7: *
8: * Redistribution and use in source and binary forms, with or without
9: * modification, are permitted provided that the following conditions
10: * are met:
11: * 1. Redistributions of source code must retain the above copyright
12: * notice, this list of conditions and the following disclaimer.
13: * 2. Redistributions in binary form must reproduce the above copyright
14: * notice, this list of conditions and the following disclaimer in the
15: * documentation and/or other materials provided with the distribution.
16: * 3. All advertising materials mentioning features or use of this software
17: * must display the following acknowledgement:
18: * This product includes software developed by the University of
19: * California, Berkeley and its contributors.
20: * 4. Neither the name of the University nor the names of its contributors
21: * may be used to endorse or promote products derived from this software
22: * without specific prior written permission.
23: *
24: * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
25: * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
26: * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
27: * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
28: * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
29: * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
30: * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
31: * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
32: * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
33: * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34: * SUCH DAMAGE.
35: *
36: * @(#)README 8.1 (Berkeley) 6/6/93
37: */
38:
39: Compilation notes:
40:
41: Compiler options:
42:
43: BYTE_ORDER (used only in ww.h)
44: It should already be defined in machine/endian.h.
45: The code knows about BIG_ENDIAN, LITTLE_ENDIAN, and PDP_ENDIAN.
46: It only cares about byte order in words, so PDP_ENDIAN
47: is the same as LITTLE_ENDIAN.
48: OLD_TTY
49: If you don't have Posix termios, then define this.
50: VMIN_BUG
51: Even if you have Posix termios, define this if the MIN and TIME
52: feature in noncanonical mode doesn't work correctly.
53:
54: Ok, there's another one, STR_DEBUG. It turns on consistency checks
55: in the string allocator. It's been left on since performace doesn't
56: seem to suffer. There's an abort() somewhere when an inconsistency
57: is found. It hasn't happened in years.
58:
59: The file local.h contains locally tunable constants.
60:
61: The makefile used to be updated with mkmf; it has been changed
62: at various times to use cpp -M and, currently, mkdep. The only library
63: it needs is termcap.
64:
65: Window, as is, only runs on 4.3 (or later) machines.
66:
67: On 4.2 machines, at least these modifications must be done:
68:
69: delete uses of window size ioctls: TIOCGWINSZ, TIOCSWINSZ,
70: struct winsize
71: add to ww.h
72: typedef int fd_set;
73: #define FD_ZERO(s) (*(s) = 0)
74: #define FD_SET(b, s) (*(s) |= 1 << (b))
75: #define FD_ISSET(b, s) (*(s) & 1 << (b))
76: add to ww.h
77: #define sigmask(s) (1 << (s) - 1)
78:
79:
80: A few notes about the internals:
81:
82: The window package. Windows are opened by calling wwopen().
83: Wwwrite() is the primitive for writing to windows. Wwputc(), wwputs(),
84: and wwprintf() are also supported. Some of the outputs to windows are
85: delayed. Wwupdate() updates the terminal to match the internal screen
86: buffer. Wwspawn() spawns a child process on the other end of a window,
87: with its environment tailored to the window. Visible windows are
88: doubly linked in the order of their overlap. Wwadd() inserts a window
89: into the list at a given place. Wwdelete() deletes it. Windows not in
90: the list are not visible, though wwwrite() still works. Window was
91: written before the days of X and Sunview, so some of the terminology
92: is not standard.
93:
94: Most functions return -1 on error. Wwopen() returns the null
95: pointer. An error number is saved in wwerrno. Wwerror() returns an
96: error string based on wwerrno suitable for printing.
97:
98: The terminal drivers perform all output to the physical terminal,
99: including special functions like character and line insertion and
100: deletion. The window package keeps a list of known terminals. At
101: initialization time, the terminal type is matched against the list to
102: find the right terminal driver to use. The last driver, the generic
103: driver, matches all terminals and uses the termcap database. The
104: interface between the window package the terminal driver is the `tt'
105: structure. It contains pointers to functions to perform special
106: functions and terminal output, as well as flags about the
107: characteristics of the terminal. Most of these ideas are borrowed
108: from the Maryland window package, which in turn is based on Goslin's
109: Emacs.
110:
111: The IO system is semi-synchronous. Terminal input is signal
112: driven, and everything else is done synchronously with a single
113: select(). It is roughly event-driven, though not in a clean way.
114:
115: Normally, in both conversation mode and command mode, window
116: sleeps in a select() in wwiomux() waiting for data from the
117: pseudo-terminals. At the same time, terminal input causes SIGIO which
118: is caught by wwrint(). The select() returns when at least one of the
119: pseudo-terminals becomes ready for reading.
120:
121: Wwrint() is the interrupt handler for tty input. It reads input
122: into a linear buffer accessed through four pointers:
123:
124: +-------+--------------+----------------+
125: | empty | data | empty |
126: +-------+--------------+----------------+
127: ^ ^ ^ ^
128: | | | |
129: wwib wwibp wwibq wwibe
130:
131: Wwrint() appends characters at the end and increments wwibq (*wwibq++
132: = c), and characters are taken off the buffer at wwibp using the
133: wwgetc() and wwpeekc() macros. As is the convention in C, wwibq
134: and wwibe point to one position beyond the end. In addition,
135: wwrint() will do a longjmp(wwjmpbuf) if wwsetjmp is true. This is
136: used by wwiomux() to interrupt the select() which would otherwise
137: resume after the interrupt. (Actually, I hear this is not true,
138: but the longjmp feature is used to avoid a race condition as well.
139: Anyway, it means I didn't have to depend on a feature in a
140: daily-changing kernel, but that's another story.) The macro
141: wwinterrupt() returns true if the input buffer is non-empty.
142: Wwupdate(), wwwrite(), and wwiomux() check this condition and will
143: return at the first convenient opportunity when it becomes true.
144: In the case of wwwrite(), the flag ww_nointr in the window structure
145: overrides this. This feature allows the user to interrupt lengthy
146: outputs safely. The structure of the input buffer is designed to
147: avoid race conditions without blocking interrupts.
148:
149: Actually, wwsetjmp and wwinterrupt() are part of a software
150: interrupt scheme used by the two interrupt catchers wwrint() and
151: wwchild(). Asserting the interrupt lets the synchronous parts of
152: the program know that there's an interesting asynchronous condition
153: (i.e., got a keyboard character, or a child process died) that they
154: might want to process before anything else. The synchronous routines
155: can check for this condition with wwinterrupt() or by arranging
156: that a longjmp() be done.
157:
158: Wwiomux() copies pseudo-terminal output into their corresponding
159: windows. Without anything to do, it blocks in a select(), waiting for
160: read ready on pseudo-terminals. Reads are done into per-window buffers
161: in the window structures. When there is at least one buffer non-empty,
162: wwiomux() finds the top most of these windows and writes it using
163: wwwrite(). Then the process is repeated. A non-blocking select() is
164: done after a wwwrite() to pick up any output that may have come in
165: during the write, which may take a long time. Specifically, we use
166: this to stop output or flush buffer when a pseudo-terminal tells us to
167: (we use pty packet mode). The select() blocks only when all of the
168: windows' buffers are empty. A wwupdate() is done prior to this, which
169: is the only time the screen is guaranteed to be completely up to date.
170: Wwiomux() loops until wwinterrupt() becomes true.
171:
172: The top level routine for all this is mloop(). In conversation
173: mode, it simply calls wwiomux(), which only returns when input is
174: available. The input buffer is then written to the pseudo-terminal of
175: the current window. If the escape character is found in the input,
176: command mode is entered. Otherwise, the process is repeated. In
177: command mode, control is transferred to docmd() which returns only when
178: conversation mode is reentered. Docmd() and other command processing
179: routines typically wait for input in a loop:
180:
181: while (wwpeekc() < 0)
182: wwiomux();
183:
184: When the loop terminates, wwgetc() is used to read the input buffer.
185:
186: Output to the physical terminal is handled by the lowest level
187: routines of the window package, in the files ttoutput.c and tt.h. The
188: standard IO package is not used, to get better control over buffering
189: and to use non-blocking reads in wwrint(). The buffer size is set to
190: approximately one second of output time, based on the baudrate.
191:
192: The result of all this complexity is faster response time,
193: especially in output stopping and flushing. Wwwrite() checks
194: wwinterrupt() after every line. It also calls wwupdate() for each line
195: it writes. The output buffer is limited to one second of output time.
196: Thus, there is usually only a delay of one to two lines plus one second
197: after a ^C or ^S. Also, commands that produce lengthy output can be
198: aborted without actually showing all of it on the terminal. (Try the
199: '?' command followed by escape immediately.)