/* $OpenBSD: arithmetic.c,v 1.21 2015/10/24 17:13:27 mmcc Exp $ */ /* * Copyright (c) 1989, 1993 * The Regents of the University of California. All rights reserved. * * This code is derived from software contributed to Berkeley by * Eamonn McManus of Trinity College Dublin. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ /* * By Eamonn McManus, Trinity College Dublin . * * The operation of this program mimics that of the standard Unix game * `arithmetic'. I've made it as close as I could manage without examining * the source code. The principal differences are: * * The method of biasing towards numbers that had wrong answers in the past * is different; original `arithmetic' seems to retain the bias forever, * whereas this program lets the bias gradually decay as it is used. * * Original `arithmetic' delays for some period (3 seconds?) after printing * the score. I saw no reason for this delay, so I scrapped it. * * There is no longer a limitation on the maximum range that can be supplied * to the program. The original program required it to be less than 100. * Anomalous results may occur with this program if ranges big enough to * allow overflow are given. * * I have obviously not attempted to duplicate bugs in the original. It * would go into an infinite loop if invoked as `arithmetic / 0'. It also * did not recognise an EOF in its input, and would continue trying to read * after it. It did not check that the input was a valid number, treating any * garbage as 0. Finally, it did not flush stdout after printing its prompt, * so in the unlikely event that stdout was not a terminal, it would not work * properly. */ #include #include #include #include #include #include #include #include #include int getrandom(int, int, int); void intr(int); int opnum(int); void penalise(int, int, int); int problem(void); void showstats(void); void usage(void); const char keylist[] = "+-x/"; const char defaultkeys[] = "+-"; const char *keys = defaultkeys; int nkeys = sizeof(defaultkeys) - 1; int rangemax = 10; int nright, nwrong; time_t qtime; #define NQUESTS 20 /* * Select keys from +-x/ to be asked addition, subtraction, multiplication, * and division problems. More than one key may be given. The default is * +-. Specify a range to confine the operands to 0 - range. Default upper * bound is 10. After every NQUESTS questions, statistics on the performance * so far are printed. */ int main(int argc, char *argv[]) { extern char *optarg; extern int optind; int ch, cnt; if (pledge("stdio", NULL) == -1) err(1, "pledge"); while ((ch = getopt(argc, argv, "hr:o:")) != -1) switch(ch) { case 'o': { const char *p; for (p = keys = optarg; *p; ++p) if (!strchr(keylist, *p)) errx(1, "unknown key."); nkeys = p - optarg; break; } case 'r': if ((rangemax = atoi(optarg)) <= 0) errx(1, "invalid range."); break; case '?': case 'h': default: usage(); } if (argc -= optind) usage(); (void)signal(SIGINT, intr); /* Now ask the questions. */ for (;;) { for (cnt = NQUESTS; cnt--;) if (problem() == EOF) intr(0); /* Print score and exit */ showstats(); } /* NOTREACHED */ } /* Handle interrupt character. Print score and exit. */ void intr(int dummy) { showstats(); exit(0); } /* Print score. Original `arithmetic' had a delay after printing it. */ void showstats(void) { if (nright + nwrong > 0) { (void)printf("\n\nRights %d; Wrongs %d; Score %d%%", nright, nwrong, (int)(100L * nright / (nright + nwrong))); if (nright > 0) (void)printf("\nTotal time %ld seconds; %.1f seconds per problem\n\n", (long)qtime, (float)qtime / nright); } (void)printf("\n"); } /* * Pick a problem and ask it. Keeps asking the same problem until supplied * with the correct answer, or until EOF or interrupt is typed. Problems are * selected such that the right operand and either the left operand (for +, x) * or the correct result (for -, /) are in the range 0 to rangemax. Each wrong * answer causes the numbers in the problem to be penalised, so that they are * more likely to appear in subsequent problems. */ int problem(void) { char *p; time_t start, finish; int left, op, right, result; char line[80]; op = keys[arc4random_uniform(nkeys)]; if (op != '/') right = getrandom(rangemax + 1, op, 1); retry: /* Get the operands. */ switch (op) { case '+': left = getrandom(rangemax + 1, op, 0); result = left + right; break; case '-': result = getrandom(rangemax + 1, op, 0); left = right + result; break; case 'x': left = getrandom(rangemax + 1, op, 0); result = left * right; break; case '/': right = getrandom(rangemax, op, 1) + 1; result = getrandom(rangemax + 1, op, 0); left = right * result + arc4random_uniform(right); break; } /* * A very big maxrange could cause negative values to pop * up, owing to overflow. */ if (result < 0 || left < 0) goto retry; (void)printf("%d %c %d = ", left, op, right); (void)fflush(stdout); (void)time(&start); /* * Keep looping until the correct answer is given, or until EOF or * interrupt is typed. */ for (;;) { if (!fgets(line, sizeof(line), stdin)) { (void)printf("\n"); return(EOF); } for (p = line; isspace((unsigned char)*p); ++p); if (!isdigit((unsigned char)*p)) { (void)printf("Please type a number.\n"); continue; } if (atoi(p) == result) { (void)printf("Right!\n"); ++nright; break; } /* Wrong answer; penalise and ask again. */ (void)printf("What?\n"); ++nwrong; penalise(right, op, 1); if (op == 'x' || op == '+') penalise(left, op, 0); else penalise(result, op, 0); } /* * Accumulate the time taken. Obviously rounding errors happen here; * however they should cancel out, because some of the time you are * charged for a partially elapsed second at the start, and some of * the time you are not charged for a partially elapsed second at the * end. */ (void)time(&finish); qtime += finish - start; return(0); } /* * Here is the code for accumulating penalties against the numbers for which * a wrong answer was given. The right operand and either the left operand * (for +, x) or the result (for -, /) are stored in a list for the particular * operation, and each becomes more likely to appear again in that operation. * Initially, each number is charged a penalty of WRONGPENALTY, giving it that * many extra chances of appearing. Each time it is selected because of this, * its penalty is decreased by one; it is removed when it reaches 0. * * The penalty[] array gives the sum of all penalties in the list for * each operation and each operand. The penlist[] array has the lists of * penalties themselves. */ int penalty[sizeof(keylist) - 1][2]; struct penalty { int value, penalty; /* Penalised value and its penalty. */ struct penalty *next; } *penlist[sizeof(keylist) - 1][2]; #define WRONGPENALTY 5 /* Perhaps this should depend on maxrange. */ /* * Add a penalty for the number `value' to the list for operation `op', * operand number `operand' (0 or 1). If we run out of memory, we just * forget about the penalty (how likely is this, anyway?). */ void penalise(int value, int op, int operand) { struct penalty *p; op = opnum(op); if ((p = malloc(sizeof(*p))) == NULL) return; p->next = penlist[op][operand]; penlist[op][operand] = p; penalty[op][operand] += p->penalty = WRONGPENALTY; p->value = value; } /* * Select a random value from 0 to maxval - 1 for operand `operand' (0 or 1) * of operation `op'. The random number we generate is either used directly * as a value, or represents a position in the penalty list. If the latter, * we find the corresponding value and return that, decreasing its penalty. */ int getrandom(int maxval, int op, int operand) { int value; struct penalty **pp, *p; op = opnum(op); value = arc4random_uniform(maxval + penalty[op][operand]); /* * 0 to maxval - 1 is a number to be used directly; bigger values * are positions to be located in the penalty list. */ if (value < maxval) return(value); value -= maxval; /* * Find the penalty at position `value'; decrement its penalty and * delete it if it reaches 0; return the corresponding value. */ for (pp = &penlist[op][operand]; (p = *pp) != NULL; pp = &p->next) { if (p->penalty > value) { value = p->value; penalty[op][operand]--; if (--(p->penalty) <= 0) { p = p->next; (void)free((char *)*pp); *pp = p; } return(value); } value -= p->penalty; } /* * We can only get here if the value from the penalty[] array doesn't * correspond to the actual sum of penalties in the list. Provide an * obscure message. */ errx(1, "bug: inconsistent penalties."); /* NOTREACHED */ } /* Return an index for the character op, which is one of [+-x/]. */ int opnum(int op) { char *p; if (op == 0 || (p = strchr(keylist, op)) == NULL) errx(1, "bug: op %c not in keylist %s.", op, keylist); return(p - keylist); } /* Print usage message and quit. */ void usage(void) { (void)fprintf(stderr, "usage: arithmetic [-o +-x/] [-r range]\n"); exit(1); }