1 | <HTML>
|
---|
2 | <HEAD>
|
---|
3 | <link rel="SHORTCUT ICON" href="http://www.cons.org/favicon.ico">
|
---|
4 | <TITLE>Proper handling of SIGINT/SIGQUIT [http://www.cons.org/cracauer/sigint.html]</TITLE>
|
---|
5 | <!-- Created by: GNU m4 using $Revision: 1.20 $ of crawww.m4lib on 11-Feb-2005 -->
|
---|
6 | <BODY BGCOLOR="#fff8e1">
|
---|
7 | <CENTER><H2>Proper handling of SIGINT/SIGQUIT</H2></CENTER>
|
---|
8 | <img src=linie.png width="100%" alt=" ">
|
---|
9 | <P>
|
---|
10 |
|
---|
11 | <table border=1 cellpadding=4>
|
---|
12 | <tr><th valign=top align=left>Abstract: </th>
|
---|
13 | <td valign=top align=left>
|
---|
14 | In UNIX terminal sessions, you usually have a key like
|
---|
15 | <code>C-c</code> (Control-C) to immediately end whatever program you
|
---|
16 | have running in the foreground. This should work even when the program
|
---|
17 | you called has called other programs in turn. Everything should be
|
---|
18 | aborted, giving you your command prompt back, no matter how deep the
|
---|
19 | call stack is.
|
---|
20 |
|
---|
21 | <p>Basically, it's trivial. But the existence of interactive
|
---|
22 | applications that use SIGINT and/or SIGQUIT for other purposes than a
|
---|
23 | complete immediate abort make matters complicated, and - as was to
|
---|
24 | expect - left us with several ways to solve the problems. Of course,
|
---|
25 | existing shells and applications follow different ways.
|
---|
26 |
|
---|
27 | <P>This Web pages outlines different ways to solve the problem and
|
---|
28 | argues that only one of them can do everything right, although it
|
---|
29 | means that we have to fix some existing software.
|
---|
30 |
|
---|
31 |
|
---|
32 |
|
---|
33 | </td></tr><tr><th valign=top align=left>Intended audience: </th>
|
---|
34 | <td valign=top align=left>Programmers who implement programs that catch SIGINT/SIGQUIT.
|
---|
35 | <BR>Programmers who implements shells or shell-like programs that
|
---|
36 | execute batches of programs.
|
---|
37 |
|
---|
38 | <p>Users who have problems problems getting rid of runaway shell
|
---|
39 | scripts using <code>Control-C</code>. Or have interactive applications
|
---|
40 | that don't behave right when sending SIGINT. Examples are emacs'es
|
---|
41 | that die on Control-g or shellscript statements that sometimes are
|
---|
42 | executed and sometimes not, apparently not determined by the user's
|
---|
43 | intention.
|
---|
44 |
|
---|
45 |
|
---|
46 | </td></tr><tr><th valign=top align=left>Required knowledge: </th>
|
---|
47 | <td valign=top align=left>You have to know what it means to catch SIGINT or SIGQUIT and how
|
---|
48 | processes are waiting for other processes (childs) they spawned.
|
---|
49 |
|
---|
50 |
|
---|
51 | </td></tr></table>
|
---|
52 | <img src=linie.png width="100%" alt=" ">
|
---|
53 |
|
---|
54 |
|
---|
55 | <H3>Basic concepts</H3>
|
---|
56 |
|
---|
57 | What technically happens when you press Control-C is that all programs
|
---|
58 | running in the foreground in your current terminal (or virtual
|
---|
59 | terminal) get the signal SIGINT sent.
|
---|
60 |
|
---|
61 | <p>You may change the key that triggers the signal using
|
---|
62 | <code>stty</code> and running programs may remap the SIGINT-sending
|
---|
63 | key at any time they like, without your intervention and without
|
---|
64 | asking you first.
|
---|
65 |
|
---|
66 | <p>The usual reaction of a running program to SIGINT is to exit.
|
---|
67 | However, not all program do an exit on SIGINT, programs are free to
|
---|
68 | use the signal for other actions or to ignore it at all.
|
---|
69 |
|
---|
70 | <p>All programs running in the foreground receive the signal. This may
|
---|
71 | be a nested "stack" of programs: You started a program that started
|
---|
72 | another and the outer is waiting for the inner to exit. This nesting
|
---|
73 | may be arbitrarily deep.
|
---|
74 |
|
---|
75 | <p>The innermost program is the one that decides what to do on SIGINT.
|
---|
76 | It may exit, do something else or do nothing. Still, when the user hit
|
---|
77 | SIGINT, all the outer programs are awaken, get the signal and may
|
---|
78 | react on it.
|
---|
79 |
|
---|
80 | <H3>What we try to achieve</H3>
|
---|
81 |
|
---|
82 | The problem is with shell scripts (or similar programs that call
|
---|
83 | several subprograms one after another).
|
---|
84 |
|
---|
85 | <p>Let us consider the most basic script:
|
---|
86 | <PRE>
|
---|
87 | #! /bin/sh
|
---|
88 | program1
|
---|
89 | program2
|
---|
90 | </PRE>
|
---|
91 | and the usual run looks like this:
|
---|
92 | <PRE>
|
---|
93 | $ sh myscript
|
---|
94 | [output of program1]
|
---|
95 | [output of program2]
|
---|
96 | $
|
---|
97 | </PRE>
|
---|
98 |
|
---|
99 | <p>Let us assume that both programs do nothing special on SIGINT, they
|
---|
100 | just exit.
|
---|
101 |
|
---|
102 | <p>Now imagine the user hits C-c while a shellscript is executing its
|
---|
103 | first program. The following programs receive SIGINT: program1 and
|
---|
104 | also the shell executing the script. program1 exits.
|
---|
105 |
|
---|
106 | <p>But what should the shell do? If we say that it is only the
|
---|
107 | innermost's programs business to react on SIGINT, the shell will do
|
---|
108 | nothing special (not exit) and it will continue the execution of the
|
---|
109 | script and run program2. But this is wrong: The user's intention in
|
---|
110 | hitting C-c is to abort the whole script, to get his prompt back. If
|
---|
111 | he hits C-c while the first program is running, he does not want
|
---|
112 | program2 to be even started.
|
---|
113 |
|
---|
114 | <p>here is what would happen if the shell doesn't do anything:
|
---|
115 | <PRE>
|
---|
116 | $ sh myscript
|
---|
117 | [first half of program1's output]
|
---|
118 | C-c [users presses C-c]
|
---|
119 | [second half of program1's output will not be displayed]
|
---|
120 | [output of program2 will appear]
|
---|
121 | </PRE>
|
---|
122 |
|
---|
123 |
|
---|
124 | <p>Consider a more annoying example:
|
---|
125 | <pre>
|
---|
126 | #! /bin/sh
|
---|
127 | # let's assume there are 300 *.dat files
|
---|
128 | for file in *.dat ; do
|
---|
129 | dat2ascii $dat
|
---|
130 | done
|
---|
131 | </pre>
|
---|
132 |
|
---|
133 | If your shell wouldn't end if the user hits <code>C-c</code>,
|
---|
134 | <code>C-c</code> would just end <strong>one</strong> dat2ascii run and
|
---|
135 | the script would continue. Thus, you had to hit <code>C-c</code> up to
|
---|
136 | 300 times to end this script.
|
---|
137 |
|
---|
138 | <H3>Alternatives to do so</H3>
|
---|
139 |
|
---|
140 | <p>There are several ways to handle abortion of shell scripts when
|
---|
141 | SIGINT is received while a foreground child runs:
|
---|
142 |
|
---|
143 | <menu>
|
---|
144 |
|
---|
145 | <li>As just outlined, the shellscript may just continue, ignoring the
|
---|
146 | fact that the user hit <code>C-c</code>. That way, your shellscript -
|
---|
147 | including any loops - would continue and you had no chance of aborting
|
---|
148 | it except using the kill command after finding out the outermost
|
---|
149 | shell's PID. This "solution" will not be discussed further, as it is
|
---|
150 | obviously not desirable.
|
---|
151 |
|
---|
152 | <p><li>The shell itself exits immediately when it receives SIGINT. Not
|
---|
153 | only the program called will exit, but the calling (the
|
---|
154 | script-executing) shell. The first variant is to exit the shell (and
|
---|
155 | therefore discontinuing execution of the script) immediately, while
|
---|
156 | the background program may still be executing (remember that although
|
---|
157 | the shell is just waiting for the called program to exit, it is woken
|
---|
158 | up and may act). I will call the way of doing things the "IUE" (for
|
---|
159 | "immediate unconditional exit") for the rest of this document.
|
---|
160 |
|
---|
161 | <p><li>As a variant of the former, when the shell receives SIGINT
|
---|
162 | while it is waiting for a child to exit, the shell does not exit
|
---|
163 | immediately. but it remembers the fact that a SIGINT happened. After
|
---|
164 | the called program exits and the shell's wait ends, the shell will
|
---|
165 | exit itself and hence discontinue the script. I will call the way of
|
---|
166 | doing things the "WUE" (for "wait and unconditional exit") for the
|
---|
167 | rest of this document.
|
---|
168 |
|
---|
169 | <p><li>There is also a way that the calling shell can tell whether the
|
---|
170 | called program exited on SIGINT and if it ignored SIGINT (or used it
|
---|
171 | for other purposes). As in the <sl>WUE</sl> way, the shell waits for
|
---|
172 | the child to complete. It figures whether the program was ended on
|
---|
173 | SIGINT and if so, it discontinue the script. If the program did any
|
---|
174 | other exit, the script will be continued. I will call the way of doing
|
---|
175 | things the "WCE" (for "wait and cooperative exit") for the rest of
|
---|
176 | this document.
|
---|
177 |
|
---|
178 | </menu>
|
---|
179 |
|
---|
180 | <H3>The problem</H3>
|
---|
181 |
|
---|
182 | On first sight, all three solutions (IUE, WUE and WCE) all seem to do
|
---|
183 | what we want: If C-c is hit while the first program of the shell
|
---|
184 | script runs, the script is discontinued. The user gets his prompt back
|
---|
185 | immediately. So what are the difference between these way of handling
|
---|
186 | SIGINT?
|
---|
187 |
|
---|
188 | <p>There are programs that use the signal SIGINT for other purposes
|
---|
189 | than exiting. They use it as a normal keystroke. The user is expected
|
---|
190 | to use the key that sends SIGINT during a perfectly normal program
|
---|
191 | run. As a result, the user sends SIGINT in situations where he/she
|
---|
192 | does not want the program or the script to end.
|
---|
193 |
|
---|
194 | <p>The primary example is the emacs editor: C-g does what ESC does in
|
---|
195 | other applications: It cancels a partially executed or prepared
|
---|
196 | operation. Technically, emacs remaps the key that sends SIGINT from
|
---|
197 | C-c to C-g and catches SIGINT.
|
---|
198 |
|
---|
199 | <p>Remember that the SIGINT is sent to all programs running in the
|
---|
200 | foreground. If emacs is executing from a shell script, both emacs and
|
---|
201 | the shell get SIGINT. emacs is the program that decides what to do:
|
---|
202 | Exit on SIGINT or not. emacs decides not to exit. The problem arises
|
---|
203 | when the shell draws its own conclusions from receiving SIGINT without
|
---|
204 | consulting emacs for its opinion.
|
---|
205 |
|
---|
206 | <p>Consider this script:
|
---|
207 | <PRE>
|
---|
208 | #! /bin/sh
|
---|
209 | emacs /tmp/foo
|
---|
210 | cp /tmp/foo /home/user/mail/sent
|
---|
211 | </PRE>
|
---|
212 |
|
---|
213 | <p>If C-g is used in emacs, both the shell and emacs will received
|
---|
214 | SIGINT. Emacs will not exit, the user used C-g as a normal editing
|
---|
215 | keystroke, he/she does not want the script to be aborted on C-g.
|
---|
216 |
|
---|
217 | <p>The central problem is that the second command (cp) may
|
---|
218 | unintentionally be killed when the shell draws its own conclusion
|
---|
219 | about the user's intention. The innermost program is the only one to
|
---|
220 | judge.
|
---|
221 |
|
---|
222 | <H3>One more example</H3>
|
---|
223 |
|
---|
224 | <p>Imagine a mail session using a curses mailer in a tty. You called
|
---|
225 | your mailer and started to compose a message. Your mailer calls emacs.
|
---|
226 | <code>C-g</code> is a normal editing key in emacs. Technically it
|
---|
227 | sends SIGINT (it was <code>C-c</code>, but emacs remapped the key) to
|
---|
228 | <menu>
|
---|
229 | <li>emacs
|
---|
230 | <li>the shell between your mailer and emacs, the one from your mailers
|
---|
231 | system("emacs /tmp/bla.44") command
|
---|
232 | <li>the mailer itself
|
---|
233 | <li>possibly another shell if your mailer was called by a shell script
|
---|
234 | or from another application using system(3)
|
---|
235 | <li>your interactive shell (which ignores it since it is interactive
|
---|
236 | and hence is not relevant to this discussion)
|
---|
237 | </menu>
|
---|
238 |
|
---|
239 | <p>If everyone just exits on SIGINT, you will be left with nothing but
|
---|
240 | your login shell, without asking.
|
---|
241 |
|
---|
242 | <p>But for sure you don't want to be dropped out of your editor and
|
---|
243 | out of your mailer back to the commandline, having your edited data
|
---|
244 | and mailer status deleted.
|
---|
245 |
|
---|
246 | <p>Understand the difference: While <code>C-g</code> is used an a kind
|
---|
247 | of abort key in emacs, it isn't the major "abort everything" key. When
|
---|
248 | you use <code>C-g</code> in emacs, you want to end some internal emacs
|
---|
249 | command. You don't want your whole emacs and mailer session to end.
|
---|
250 |
|
---|
251 | <p>So, if the shell exits immediately if the user sends SIGINT (the
|
---|
252 | second of the four ways shown above), the parent of emacs would die,
|
---|
253 | leaving emacs without the controlling tty. The user will lose it's
|
---|
254 | editing session immediately and unrecoverable. If the "main" shell of
|
---|
255 | the operating system defaults to this behavior, every editor session
|
---|
256 | that is spawned from a mailer or such will break (because it is
|
---|
257 | usually executed by system(3), which calls /bin/sh). This was the case
|
---|
258 | in FreeBSD before I and Bruce Evans changed it in 1998.
|
---|
259 |
|
---|
260 | <p>If the shell recognized that SIGINT was sent and exits after the
|
---|
261 | current foreground process exited (the third way of the four), the
|
---|
262 | editor session will not be disturbed, but things will still not work
|
---|
263 | right.
|
---|
264 |
|
---|
265 | <H3>A further look at the alternatives</H3>
|
---|
266 |
|
---|
267 | <p>Still considering this script to examine the shell's actions in the
|
---|
268 | IUE, WUE and ICE way of handling SIGINT:
|
---|
269 | <PRE>
|
---|
270 | #! /bin/sh
|
---|
271 | emacs /tmp/foo
|
---|
272 | cp /tmp/foo /home/user/mail/sent
|
---|
273 | </PRE>
|
---|
274 |
|
---|
275 | <p>The IUE ("immediate unconditional exit") way does not work at all:
|
---|
276 | emacs wants to survive the SIGINT (it's a normal editing key for
|
---|
277 | emacs), but its parent shell unconditionally thinks "We received
|
---|
278 | SIGINT. Abort everything. Now.". The shell will exit even before emacs
|
---|
279 | exits. But this will leave emacs in an unusable state, since the death
|
---|
280 | of its calling shell will leave it without required resources (file
|
---|
281 | descriptors). This way does not work at all for shellscripts that call
|
---|
282 | programs that use SIGINT for other purposes than immediate exit. Even
|
---|
283 | for programs that exit on SIGINT, but want to do some cleanup between
|
---|
284 | the signal and the exit, may fail before they complete their cleanup.
|
---|
285 |
|
---|
286 | <p>It should be noted that this way has one advantage: If a child
|
---|
287 | blocks SIGINT and does not exit at all, this way will get control back
|
---|
288 | to the user's terminal. Since such programs should be banned from your
|
---|
289 | system anyway, I don't think that weighs against the disadvantages.
|
---|
290 |
|
---|
291 | <p>WUE ("wait and unconditional exit") is a little more clever: If C-g
|
---|
292 | was used in emacs, the shell will get SIGINT. It will not immediately
|
---|
293 | exit, but remember the fact that a SIGINT happened. When emacs ends
|
---|
294 | (maybe a long time after the SIGINT), it will say "Ok, a SIGINT
|
---|
295 | happened sometime while the child was executing, the user wants the
|
---|
296 | script to be discontinued". It will then exit. The cp will not be
|
---|
297 | executed. But that's bad. The "cp" will be executed when the emacs
|
---|
298 | session ended without the C-g key ever used, but it will not be
|
---|
299 | executed when the user used C-g at least one time. That is clearly not
|
---|
300 | desired. Since C-g is a normal editing key in emacs, the user expects
|
---|
301 | the rest of the script to behave identically no matter what keys he
|
---|
302 | used.
|
---|
303 |
|
---|
304 | <p>As a result, the "WUE" way is better than the "IUE" way in that it
|
---|
305 | does not break SIGINT-using programs completely. The emacs session
|
---|
306 | will end undisturbed. But it still does not support scripts where
|
---|
307 | other actions should be performed after a program that use SIGINT for
|
---|
308 | non-exit purposes. Since the behavior is basically undeterminable for
|
---|
309 | the user, this can lead to nasty surprises.
|
---|
310 |
|
---|
311 | <p>The "WCE" way fixes this by "asking" the called program whether it
|
---|
312 | exited on SIGINT or not. While emacs receives SIGINT, it does not exit
|
---|
313 | on it and a calling shell waiting for its exit will not be told that
|
---|
314 | it exited on SIGINT. (Although it receives SIGINT at some point in
|
---|
315 | time, the system does not enforce that emacs will exit with
|
---|
316 | "I-exited-on-SIGINT" status. This is under emacs' control, see below).
|
---|
317 |
|
---|
318 | <p>this still work for the normal script without SIGINT-using
|
---|
319 | programs:</p>
|
---|
320 | <PRE>
|
---|
321 | #! /bin/sh
|
---|
322 | program1
|
---|
323 | program2
|
---|
324 | </PRE>
|
---|
325 |
|
---|
326 | Unless program1 and program2 mess around with signal handling, the
|
---|
327 | system will tell the calling shell whether the programs exited
|
---|
328 | normally or as a result of SIGINT.
|
---|
329 |
|
---|
330 | <p>The "WCE" way then has an easy way to things right: When one called
|
---|
331 | program exited with "I-exited-on-SIGINT" status, it will discontinue
|
---|
332 | the script after this program. If the program ends without this
|
---|
333 | status, the next command in the script is started.
|
---|
334 |
|
---|
335 | <p>It is important to understand that a shell in "WCE" modus does not
|
---|
336 | need to listen to the SIGINT signal at all. Both in the
|
---|
337 | "emacs-then-cp" script and in the "several-normal-programs" script, it
|
---|
338 | will be woken up and receive SIGINT when the user hits the
|
---|
339 | corresponding key. But the shell does not need to react on this event
|
---|
340 | and it doesn't need to remember the event of any SIGINT, either.
|
---|
341 | Telling whether the user wants to end a script is done by asking that
|
---|
342 | program that has to decide, that program that interprets keystrokes
|
---|
343 | from the user, the innermost program.
|
---|
344 |
|
---|
345 | <H3>So everything is well with WCE?</H3>
|
---|
346 |
|
---|
347 | Well, almost.
|
---|
348 |
|
---|
349 | <p>The problem with the "WCE" modus is that there are broken programs
|
---|
350 | that do not properly communicate the required information up to the
|
---|
351 | calling program.
|
---|
352 |
|
---|
353 | <p>Unless a program messes with signal handling, the system does this
|
---|
354 | automatically.
|
---|
355 |
|
---|
356 | <p>There are programs that want to exit on SIGINT, but they don't let
|
---|
357 | the system do the automatic exit, because they want to do some
|
---|
358 | cleanup. To do so, they catch SIGINT, do the cleanup and then exit by
|
---|
359 | themselves.
|
---|
360 |
|
---|
361 | <p>And here is where the problem arises: Once they catch the signal,
|
---|
362 | the system will no longer communicate the "I-exited-on-SIGINT" status
|
---|
363 | to the calling program automatically. Even if the program exit
|
---|
364 | immediately in the signal handler of SIGINT. Once it catches the
|
---|
365 | signal, it has to take care of communicating the signal status
|
---|
366 | itself.
|
---|
367 |
|
---|
368 | <p>Some programs don't do this. On SIGINT, they do cleanup and exit
|
---|
369 | immediatly, but the calling shell isn't told about the non-normal exit
|
---|
370 | and it will call the next program in the script.
|
---|
371 |
|
---|
372 | <p>As a result, the user hits SIGINT and while one program exits, the
|
---|
373 | shellscript continues. To him/her it looks like the shell fails to
|
---|
374 | obey to his abortion command.
|
---|
375 |
|
---|
376 | <p>Both IUE or WUE shell would not have this problem, since they
|
---|
377 | discontinue the script on their own. But as I said, they don't support
|
---|
378 | programs using SIGINT for non-exiting purposes, no matter whether
|
---|
379 | these programs properly communicate their signal status to the calling
|
---|
380 | shell or not.
|
---|
381 |
|
---|
382 | <p>Since some shell in wide use implement the WUE way (and some even
|
---|
383 | IUE), there is a considerable number of broken programs out there that
|
---|
384 | break WCE shells. The programmers just don't recognize it if their
|
---|
385 | shell isn't WCE.
|
---|
386 |
|
---|
387 | <H3>How to be a proper program</H3>
|
---|
388 |
|
---|
389 | <p>(Short note in advance: What you need to achieve is that
|
---|
390 | WIFSIGNALED(status) is true in the calling program and that
|
---|
391 | WTERMSIG(status) returns SIGINT.)
|
---|
392 |
|
---|
393 | <p>If you don't catch SIGINT, the system automatically does the right
|
---|
394 | thing for you: Your program exits and the calling program gets the
|
---|
395 | right "I-exited-on-SIGINT" status after waiting for your exit.
|
---|
396 |
|
---|
397 | <p>But once you catch SIGINT, you have to act.
|
---|
398 |
|
---|
399 | <p>Decide whether the SIGINT is used for exit/abort purposes and hence
|
---|
400 | a shellscript calling this program should discontinue. This is
|
---|
401 | hopefully obvious. If you just need to do some cleanup on SIGINT, but
|
---|
402 | then exit immediately, the answer is "yes".
|
---|
403 |
|
---|
404 | <p>If so, you have to tell the calling program about it by exiting
|
---|
405 | with the "I-exited-on-SIGINT" status.
|
---|
406 |
|
---|
407 | <p>There is no other way of doing this than to kill yourself with a
|
---|
408 | SIGINT signal. Do it by resetting the SIGINT handler to SIG_DFL, then
|
---|
409 | send yourself the signal.
|
---|
410 |
|
---|
411 | <PRE>
|
---|
412 | void sigint_handler(int sig)
|
---|
413 | {
|
---|
414 | <do some cleanup>
|
---|
415 | signal(SIGINT, SIG_DFL);
|
---|
416 | kill(getpid(), SIGINT);
|
---|
417 | }
|
---|
418 | </PRE>
|
---|
419 |
|
---|
420 | Notes:
|
---|
421 |
|
---|
422 | <MENU>
|
---|
423 |
|
---|
424 | <LI>You cannot "fake" the proper exit status by an exit(3) with a
|
---|
425 | special numeric value. People often assume this since the manuals for
|
---|
426 | shells often list some return value for exactly this. But this is just
|
---|
427 | a convention for your shell script. It does not work from one UNIX API
|
---|
428 | program to another.
|
---|
429 |
|
---|
430 | <P>All that happens is that the shell sets the "$?" variable to a
|
---|
431 | special numeric value for the convenience of your script, because your
|
---|
432 | script does not have access to the lower-lever UNIX status evaluation
|
---|
433 | functions. This is just an agreement between your script and the
|
---|
434 | executing shell, it does not have any meaning in other contexts.
|
---|
435 |
|
---|
436 | <P><LI>Do not use kill(0, SIGINT) without consulting the manul for
|
---|
437 | your OS implementation. I.e. on BSD, this would not send the signal to
|
---|
438 | the current process, but to all processes in the group.
|
---|
439 |
|
---|
440 | <P><LI>POSIX 1003.1 allows all these calls to appear in signal
|
---|
441 | handlers, so it is portable.
|
---|
442 |
|
---|
443 | </MENU>
|
---|
444 |
|
---|
445 | <p>In a bourne shell script, you can catch signals using the
|
---|
446 | <code>trap</code> command. Here, the same as for C programs apply. If
|
---|
447 | the intention of SIGINT is to end your program, you have to exit in a
|
---|
448 | way that the calling programs "sees" that you have been killed. If
|
---|
449 | you don't catch SIGINT, this happend automatically, but of you catch
|
---|
450 | SIGINT, i.e. to do cleanup work, you have to end the program by
|
---|
451 | killing yourself, not by calling exit.
|
---|
452 |
|
---|
453 | <p>Consider this example from FreeBSD's <code>mkdep</code>, which is a
|
---|
454 | bourne shell script.
|
---|
455 |
|
---|
456 | <pre>
|
---|
457 | TMP=_mkdep$$
|
---|
458 | trap 'rm -f $TMP ; trap 2 ; kill -2 $$' 1 2 3 13 15
|
---|
459 | </pre>
|
---|
460 |
|
---|
461 | Yes, you have to do it the hard way. It's even more annoying in shell
|
---|
462 | scripts than in C programs since you can't "pre-delete" temporary
|
---|
463 | files (which isn't really portable in C, though).
|
---|
464 |
|
---|
465 | <P>All this applies to programs in all languages, not only C and
|
---|
466 | bourne shell. Every language implementation that lets you catch SIGINT
|
---|
467 | should also give you the option to reset the signal and kill yourself.
|
---|
468 |
|
---|
469 | <P>It is always desireable to exit the right way, even if you don't
|
---|
470 | expect your usual callers to depend on it, some unusual one will come
|
---|
471 | along. This proper exit status will be needed for WCE and will not
|
---|
472 | hurt when the calling shell uses IUE or WUE.
|
---|
473 |
|
---|
474 | <H3>How to be a proper shell</H3>
|
---|
475 |
|
---|
476 | All this applies only for the script-executing case. Most shells will
|
---|
477 | also have interactive modes where things are different.
|
---|
478 |
|
---|
479 | <MENU>
|
---|
480 |
|
---|
481 | <LI>Do nothing special when SIGINT appears while you wait for a child.
|
---|
482 | You don't even have to remember that one happened.
|
---|
483 |
|
---|
484 | <P><LI>Wait for child to exit, get the exit status. Do not truncate it
|
---|
485 | to type char.
|
---|
486 |
|
---|
487 | <P><LI>Look at WIFSIGNALED(status) and WTERMSIG(status) to tell
|
---|
488 | whether the child says "I exited on SIGINT: in my opinion the user
|
---|
489 | wants the shellscript to be discontinued".
|
---|
490 |
|
---|
491 | <P><LI>If the latter applies, discontinue the script.
|
---|
492 |
|
---|
493 | <P><LI>Exit. But since a shellscript may in turn be called by a
|
---|
494 | shellscript, you need to make sure that you properly communicate the
|
---|
495 | discontinue intention to the calling program. As in any other program
|
---|
496 | (see above), do
|
---|
497 |
|
---|
498 | <PRE>
|
---|
499 | signal(SIGINT, SIG_DFL);
|
---|
500 | kill(getpid(), SIGINT);
|
---|
501 | </PRE>
|
---|
502 |
|
---|
503 | </MENU>
|
---|
504 |
|
---|
505 | <H3>Other remarks</H3>
|
---|
506 |
|
---|
507 | Although this web page talks about SIGINT only, almost the same issues
|
---|
508 | apply to SIGQUIT, including proper exiting by killing yourself after
|
---|
509 | catching the signal and proper reaction on the WIFSIGNALED(status)
|
---|
510 | value. One notable difference for SIGQUIT is that you have to make
|
---|
511 | sure that not the whole call tree dumps core.
|
---|
512 |
|
---|
513 | <H3>What to fight</H3>
|
---|
514 |
|
---|
515 | Make sure all programs <em>really</em> kill themselves if they react
|
---|
516 | to SIGINT or SIGQUIT and intend to abort their operation as a result
|
---|
517 | of this signal. Programs that don't use SIGINT/SIGQUIT as a
|
---|
518 | termination trigger - but as part of normal operation - don't kill
|
---|
519 | themselves, but do a normal exit instead.
|
---|
520 |
|
---|
521 | <p>Make sure people understand why you can't fake an exit-on-signal by
|
---|
522 | doing exit(...) using any numerical status.
|
---|
523 |
|
---|
524 | <p>Make sure you use a shell that behaves right. Especially if you
|
---|
525 | develop programs, since it will help seeing problems.
|
---|
526 |
|
---|
527 | <H3>Concrete examples how to fix programs:</H3>
|
---|
528 | <ul>
|
---|
529 |
|
---|
530 | <li>The fix for FreeBSD's
|
---|
531 | <A HREF="http://www.freebsd.org/cgi/cvsweb.cgi/src/usr.bin/time/time.c.diff?r1=1.10&r2=1.11">time(1)</A>. This fix is the best example, it's quite short and clear and
|
---|
532 | it fixes a case where someone tried to fake signal exit status by a
|
---|
533 | numerical value. And the complete program is small.
|
---|
534 |
|
---|
535 | <p><li>Fix for FreeBSD's
|
---|
536 | <A HREF="http://www.freebsd.org/cgi/cvsweb.cgi/src/usr.bin/truss/main.c.diff?r1=1.9&r2=1.10">truss(1)</A>.
|
---|
537 |
|
---|
538 | <p><li>The fix for FreeBSD's
|
---|
539 | <A HREF="http://www.freebsd.org/cgi/cvsweb.cgi/src/usr.bin/mkdep/mkdep.gcc.sh.diff?r1=1.8.2.1&r2=1.8.2.2">mkdep(1)</A>, a shell script.
|
---|
540 |
|
---|
541 |
|
---|
542 | <p><li>Fix for FreeBSD's make(1), <A HREF="http://www.freebsd.org/cgi/cvsweb.cgi/src/usr.bin/make/job.c.diff?r1=1.9&r2=1.10">part 1</A>,
|
---|
543 | <A HREF="http://www.freebsd.org/cgi/cvsweb.cgi/src/usr.bin/make/compat.c.diff?r1=1.10&r2=1.11">part 2</A>.
|
---|
544 |
|
---|
545 | </ul>
|
---|
546 |
|
---|
547 | <H3>Testsuite for shells</H3>
|
---|
548 |
|
---|
549 | I have a collection of shellscripts that test shells for the
|
---|
550 | behavior. See my <A HREF="download/">download dir</A> to get the newest
|
---|
551 | "sh-interrupt" files, either as a tarfile or as individual file for
|
---|
552 | online browsing. This isn't really documented, besides from the
|
---|
553 | comments the scripts echo.
|
---|
554 |
|
---|
555 | <H3>Appendix 1 - table of implementation choices</H3>
|
---|
556 |
|
---|
557 | <table border cellpadding=2>
|
---|
558 |
|
---|
559 | <tr valign=top>
|
---|
560 | <th>Method sign</th>
|
---|
561 | <th>Does what?</th>
|
---|
562 | <th>Example shells that implement it:</th>
|
---|
563 | <th>What happens when a shellscript called emacs, the user used
|
---|
564 | <code>C-g</code> and the script has additional commands in it?</th>
|
---|
565 | <th>What happens when a shellscript called emacs, the user did not use
|
---|
566 | <code>C-c</code> and the script has additional commands in it?</th>
|
---|
567 | <th>What happens if a non-interactive child catches SIGINT?</th>
|
---|
568 | <th>To behave properly, childs must do what?</th>
|
---|
569 | </tr>
|
---|
570 |
|
---|
571 | <tr valign=top align=left>
|
---|
572 | <td>IUE</td>
|
---|
573 | <td>The shell executing a script exits immediately if it receives
|
---|
574 | SIGINT.</td>
|
---|
575 | <td>4.4BSD ash (ash), NetBSD, FreeBSD prior to 3.0/22.8</td>
|
---|
576 | <td>The editor session is lost and subsequent commands are not
|
---|
577 | executed.</td>
|
---|
578 | <td>The editor continues as normal and the subsequent commands are
|
---|
579 | executed. </td>
|
---|
580 | <td>The scripts ends immediately, returning to the caller even before
|
---|
581 | the current foreground child of the shell exits. </td>
|
---|
582 | <td>It doesn't matter what the child does or how it exits, even if the
|
---|
583 | child continues to operate, the shell returns. </td>
|
---|
584 | </tr>
|
---|
585 |
|
---|
586 | <tr valign=top align=left>
|
---|
587 | <td>WUE</td>
|
---|
588 | <td>If the shell executing a script received SIGINT while a foreground
|
---|
589 | process was running, it will exit after that child's exit.</td>
|
---|
590 | <td>pdksh (OpenBSD /bin/sh)</td>
|
---|
591 | <td>The editor continues as normal, but subsequent commands from the
|
---|
592 | script are not executed.</td>
|
---|
593 | <td>The editor continues as normal and subsequent commands are
|
---|
594 | executed. </td>
|
---|
595 | <td>The scripts returns to its caller after the current foreground
|
---|
596 | child exits, no matter how the child exited. </td>
|
---|
597 | <td>It doesn't matter how the child exits (signal status or not), but
|
---|
598 | if it doesn't return at all, the shell will not return. In no case
|
---|
599 | will further commands from the script be executed. </td>
|
---|
600 | </tr>
|
---|
601 |
|
---|
602 | <tr valign=top align=left>
|
---|
603 | <td>WCE</td>
|
---|
604 | <td>The shell exits if a child signaled that it was killed on a
|
---|
605 | signal (either it had the default handler for SIGINT or it killed
|
---|
606 | itself). </td>
|
---|
607 | <td>bash (Linux /bin/sh), most commercial /bin/sh, FreeBSD /bin/sh
|
---|
608 | from 3.0/2.2.8.</td>
|
---|
609 | <td>The editor continues as normal and subsequent commands are
|
---|
610 | executed. </td>
|
---|
611 | <td>The editor continues as normal and subsequent commands are
|
---|
612 | executed. </td>
|
---|
613 | <td>The scripts returns to its caller after the current foreground
|
---|
614 | child exits, but only if the child exited with signal status. If
|
---|
615 | the child did a normal exit (even if it received SIGINT, but catches
|
---|
616 | it), the script will continue. </td>
|
---|
617 | <td>The child must be implemented right, or the user will not be able
|
---|
618 | to break shell scripts reliably.</td>
|
---|
619 | </tr>
|
---|
620 |
|
---|
621 | </table>
|
---|
622 |
|
---|
623 | <P><img src=linie.png width="100%" alt=" ">
|
---|
624 | <BR>©2005 Martin Cracauer <cracauer @ cons.org>
|
---|
625 | <A HREF="http://www.cons.org/cracauer/">http://www.cons.org/cracauer/</A>
|
---|
626 | <BR>Last changed: $Date: 2005/02/11 21:44:43 $
|
---|
627 | </BODY></HTML>
|
---|