src/usr.bin/window/README - annotate

Return to README CVS log

Up to [local] / src / usr.bin / window

Annotation of src/usr.bin/window/README, Revision 1.4

1.4 ! millert 1: /* $OpenBSD: README,v 1.3 1997/02/25 00:03:51 downsj Exp $ */
1.2 deraadt 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.
1.4 ! millert 18: * 3. Neither the name of the University nor the names of its contributors
1.1 deraadt 19: * may be used to endorse or promote products derived from this software
20: * without specific prior written permission.
21: *
22: * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23: * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24: * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25: * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26: * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27: * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28: * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29: * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30: * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31: * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32: * SUCH DAMAGE.
33: *
34: * @(#)README 8.1 (Berkeley) 6/6/93
35: */
36:
37: Compilation notes:
38:
39: Compiler options:
40:
41: BYTE_ORDER (used only in ww.h)
42: It should already be defined in machine/endian.h.
43: The code knows about BIG_ENDIAN, LITTLE_ENDIAN, and PDP_ENDIAN.
44: It only cares about byte order in words, so PDP_ENDIAN
45: is the same as LITTLE_ENDIAN.
46: OLD_TTY
47: If you don't have Posix termios, then define this.
48: VMIN_BUG
49: Even if you have Posix termios, define this if the MIN and TIME
50: feature in noncanonical mode doesn't work correctly.
51:
52: Ok, there's another one, STR_DEBUG. It turns on consistency checks
53: in the string allocator. It's been left on since performace doesn't
54: seem to suffer. There's an abort() somewhere when an inconsistency
55: is found. It hasn't happened in years.
56:
57: The file local.h contains locally tunable constants.
58:
59: The makefile used to be updated with mkmf; it has been changed
60: at various times to use cpp -M and, currently, mkdep. The only library
61: it needs is termcap.
62:
63: Window, as is, only runs on 4.3 (or later) machines.
64:
65: On 4.2 machines, at least these modifications must be done:
66:
67: delete uses of window size ioctls: TIOCGWINSZ, TIOCSWINSZ,
68: struct winsize
69: add to ww.h
70: typedef int fd_set;
71: #define FD_ZERO(s) (*(s) = 0)
72: #define FD_SET(b, s) (*(s) |= 1 << (b))
73: #define FD_ISSET(b, s) (*(s) & 1 << (b))
74: add to ww.h
75: #define sigmask(s) (1 << (s) - 1)
76:
77:
78: A few notes about the internals:
79:
80: The window package. Windows are opened by calling wwopen().
81: Wwwrite() is the primitive for writing to windows. Wwputc(), wwputs(),
82: and wwprintf() are also supported. Some of the outputs to windows are
83: delayed. Wwupdate() updates the terminal to match the internal screen
84: buffer. Wwspawn() spawns a child process on the other end of a window,
85: with its environment tailored to the window. Visible windows are
86: doubly linked in the order of their overlap. Wwadd() inserts a window
87: into the list at a given place. Wwdelete() deletes it. Windows not in
88: the list are not visible, though wwwrite() still works. Window was
89: written before the days of X and Sunview, so some of the terminology
90: is not standard.
91:
92: Most functions return -1 on error. Wwopen() returns the null
93: pointer. An error number is saved in wwerrno. Wwerror() returns an
94: error string based on wwerrno suitable for printing.
95:
96: The terminal drivers perform all output to the physical terminal,
97: including special functions like character and line insertion and
98: deletion. The window package keeps a list of known terminals. At
99: initialization time, the terminal type is matched against the list to
100: find the right terminal driver to use. The last driver, the generic
101: driver, matches all terminals and uses the termcap database. The
102: interface between the window package the terminal driver is the `tt'
103: structure. It contains pointers to functions to perform special
104: functions and terminal output, as well as flags about the
105: characteristics of the terminal. Most of these ideas are borrowed
106: from the Maryland window package, which in turn is based on Goslin's
107: Emacs.
108:
109: The IO system is semi-synchronous. Terminal input is signal
110: driven, and everything else is done synchronously with a single
111: select(). It is roughly event-driven, though not in a clean way.
112:
113: Normally, in both conversation mode and command mode, window
114: sleeps in a select() in wwiomux() waiting for data from the
115: pseudo-terminals. At the same time, terminal input causes SIGIO which
116: is caught by wwrint(). The select() returns when at least one of the
117: pseudo-terminals becomes ready for reading.
118:
119: Wwrint() is the interrupt handler for tty input. It reads input
120: into a linear buffer accessed through four pointers:
121:
122: +-------+--------------+----------------+
123: | empty | data | empty |
124: +-------+--------------+----------------+
125: ^ ^ ^ ^
126: | | | |
127: wwib wwibp wwibq wwibe
128:
129: Wwrint() appends characters at the end and increments wwibq (*wwibq++
130: = c), and characters are taken off the buffer at wwibp using the
131: wwgetc() and wwpeekc() macros. As is the convention in C, wwibq
132: and wwibe point to one position beyond the end. In addition,
133: wwrint() will do a longjmp(wwjmpbuf) if wwsetjmp is true. This is
134: used by wwiomux() to interrupt the select() which would otherwise
135: resume after the interrupt. (Actually, I hear this is not true,
136: but the longjmp feature is used to avoid a race condition as well.
137: Anyway, it means I didn't have to depend on a feature in a
138: daily-changing kernel, but that's another story.) The macro
139: wwinterrupt() returns true if the input buffer is non-empty.
140: Wwupdate(), wwwrite(), and wwiomux() check this condition and will
141: return at the first convenient opportunity when it becomes true.
142: In the case of wwwrite(), the flag ww_nointr in the window structure
143: overrides this. This feature allows the user to interrupt lengthy
144: outputs safely. The structure of the input buffer is designed to
145: avoid race conditions without blocking interrupts.
146:
147: Actually, wwsetjmp and wwinterrupt() are part of a software
148: interrupt scheme used by the two interrupt catchers wwrint() and
149: wwchild(). Asserting the interrupt lets the synchronous parts of
150: the program know that there's an interesting asynchronous condition
151: (i.e., got a keyboard character, or a child process died) that they
152: might want to process before anything else. The synchronous routines
153: can check for this condition with wwinterrupt() or by arranging
154: that a longjmp() be done.
155:
156: Wwiomux() copies pseudo-terminal output into their corresponding
157: windows. Without anything to do, it blocks in a select(), waiting for
158: read ready on pseudo-terminals. Reads are done into per-window buffers
159: in the window structures. When there is at least one buffer non-empty,
160: wwiomux() finds the top most of these windows and writes it using
161: wwwrite(). Then the process is repeated. A non-blocking select() is
162: done after a wwwrite() to pick up any output that may have come in
163: during the write, which may take a long time. Specifically, we use
164: this to stop output or flush buffer when a pseudo-terminal tells us to
165: (we use pty packet mode). The select() blocks only when all of the
166: windows' buffers are empty. A wwupdate() is done prior to this, which
167: is the only time the screen is guaranteed to be completely up to date.
168: Wwiomux() loops until wwinterrupt() becomes true.
169:
170: The top level routine for all this is mloop(). In conversation
171: mode, it simply calls wwiomux(), which only returns when input is
172: available. The input buffer is then written to the pseudo-terminal of
173: the current window. If the escape character is found in the input,
174: command mode is entered. Otherwise, the process is repeated. In
175: command mode, control is transferred to docmd() which returns only when
176: conversation mode is reentered. Docmd() and other command processing
177: routines typically wait for input in a loop:
178:
179: while (wwpeekc() < 0)
180: wwiomux();
181:
182: When the loop terminates, wwgetc() is used to read the input buffer.
183:
184: Output to the physical terminal is handled by the lowest level
185: routines of the window package, in the files ttoutput.c and tt.h. The
186: standard IO package is not used, to get better control over buffering
187: and to use non-blocking reads in wwrint(). The buffer size is set to
188: approximately one second of output time, based on the baudrate.
189:
190: The result of all this complexity is faster response time,
191: especially in output stopping and flushing. Wwwrite() checks
192: wwinterrupt() after every line. It also calls wwupdate() for each line
193: it writes. The output buffer is limited to one second of output time.
194: Thus, there is usually only a delay of one to two lines plus one second
195: after a ^C or ^S. Also, commands that produce lengthy output can be
196: aborted without actually showing all of it on the terminal. (Try the
197: '?' command followed by escape immediately.)