scripts.ps

物业管理和办公自动化系统

0 F
1.67 (This script begins by disabling echo so that the password can be typed unseen.  Unlike) 72 280.5 P
5 11 Q
1.7 (exec) 516.8 280.5 P
0 12 Q
-0.18 (which manipulates its standard I/O so that it is accessible to) 72 266.5 P
2 F
-0.18 (expect) 359.53 266.5 P
0 F
-0.18 (, the) 392.17 266.5 P
5 11 Q
-0.18 (system) 415.46 266.5 P
0 12 Q
-0.18 ( command does no) 450.26 266.5 P
(manipulation, thereby allowing) 72 252.5 T
2 F
(stty) 225.24 252.5 T
0 F
( to af) 243.89 252.5 T
(fect the terminal.) 268.32 252.5 T
5 11 Q
0.68 (fork) 72 228.5 P
0 12 Q
0.67 ( literally causes) 90.31 228.5 P
2 F
0.67 (expect) 169.93 228.5 P
0 F
0.67 ( to fork.  Like the UNIX system call of the same name, it returns the) 202.56 228.5 P
0.19 (child process ID to the parent.  Since the parent has nothing else to do, it immediately exits.  The) 72 214.5 P
0.37 (shell will detect this as normal program termination.  Meanwhile,) 72 200.5 P
5 11 Q
0.37 (disconnect) 392.8 200.5 P
0 12 Q
0.37 ( breaks the associa-) 445.31 200.5 P
0.16 (tion between the child process and the terminal so that the rest of the script can continue immune) 72 186.5 P
(to the user logging out.) 72 172.5 T
1.2 (This paradigm provides a secure way of starting long-running background processes which re-) 72 148.5 P
0.31 (quire passwords.  This works well with security schemes such as MIT\325) 72 134.5 P
0.31 (s Kerberos system.  In or-) 414.54 134.5 P
1.3 (der to run a process authenticated by Kerberos, all that is necessary is to spawn) 72 120.5 P
2 F
1.3 (kinit) 473.81 120.5 P
0 F
1.3 ( to get a) 497.81 120.5 P
(ticket, and similarly) 72 106.5 T
2 F
(kdestr) 170.61 106.5 T
(oy) 203.04 106.5 T
0 F
( when the ticket is no longer needed.) 215.03 106.5 T
-0.19 (Before) 72 82.5 P
2 F
-0.19 (expect) 107.44 82.5 P
0 F
-0.19 (, there was no way to achieve such results.  The choice was either in\337exibility or in-) 140.07 82.5 P
(security) 72 68.5 T
(.) 109.19 68.5 T
2 F
(expect) 118.19 68.5 T
0 F
( has made this choice unnecessary) 150.82 68.5 T
(, and given us the best of both worlds.) 314.24 68.5 T
FMENDPAGE
%%EndPage: "17" 16
%%Page: "16" 16
612 792 0 FMBEGINPAGE
72 36 504 47.95 R
7 X
0 K
V
0 12 Q
0 X
(16) 72 39.95 T
0 10 Q
(Don Libes) 108 39.95 T
72 740.81 504 756 R
7 X
V
72 63 540 720 R
V
4 12 Q
0 X
(break) 207 712 T
(}) 162 698.5 T
(}) 115.2 685 T
0 F
-0.18 (This script uses) 72 661.5 P
2 F
-0.18 (ed) 149.09 661.5 P
0 F
-0.18 ( although any editor could be used.  First) 161.09 661.5 P
2 F
-0.18 (ed) 358.71 661.5 P
0 F
-0.18 ( is directed to search for the printer) 370.7 661.5 P
-0.18 (.) 537 661.5 P
-0.07 (Once the printer is found, returns are sent to get the successive lines until the value is either locat-) 72 647.5 P
(ed or no more lines remain.) 72 633.5 T
0.19 (Using a specialized tool such as) 72 609.5 P
2 F
0.19 (awk) 229.03 609.5 P
0 F
0.19 ( might seem like a better alternative, except if you aren\325) 250.35 609.5 P
0.19 (t fa-) 520.16 609.5 P
0.94 (miliar with) 72 595.5 P
2 F
0.94 (awk) 129.84 595.5 P
0 F
0.94 (\325) 151.17 595.5 P
0.94 (s style of processing.  While the same claim could be made about) 154.5 595.5 P
2 F
0.94 (expect) 483.1 595.5 P
0 F
0.94 (, this) 515.74 595.5 P
1.56 (script illustrates the idea that \050ignoring syntax dif) 72 581.5 P
1.56 (ferences\051 you can automate a procedure you) 318.86 581.5 P
(know how to do interactively by simply translating it into) 72 567.5 T
5 11 Q
(send) 351.18 567.5 T
0 12 Q
(/) 375 567.5 T
5 11 Q
(expect) 378.33 567.5 T
0 12 Q
( sequences.) 410.7 567.5 T
3 16 Q
(12.  Is expect too fast?) 108 526.83 T
0 12 Q
0.18 (The previous example demonstrated how) 72 501.5 P
2 F
0.18 (expect) 273.74 501.5 P
0 F
0.18 ( can use an editor to read a \336le.) 306.38 501.5 P
2 F
0.18 (expect) 463.7 501.5 P
0 F
0.18 ( has sim-) 496.33 501.5 P
-0.24 (pler ways of reading \336les.  For instance the command) 72 487.5 P
4 F
-0.57 (send [exec cat /etc/motd]) 329.75 487.5 P
0 F
-0.24 ( writes) 507.93 487.5 P
0.37 (the contents of) 72 473.5 P
2 F
0.37 (/etc/motd) 146.73 473.5 P
0 F
0.37 ( to the current process.  Calling a UNIX program to read a \336le may not) 194.69 473.5 P
0.61 (seem like a fast method but it is a lot faster than having a user type it in.  In a window environ-) 72 459.5 P
0.07 (ment, cutting and pasting is an alternative, but this takes a lar) 72 445.5 P
0.07 (ge amount of time also.  Realistical-) 366.03 445.5 P
(ly) 72 431.5 T
(, blazing speed is hardly needed in a program that simulates users.) 80.55 431.5 T
-0.24 (The speed of) 72 407.5 P
2 F
-0.24 (expect) 136.22 407.5 P
0 F
-0.24 ( operations is described by Libes [2].  One side-ef) 168.86 407.5 P
-0.24 (fect not discussed is that) 405.63 407.5 P
2 F
-0.24 (ex-) 524.68 407.5 P
-0.22 (pect) 72 393.5 P
0 F
-0.22 ( can overrun input buf) 93.31 393.5 P
-0.22 (fers designed for human typists.) 198.8 393.5 P
5 11 Q
-0.22 (send) 357.03 393.5 P
0 12 Q
-0.22 ( supports a) 380.85 393.5 P
1 F
-0.22 (slow) 435.15 393.5 P
0 F
-0.22 ( option \050) 457.15 393.5 P
4 F
-0.53 (send -) 497.35 393.5 P
-0.18 (s) 72 379.5 P
0 F
-0.08 (\051 speci\336cally to avoid this problem.  It is controlled by parameters which describe the number of) 79.2 379.5 P
(bytes to send atomically and a length of time to wait between each packet.) 72 365.5 T
5 11 Q
0.5 (send) 72 341.5 P
0 12 Q
0.49 ( also supports a simulation of actual human typing speed \050) 95.82 341.5 P
4 F
1.18 (send -h) 379.89 341.5 P
0 F
0.49 (\051 according to a modi-) 431.45 341.5 P
0.8 (\336ed W) 72 327.5 P
0.8 (eibull distribution [8], a common statistical tool to simulate interarrival times.  The algo-) 104.15 327.5 P
3.53 (rithm is driven by a random number generator and several user) 72 313.5 P
3.53 (-chosen parameters.  The) 408.85 313.5 P
1.31 (parameters describe two average character interarrival times \050default and word endings\051, mini-) 72 299.5 P
0.22 (mum and maximum interarrival times, and a variability \322shape\323.  Errors are not simulated as this) 72 285.5 P
0.67 (can be done by the user directly) 72 271.5 P
0.67 (.  Simplistic errors may be generated by embedding typing mis-) 227.8 271.5 P
0.58 (takes and corrections \050if desired\051 in a) 72 257.5 P
5 11 Q
0.59 (send) 256.93 257.5 P
0 12 Q
0.58 ( ar) 280.74 257.5 P
0.58 (gument.  A more sophisticated approach could use) 293.43 257.5 P
(an expert system as a coroutine.) 72 243.5 T
3 16 Q
(13.  Security) 108 202.83 T
0 12 Q
0.21 (The) 72 177.5 P
2 F
0.21 (passwd) 93.86 177.5 P
0 F
0.21 ( script shown earlier was designed solely to be used  for conformance testing.  Many) 131.18 177.5 P
1.82 (system administrators want such a script to embed in a comprehensive) 72 163.5 P
1 F
1.82 (adduser) 432.77 163.5 P
0 F
1.82 ( script, which) 471.41 163.5 P
-0.29 (would set up everything that a generic new user needs including an account and password.  Unfor-) 72 149.5 P
0.34 (tunately) 72 135.5 P
0.34 (, calling the) 109.86 135.5 P
2 F
0.34 (passwd) 170.18 135.5 P
0 F
0.34 ( script from another script reopens the very problem that the) 207.5 135.5 P
2 F
0.34 (passwd) 502.68 135.5 P
0 F
-0.01 (program was designed to solve.  Passwords should not be used as ar) 72 121.5 P
-0.01 (guments to programs because) 397.8 121.5 P
(they can be seen by) 72 107.5 T
2 F
(ps) 168.92 107.5 T
0 F
( and similar programs.) 180.25 107.5 T
FMENDPAGE
%%EndPage: "16" 15
%%Page: "15" 15
612 792 0 FMBEGINPAGE
72 740.81 540 756 R
7 X
0 K
V
72 36 540 47.95 R
V
0 10 Q
0 X
(expect: Scripts for Controlling Interactive Processes) 295.54 39.95 T
0 12 Q
(15) 528.01 39.95 T
72 63 540 720 R
7 X
V
0 X
0.05 (This script has two) 72 712 P
5 11 Q
0.05 (interact) 166.14 712 P
0 12 Q
0.05 (s.  The user switches between them by typing) 202.16 712 P
4 F
0.12 (~d) 423.77 712 P
0 F
0.05 ( \050for Dvorak\051 and) 438.17 712 P
4 F
0.12 (~q) 525.61 712 P
0 F
0.63 (\050for qwerty\051.  The Dvorak translation occurs in the procedure) 72 698 P
5 11 Q
0.64 (dvorak) 375.05 698 P
0 12 Q
0.63 ( de\336ned with) 408.02 698 P
5 11 Q
0.64 (proc) 475.54 698 P
0 12 Q
0.63 (.  W) 496.91 698 P
0.63 (ithin) 518 698 P
5 11 Q
0.07 (dvorak) 72 684 P
0 12 Q
0.07 (, an) 104.97 684 P
5 11 Q
0.07 (interact) 125.43 684 P
0 12 Q
0.07 ( gives each character an action that corresponds to sending its Dvorak counter-) 161.45 684 P
0.42 (part instead.  Nothing has to be sent to the user) 72 670 P
0.42 (, since the character will be echoed \050if necessary\051) 300.55 670 P
(by the current process.) 72 656 T
0.43 (The) 72 632 P
4 F
1.04 (return continue) 94.08 632 P
0 F
0.43 ( action for) 203.06 632 P
4 F
1.04 (~q) 256.65 632 P
0 F
0.43 ( causes the Dvorak) 271.04 632 P
5 11 Q
0.44 (interact) 366.69 632 P
0 12 Q
0.43 ( to return the value) 402.71 632 P
5 11 Q
0.44 (continue) 498.47 632 P
0 12 Q
0.55 (to its caller) 72 618 P
0.55 (.) 125.74 618 P
5 11 Q
0.56 (interact) 135.83 618 P
0 12 Q
0.55 (\325) 171.86 618 P
0.55 (s caller happens to be an earlier) 175.19 618 P
5 11 Q
0.56 (interact) 333.24 618 P
0 12 Q
0.55 ( \050at the bottom of the script\051 which) 369.26 618 P
1.43 (evaluates the) 72 604 P
5 11 Q
1.46 (continue) 140.14 604 P
0 12 Q
1.43 ( and literally continues.  This isn\325) 181.67 604 P
1.43 (t anything magical.  They are just T) 350.94 604 P
1.43 (cl) 531.34 604 P
(commands that are appropriately handled.) 72 590 T
-0.01 (The script chooses the desired shell by examining the) 72 566 P
5 11 Q
-0.01 (SHELL) 331.38 566 P
0 12 Q
-0.01 ( environment variable.  Since) 366.18 566 P
2 F
-0.01 (print-) 510.02 566 P
-0.13 (env) 72 552 P
0 F
-0.13 ( appends a newline to the end of its output, this has to be stripped of) 89.99 552 P
-0.13 (f and is done here by) 414.31 552 P
5 11 Q
-0.13 (scan) 516.8 552 P
0 12 Q
(\320 an equivalent to scanf in the C programming language.) 72 538 T
0.37 (This script is excessive and is not at all what this feature of) 72 514 P
5 11 Q
0.37 (interact) 362.92 514 P
0 12 Q
0.37 ( was intended for) 398.94 514 P
0.37 (.  Neverthe-) 482.32 514 P
(less it works and demonstrates a number of interesting aspects.) 72 500 T
3 16 Q
(1) 108 459.33 T
(1.  Non-interactive programs are controlled dif) 115.12 459.33 T
(ferently) 422.66 459.33 T
0 12 Q
1.01 (Some interactive programs have non-interactive alternatives.  However) 72 434 P
1.01 (, it is often the case that) 420.33 434 P
(these alternatives are controlled in a way quite unlike the original interactive program.  Thus, you) 72 420 T
(need to learn two ways of doing things: interactively and non-interactively) 72 406 T
(.) 428.97 406 T
-0.13 (For example, suppose you want to locate a printer server) 72 382 P
-0.13 (.  This is described by the) 341.99 382 P
4 F
-0.3 (rm) 467.03 382 P
0 F
-0.13 ( value in the) 481.42 382 P
2 F
0.12 (printcap) 72 368 P
0 F
0.12 ( \336le.  Interactively) 115.98 368 P
0.12 (, you might use an editor) 203.49 368 P
0.12 (, or even,) 322.88 368 P
2 F
0.12 (mor) 370.86 368 P
0.12 (e) 391.96 368 P
0 F
0.12 (, to search the \336le for the cor-) 397.28 368 P
-0.12 (rect printer and then begin scanning for the) 72 354 P
4 F
-0.28 (rm) 280.89 354 P
0 F
-0.12 ( \336eld.  T) 295.28 354 P
-0.12 (o automate this, you must switch to a com-) 334.73 354 P
(pletely dif) 72 340 T
(ferent program, such as) 120.75 340 T
2 F
(awk) 236.32 340 T
0 F
(.) 257.65 340 T
-0.29 (Alternatively) 72 316 P
-0.29 (, you could just translate the interaction you were doing by hand into) 134.51 316 P
5 11 Q
-0.3 (send) 463.79 316 P
0 12 Q
-0.29 (/) 487.61 316 P
5 11 Q
-0.3 (expect) 490.95 316 P
0 12 Q
-0.29 ( se-) 523.31 316 P
0.01 (quences.  The following fragment does exactly this.  It was used by a lar) 72 302 P
0.01 (ger script that manipulat-) 419.04 302 P
(ed printers by running) 72 288 T
2 F
(lpc) 181.26 288 T
0 F
(, the interactive interface to the BSD line printer system.) 196.59 288 T
4 F
(spawn ed /etc/printcap) 115.2 268 T
(expect {*\134n}) 115.2 254.5 T
(    ;# discard character count) 207 254.5 T
(send "/$printer/\134r") 115.2 241 T
(for {} {1} {} {) 115.2 227.5 T
(expect "*\134r\134n*:rm=*\134n*" {) 162 214 T
(# found rm, now get value) 207 200.5 T
(set i [string first :rm= $expect_match]) 207 187 T
(scan [range $expect_match [expr $i+4] end c] \134) 207 173.5 T
("%\134[^:\134]" server) 252 160 T
(break) 207 146.5 T
(} "*\134r\134n*\134\134\134r\134n" {) 162 133 T
(# look at next line of entry) 207 119.5 T
(send "\134r") 207 106 T
(} "*\134r\134n*\134n" {) 162 92.5 T
(# no more lines in entry - give up) 207 79 T
FMENDPAGE
%%EndPage: "15" 14
%%Page: "14" 14
612 792 0 FMBEGINPAGE
72 36 504 47.95 R
7 X
0 K
V
0 12 Q
0 X
(14) 72 39.95 T
0 10 Q
(Don Libes) 108 39.95 T
72 740.81 504 756 R
7 X
V
72 63 540 720 R
V
0 12 Q
0 X
0.69 (to the) 72 712 P
2 F
0.69 (expect) 103.37 712 P
0 F
0.69 ( interpreter and invoke a procedure to answer the remaining questions without fur-) 136 712 P
(ther interaction from you.  This can be made as complex as you like.) 72 698 T
0.22 (The ar) 72 674 P
0.22 (guments to) 102.97 674 P
5 11 Q
0.22 (interact) 159.38 674 P
0 12 Q
0.22 ( are actually string-action pairs.  \050The default action is to invoke the in-) 195.4 674 P
0.73 (terpreter interactively) 72 660 P
0.73 (.\051  This generalized mechanism can support all the usual styles of escapes.) 175.52 660 P
1.11 (such as) 72 646 P
2 F
1.11 (tip) 112.19 646 P
0 F
1.11 (\325) 126.19 646 P
1.11 (s) 129.52 646 P
5 11 Q
1.13 (~) 138.3 646 P
0 12 Q
1.11 (-pre\336xed commands or) 144.72 646 P
2 F
1.11 (csh) 262.3 646 P
0 F
1.11 (\325) 278.96 646 P
1.11 (s single-character job control keys.  Actions may be) 282.29 646 P
-0.12 (any) 72 632 P
2 F
-0.12 (expect) 92.2 632 P
0 F
-0.12 ( command.  As an example, the following line maps the strings) 124.83 632 P
4 F
-0.28 (~y) 429.22 632 P
0 F
-0.12 (,) 442.83 632 P
4 F
-0.28 (~a) 448.71 632 P
0 F
-0.12 (, and the) 463.1 632 P
5 11 Q
-0.12 (^C) 506.72 632 P
0 12 Q
-0.12 ( and) 519.8 632 P
5 11 Q
(^Z) 72 618 T
0 12 Q
( characters.) 83.86 618 T
4 F
(interact \134) 115.2 598 T
(~y {yes} \134) 162 584.5 T
(~a {send "[exec date]"; send_user "hello world"} \134) 162 571 T
(\134Cc {exit} \134) 162 557.5 T
(\134Cz {exec kill -STOP 0}) 162 544 T
0 F
0.11 (When) 72 520.5 P
4 F
0.26 (~y) 103.75 520.5 P
0 F