cq.c File Reference

Go to the source code of this file.

Data Structures
struct	defs
Functions
	array (a, size, start)
	clobber (int x, int *y)
int *	fip (int x)
	glork (int x)
	main (int n, char **args)
	McCarthy (int x)
	one ()
long	pow2 (long n)
	regc ()
	regi ()
	regp ()
	s22 (struct defs *pd0)
	s241 (struct defs *pd0)
	s243 (struct defs *pd0)
	s244 (struct defs *pd0)
	s25 (struct defs *pd0)
	s26 (struct defs *pd0)
	s4 (struct defs *pd0)
	s61 (struct defs *pd0)
	s626 (struct defs *pd0)
	s71 (struct defs *pd0)
	s714 (struct defs *pd0)
	s715 (struct defs *pd0)
	s715f (int x, int y, int z)
	s72 (struct defs *pd0)
	s757 (struct defs *pd0)
	s7813 (struct defs *pd0)
	s81 (struct defs *pd0)
	s84 (struct defs *pd0)
	s85 (struct defs *pd0)
	s86 (struct defs *pd0)
	s88 (struct defs *pd0)
	s9 (struct defs *pd0)
	setev ()
	sumof (char *x)
	svtest (int n)
	testev ()
	zero ()
	zerofill (char *x)
Variables
int	crc
int	dbits
float	dprec
int	extvar
int	fbits
int	flgd
int	flgl
int	flgm
int	flgs
float	fprec
int	lbits
int *	metricp
char	rfs [8]
int	rrc
int	ubits

---

Function Documentation

array (  a  ,   size,   start )

Definition at line 4878 of file cq.c. References a, and i. Referenced by array(), bbcall(), bbentry(), bbvars(), compose(), dcllocal(), dclr(), doglobal(), gettok(), initializer(), mkstr(), prof_init(), s84(), swcode(), and uid2type(). { int i; for(i=0; i<size; i++) if(a[i] != i+start) return 1; return 0; }

clobber (  int  x, int *  y )

Definition at line 1164 of file cq.c. References x, and y. Referenced by s71(). { x = 3; *y = 2; }

int* fip (  int  x  )

Definition at line 4887 of file cq.c. References y. Referenced by s84(). { static int y; y = x; return &y; }

glork (  int  x  )

Definition at line 4894 of file cq.c. Referenced by s84(). {return x;}

main (  int  n, char **  args )

Definition at line 19 of file cq.c. References defs::crc, defs::flgd, defs::flgl, defs::flgm, defs::flgs, int(), j, printf(), defs::rfs, defs::rrc, s22(), s241(), s243(), s244(), s25(), s26(), s4(), s61(), s626(), s71(), s714(), s715(), s72(), s757(), s7813(), s81(), s84(), s85(), s86(), s88(), and s9(). { /* This program performs a series of tests on a C compiler, based on information in the C REFERENCE MANUAL which appears as Appendix A to the book "The C Programming Language" by Brian W. Kernighan and Dennis M. Ritchie (Prentice-Hall, 1978, $10.95). This Appendix is hereafter referred to as "the Manual". The rules followed in writing this program are: 1. The entire program is written in legal C, according to the Manual. It should compile with no error messages, although some warning messages may be produced by some compilers. Failure to compile should be interpreted as a compiler error. 2. The program is clean, in that it does not make use of any features of the operating system on which it runs, with the sole exceptions of the printf() function, and an internal "options" routine, which is easily excised. 3. No global variables are used, except for the spec- ific purpose of testing the global variable facility. The program is divided into modules having names of the form snnn... These modules correspond to those sections of the Manual, as identified by boldface type headings, in which there is something to test. For example, s241() corresponds to section 2.4.1 of the Manual (Integer constants) and tests the facilities described therein. The module numbering scheme is ambiguous, especially when it names modules referring to more than one section; module s7813, for ex- ample, deals with sections 7.8 through 7.13. Nonetheless, it is surprisingly easy to find a section in the Manual corresponding to a section of code, and vice versa. Note also that there seem to be "holes" in the program, at least from the point of view that there exist sections in the Manual for which there is no corresponding code. Such holes arise from three causes: (a) there is nothing in that partic- ular section to test, (b) everything in that section is tested elsewhere, and (c) it was deemed advisable not to check cer- tain features like preprocessor or listing control features. Modules are called by a main program main(). The mod- ules that are called, and the sequence in which they are called, are determined by two lists in main(), in which the module names appear. The first list (an extern statement) declares the module names to be external. The second (a stat- ic int statement) names the modules and defines the sequence in which they are called. There is no need for these lists to be in the same order, but it is probably a good idea to keep them that way in the interest of clarity. Since there are no cross-linkages between modules, new modules may be added, or old ones deleted, simply by editing the lists, with one exception: section s26, which pokes around at the hardware trying to figure out the characteristics of the machine that it is running on, saves information that is subsequently used by sections s626, s72, and s757. If this program is to be broken up into smallish pieces, say for running on a microcomputer, take care to see that s26 is called before calling any of the latter three sections. The size of the lists, i.e., the number of modules to be called, is not explicitly specified as a program parameter, but is determined dynamically using the sizeof operator. Communication between the main program and the modules takes place in two ways. In all cases, a pointer to a structure is passed to the called module. The structure contains flags that will determine the type of information to be published by the module, and fields that may be written in by the module. The former include "flgm" and "flgd", which, if set to a nonzero value, specify that machine dependencies are to be announced or that error messages are to be printed, re- spectively. The called module's name, and the hardware char- acteristics probed in s26() comprise the latter. Also, in all cases, a return code is returned by the called module. A return code of zero indicates that all has gone well; nonzero indicates otherwise. Since more than one type of error may be detected by a module, the return code is a composite of error indicators, which, individually, are given as numbers that are powers of two. Thus, a return code of 10 indicates that two specific errors, 8 and 2, were detected. Whether or not the codes returned by the modules are printed by the main program is determined by setting "flgs" to 1 (resp. 0). The entire logic of the main program is contained in the half-dozen or so lines at the end. The somewhat cryptic statement: d0.rrc = (*sec[j])(pd0); in the for loop calls the modules. The rest of the code is reasonably straightforward. Finally, in each of the modules, there is the following prologue: snnn(pd0) struct defs *pd0; { static char snnner[] = "snnn,er%d\n"; static char qsnnn[8] = "snnn "; char *ps, *pt; int rc; rc = 0; ps = qsnnn; pt = pd0->rfs; while(*pt++ = *ps++); used for housekeeping, handshaking and module initialization. */ extern s22(), s241(), s243(), s244(), s25(), s26(), s4(), s61(), s626(), s71(), s72(), s757(), s7813(), s714(), s715(), s81(), s84(), s85(), s86(), s88(), s9() ; int j; static int (*sec[])() = { s22, s241, s243, s244, s25, s26, s4, s61, s626, s71, s72, s757, s7813, s714, s715, s81, s84, s85, s86, s88, s9 }; static struct defs d0, *pd0; d0.flgs = 1; /* These flags dictate */ d0.flgm = 1; /* the verbosity of */ d0.flgd = 1; /* the program. */ d0.flgl = 1; pd0 = &d0; for (j=0; j<sizeof(sec) / sizeof(sec[0]); j++) { d0.rrc = (*sec[j])(pd0); d0.crc = d0.crc+d0.rrc; if(d0.flgs != 0) printf("Section %s returned %d.\n",d0.rfs,d0.rrc); } if(d0.crc == 0) printf("\nNo errors detected.\n"); else printf("\nFailed.\n"); return 0; }

Here is the call graph for this function:

McCarthy (  int  x  )

Definition at line 1158 of file cq.c. References x. Referenced by s71(). { if(x>100) return x-10; else return McCarthy( McCarthy(x+11)); }

one (   )

Definition at line 5176 of file cq.c. Referenced by s26(), s7813(), s86(), and VectorArrayNormalize(). { return 1; }

long pow2 (  long  n  )

Definition at line 420 of file cq.c. References n, and s. Referenced by s241(). { long s; s = 1; while(n--) s = s*2; return s; }

regc (   )

Definition at line 4266 of file cq.c. References d, j, and s. Referenced by s81(). { /* char to register assignment */ /* Testing a variable whose storage class has been spec- ified as "register" is somewhat tricky, but it can be done in a fairly reliable fashion by taking advantage of our knowledge of the ways in which compilers operate. If we declare a collection of vari- ables of the same storage class, we would expect that, when storage for these variables is actually allocated, the variables will be bunched together and ordered according to one of the following criteria: (a) the order in which they were defined. (b) the order in which they are used. (c) alphabetically. (d) the order in which they appear in the compiler's symbol table. (e) some other way. Hence, if we define a sequence of variables in close alpha- betical order, and use them in the same order in which we define them, we would expect the differences between the addresses of successive variables to be constant, except in case (d) where the symbol table is a hash table, or in case (e). If a subsequence in the middle of this sequence is selected, and for this subsequence, every other variable is specified to be "register", and address differences are taken between adjacent nonregister variables, we would still expect to find constant differences if the "register" vari- ables were actually assigned to registers, and some other diff- erences if they were not. Specifically, if we had N variables specified as "register" of which the first n were actually ass- igned to registers, we would expect the sequence of differences to consist of a number of occurrences of some number, followed by N-n occurrences of some other number, followed by several occurr- ences of the first number. If we get a sequence like this, we can determine, by simple subtraction, how many (if any) variables are being assigned to registers. If we get some other sequence, we know that the test is invalid. */ char r00; char r01; char r02; char r03; register char r04; char r05; register char r06; char r07; register char r08; char r09; register char r10; char r11; register char r12; char r13; register char r14; char r15; register char r16; char r17; register char r18; char r19; register char r20; char r21; register char r22; char r23; register char r24; char r25; register char r26; char r27; register char r28; char r29; register char r30; char r31; register char r32; char r33; register char r34; char r35; char r36; char r37; char r38; int s, n1, n2, nr, j, d[22]; r00 = 0; r01 = 1; r02 = 2; r03 = 3; r04 = 4; r05 = 5; r06 = 6; r07 = 7; r08 = 8; r09 = 9; r10 = 10; r11 = 11; r12 = 12; r13 = 13; r14 = 14; r15 = 15; r16 = 16; r17 = 17; r18 = 18; r19 = 19; r20 = 20; r21 = 21; r22 = 22; r23 = 23; r24 = 24; r25 = 25; r26 = 26; r27 = 27; r28 = 28; r29 = 29; r30 = 30; r31 = 31; r32 = 32; r33 = 33; r34 = 34; r35 = 35; r36 = 36; r37 = 37; r38 = 38; d[0] = &r01 - &r00; d[1] = &r02 - &r01; d[2] = &r03 - &r02; d[3] = &r05 - &r03; d[4] = &r07 - &r05; d[5] = &r09 - &r07; d[6] = &r11 - &r09; d[7] = &r13 - &r11; d[8] = &r15 - &r13; d[9] = &r17 - &r15; d[10] = &r19 - &r17; d[11] = &r21 - &r19; d[12] = &r23 - &r21; d[13] = &r25 - &r23; d[14] = &r27 - &r25; d[15] = &r29 - &r27; d[16] = &r31 - &r29; d[17] = &r33 - &r31; d[18] = &r35 - &r33; d[19] = &r36 - &r35; d[20] = &r37 - &r36; d[21] = &r38 - &r37; /* The following FSM analyzes the string of differences. It accepts strings of the form a+b+a+ and returns 16 minus the number of bs, which is the number of variables that actually got into registers. Otherwise it signals rejection by returning -1., indicating that the test is unreliable. */ n1 = d[0]; s = 1; for (j=0; j<22; j++) switch (s) { case 1: if (d[j] != n1) { n2 = d[j]; s = 2; nr = 1; } break; case 2: if (d[j] == n1) { s = 3; break; } if (d[j] == n2) { nr = nr+1; break; } s = 4; break; case 3: if (d[j] != n1) s = 4; break; } ; if (s == 3) return 16-nr; else return -1; }

regi (   )

Definition at line 4443 of file cq.c. References d, j, and s. Referenced by s81(). { /* int to register assignment */ /* Testing a variable whose storage class has been spec- ified as "register" is somewhat tricky, but it can be done in a fairly reliable fashion by taking advantage of our knowledge of the ways in which compilers operate. If we declare a collection of vari- ables of the same storage class, we would expect that, when storage for these variables is actually allocated, the variables will be bunched together and ordered according to one of the following criteria: (a) the order in which they were defined. (b) the order in which they are used. (c) alphabetically. (d) the order in which they appear in the compiler's symbol table. (e) some other way. Hence, if we define a sequence of variables in close alpha- betical order, and use them in the same order in which we define them, we would expect the differences between the addresses of successive variables to be constant, except in case (d) where the symbol table is a hash table, or in case (e). If a subsequence in the middle of this sequence is selected, and for this subsequence, every other variable is specified to be "register", and address differences are taken between adjacent nonregister variables, we would still expect to find constant differences if the "register" vari- ables were actually assigned to registers, and some other diff- erences if they were not. Specifically, if we had N variables specified as "register" of which the first n were actually ass- igned to registers, we would expect the sequence of differences to consist of a number of occurrences of some number, followed by N-n occurrences of some other number, followed by several occurr- ences of the first number. If we get a sequence like this, we can determine, by simple subtraction, how many (if any) variables are being assigned to registers. If we get some other sequence, we know that the test is invalid. */ int r00; int r01; int r02; int r03; register int r04; int r05; register int r06; int r07; register int r08; int r09; register int r10; int r11; register int r12; int r13; register int r14; int r15; register int r16; int r17; register int r18; int r19; register int r20; int r21; register int r22; int r23; register int r24; int r25; register int r26; int r27; register int r28; int r29; register int r30; int r31; register int r32; int r33; register int r34; int r35; int r36; int r37; int r38; int s, n1, n2, nr, j, d[22]; r00 = 0; r01 = 1; r02 = 2; r03 = 3; r04 = 4; r05 = 5; r06 = 6; r07 = 7; r08 = 8; r09 = 9; r10 = 10; r11 = 11; r12 = 12; r13 = 13; r14 = 14; r15 = 15; r16 = 16; r17 = 17; r18 = 18; r19 = 19; r20 = 20; r21 = 21; r22 = 22; r23 = 23; r24 = 24; r25 = 25; r26 = 26; r27 = 27; r28 = 28; r29 = 29; r30 = 30; r31 = 31; r32 = 32; r33 = 33; r34 = 34; r35 = 35; r36 = 36; r37 = 37; r38 = 38; d[0] = &r01 - &r00; d[1] = &r02 - &r01; d[2] = &r03 - &r02; d[3] = &r05 - &r03; d[4] = &r07 - &r05; d[5] = &r09 - &r07; d[6] = &r11 - &r09; d[7] = &r13 - &r11; d[8] = &r15 - &r13; d[9] = &r17 - &r15; d[10] = &r19 - &r17; d[11] = &r21 - &r19; d[12] = &r23 - &r21; d[13] = &r25 - &r23; d[14] = &r27 - &r25; d[15] = &r29 - &r27; d[16] = &r31 - &r29; d[17] = &r33 - &r31; d[18] = &r35 - &r33; d[19] = &r36 - &r35; d[20] = &r37 - &r36; d[21] = &r38 - &r37; /* The following FSM analyzes the string of differences. It accepts strings of the form a+b+a+ and returns 16 minus the number of bs, which is the number of variables that actually got into registers. Otherwise it signals rejection by returning -1., indicating that the test is unreliable. */ n1 = d[0]; s = 1; for (j=0; j<22; j++) switch (s) { case 1: if (d[j] != n1) { n2 = d[j]; s = 2; nr = 1; } break; case 2: if (d[j] == n1) { s = 3; break; } if (d[j] == n2) { nr = nr+1; break; } s = 4; break; case 3: if (d[j] != n1) s = 4; break; } ; if (s == 3) return 16-nr; else return -1; }

regp (   )

Definition at line 4622 of file cq.c. References d, j, and s. Referenced by s81(). { /* pointer to register assignment */ /* Testing a variable whose storage class has been spec- ified as "register" is somewhat tricky, but it can be done in a fairly reliable fashion by taking advantage of our knowledge of the ways in which compilers operate. If we declare a collection of vari- ables of the same storage class, we would expect that, when storage for these variables is actually allocated, the variables will be bunched together and ordered according to one of the following criteria: (a) the order in which they were defined. (b) the order in which they are used. (c) alphabetically. (d) the order in which they appear in the compiler's symbol table. (e) some other way. Hence, if we define a sequence of variables in close alpha- betical order, and use them in the same order in which we define them, we would expect the differences between the addresses of successive variables to be constant, except in case (d) where the symbol table is a hash table, or in case (e). If a subsequence in the middle of this sequence is selected, and for this subsequence, every other variable is specified to be "register", and address differences are taken between adjacent nonregister variables, we would still expect to find constant differences if the "register" vari- ables were actually assigned to registers, and some other diff- erences if they were not. Specifically, if we had N variables specified as "register" of which the first n were actually ass- igned to registers, we would expect the sequence of differences to consist of a number of occurrences of some number, followed by N-n occurrences of some other number, followed by several occurr- ences of the first number. If we get a sequence like this, we can determine, by simple subtraction, how many (if any) variables are being assigned to registers. If we get some other sequence, we know that the test is invalid. */ int *r00; int *r01; int *r02; int *r03; register int *r04; int *r05; register int *r06; int *r07; register int *r08; int *r09; register int *r10; int *r11; register int *r12; int *r13; register int *r14; int *r15; register int *r16; int *r17; register int *r18; int *r19; register int *r20; int *r21; register int *r22; int *r23; register int *r24; int *r25; register int *r26; int *r27; register int *r28; int *r29; register int *r30; int *r31; register int *r32; int *r33; register int *r34; int *r35; int *r36; int *r37; int *r38; int s, n1, n2, nr, j, d[22]; r00 = (int *)&r00; r01 = (int *)&r01; r02 = (int *)&r02; r03 = (int *)&r03; r04 = (int *)&r05; r05 = (int *)&r05; r06 = (int *)&r07; r07 = (int *)&r07; r08 = (int *)&r09; r09 = (int *)&r09; r10 = (int *)&r11; r11 = (int *)&r11; r12 = (int *)&r13; r13 = (int *)&r13; r14 = (int *)&r15; r15 = (int *)&r15; r16 = (int *)&r17; r17 = (int *)&r17; r18 = (int *)&r19; r19 = (int *)&r19; r20 = (int *)&r21; r21 = (int *)&r21; r22 = (int *)&r23; r23 = (int *)&r23; r24 = (int *)&r25; r25 = (int *)&r25; r26 = (int *)&r27; r27 = (int *)&r27; r28 = (int *)&r29; r29 = (int *)&r29; r30 = (int *)&r31; r31 = (int *)&r31; r32 = (int *)&r33; r33 = (int *)&r33; r34 = (int *)&r35; r35 = (int *)&r35; r36 = (int *)&r36; r37 = (int *)&r37; r38 = (int *)&r38; d[0] = &r01 - &r00; d[1] = &r02 - &r01; d[2] = &r03 - &r02; d[3] = &r05 - &r03; d[4] = &r07 - &r05; d[5] = &r09 - &r07; d[6] = &r11 - &r09; d[7] = &r13 - &r11; d[8] = &r15 - &r13; d[9] = &r17 - &r15; d[10] = &r19 - &r17; d[11] = &r21 - &r19; d[12] = &r23 - &r21; d[13] = &r25 - &r23; d[14] = &r27 - &r25; d[15] = &r29 - &r27; d[16] = &r31 - &r29; d[17] = &r33 - &r31; d[18] = &r35 - &r33; d[19] = &r36 - &r35; d[20] = &r37 - &r36; d[21] = &r38 - &r37; /* The following FSM analyzes the string of differences. It accepts strings of the form a+b+a+ and returns 16 minus the number of bs, which is the number of variables that actually got into registers. Otherwise it signals rejection by returning -1., indicating that the test is unreliable. */ n1 = d[0]; s = 1; for (j=0; j<22; j++) switch (s) { case 1: if (d[j] != n1) { n2 = d[j]; s = 2; nr = 1; } break; case 2: if (d[j] == n1) { s = 3; break; } if (d[j] == n2) { nr = nr+1; break; } s = 4; break; case 3: if (d[j] != n1) s = 4; break; } ; if (s == 3) return 16-nr; else return -1; }

s22 (  struct defs *  pd0  )

Definition at line 210 of file cq.c. References A, a, and printf(). Referenced by main(). { int a234, a; int _, _234, A, rc; static char s22er[] = "s22,er%d\n"; static char qs22[8] = "s22 "; char *ps, *pt; /* Initialize */ rc = 0; ps = qs22; pt = pd0 -> rfs; while (*pt++ = *ps++); /* An identifier is a sequence of letters and digits; the first character must be a letter. The under- score _ counts as a letter. */ a=1; _=2; _234=3; a234=4; if(a+_+_234+a234 != 10) { rc = rc+1; if(pd0->flgd != 0) printf(s22er,1); } /* Upper and lower case letters are different. */ A = 2; if (A == a) { rc = rc+4; if (pd0->flgd != 0) printf(s22er,4); } return(rc); }

Here is the call graph for this function:

s241 (  struct defs *  pd0  )

Definition at line 250 of file cq.c. References d, flgd, flgm, g(), j, L, l, pow2(), printf(), and x. Referenced by main(). { long pow2(); static char s241er[] = "s241,er%d\n"; static char qs241[8] = "s241 "; char *ps, *pt; int rc, j, lrc; static long g[39] = {0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,6,0,8,0,12,0,16,0,18,0,20,0,24, 0,28,0,30,0,32,0,36}; long d[39], o[39], x[39]; rc = 0; lrc = 0; ps = qs241; pt = pd0 -> rfs; while (*pt++ = *ps++); /* An integer constant consisting of a sequence of digits is taken to be octal if it begins with 0 (digit zero), decimal otherwise. */ if ( 8 != 010 || 16 != 020 || 24 != 030 || 32 != 040 || 40 != 050 || 48 != 060 || 56 != 070 || 64 != 0100 || 72 != 0110 || 80 != 0120 || 9 != 0011 || 17 != 0021 || 25 != 0031 || 33 != 0041 || 41 != 0051 || 49 != 0061 || 57 != 0071 || 65 != 0101 || 73 != 0111 || 81 != 0121 ){ rc = rc+1; if( pd0->flgd != 0 ) printf(s241er,1); } /* A sequence of digits preceded by 0x or 0X (digit zero) is taken to be a hexadecimal integer. The hexadecimal digits include a or A through f or F with values 10 through 15. */ if ( 0x00abcdef != 0xabcdef || 0xabcdef != 0Xabcdef || 0Xabcdef != 0XAbcdef || 0XAbcdef != 0XABcdef || 0XABcdef != 0XABCdef || 0XABCdef != 0XABCDef || 0XABCDef != 0XABCDEf || 0XABCDEf != 0XABCDEF || 0xABCDEF != 11259375 ){ rc = rc+2; if( pd0->flgd != 0 ) printf(s241er,2); } /* A decimal constant whose value exceeds the largest signed machine integer is taken to be long; an octal or hex con- stant which exceeds the largest unsigned machine integer is likewise taken to be long. */ if ( sizeof 010000000000 != sizeof(long) /* 2**30 */ || sizeof 1073741824 != sizeof(long) /* ditto */ || sizeof 0x40000000 != sizeof(long) ){ /* " */ rc = rc+4; if( pd0->flgd != 0 ) printf(s241er,4); } /* A decimal, octal, or hexadecimal constant immediately followed by l (letter ell) or L is a long constant. */ if ( sizeof 67l != sizeof(long) || sizeof 67L != sizeof(long) || sizeof 067l != sizeof(long) || sizeof 067L != sizeof(long) || sizeof 0X67l != sizeof(long) || sizeof 0x67L != sizeof(long) ){ rc = rc+8; if( pd0 -> flgd != 0 ) printf(s241er,8); } /* Finally, we test to see that decimal (d), octal (o), and hexadecimal (x) constants representing the same values agree among themselves, and with computed values, at spec- ified points over an appropriate range. The points select- ed here are those with the greatest potential for caus- ing trouble, i.e., zero, 1-16, and values of 2**n and 2**n - 1 where n is some multiple of 4 or 6. Unfortunately, just what happens when a value is too big to fit in a long is undefined; however, it would be nice if what happened were at least consistent... */ for ( j=0; j<17; j++ ) g[j] = j; for ( j=18; j<39; ) { g[j] = pow2(g[j]); g[j-1] = g[j] - 1; j = j+2; } d[0] = 0; o[0] = 00; x[0] = 0x0; d[1] = 1; o[1] = 01; x[1] = 0x1; d[2] = 2; o[2] = 02; x[2] = 0x2; d[3] = 3; o[3] = 03; x[3] = 0x3; d[4] = 4; o[4] = 04; x[4] = 0x4; d[5] = 5; o[5] = 05; x[5] = 0x5; d[6] = 6; o[6] = 06; x[6] = 0x6; d[7] = 7; o[7] = 07; x[7] = 0x7; d[8] = 8; o[8] = 010; x[8] = 0x8; d[9] = 9; o[9] = 011; x[9] = 0x9; d[10] = 10; o[10] = 012; x[10] = 0xa; d[11] = 11; o[11] = 013; x[11] = 0xb; d[12] = 12; o[12] = 014; x[12] = 0xc; d[13] = 13; o[13] = 015; x[13] = 0xd; d[14] = 14; o[14] = 016; x[14] = 0xe; d[15] = 15; o[15] = 017; x[15] = 0xf; d[16] = 16; o[16] = 020; x[16] = 0x10; d[17] = 63; o[17] = 077; x[17] = 0x3f; d[18] = 64; o[18] = 0100; x[18] = 0x40; d[19] = 255; o[19] = 0377; x[19] = 0xff; d[20] = 256; o[20] = 0400; x[20] = 0x100; d[21] = 4095; o[21] = 07777; x[21] = 0xfff; d[22] = 4096; o[22] = 010000; x[22] = 0x1000; d[23] = 65535; o[23] = 0177777; x[23] = 0xffff; d[24] = 65536; o[24] = 0200000; x[24] = 0x10000; d[25] = 262143; o[25] = 0777777; x[25] = 0x3ffff; d[26] = 262144; o[26] = 01000000; x[26] = 0x40000; d[27] = 1048575; o[27] = 03777777; x[27] = 0xfffff; d[28] = 1048576; o[28] = 04000000; x[28] = 0x100000; d[29] = 16777215; o[29] = 077777777; x[29] = 0xffffff; d[30] = 16777216; o[30] = 0100000000; x[30] = 0x1000000; d[31] = 268435455; o[31] = 01777777777; x[31] = 0xfffffff; d[32] = 268435456; o[32] = 02000000000; x[32] = 0x10000000; d[33] = 1073741823; o[33] = 07777777777; x[33] = 0x3fffffff; d[34] = 1073741824; o[34] = 010000000000; x[34] = 0x40000000; d[35] = 4294967295; o[35] = 037777777777; x[35] = 0xffffffff; d[36] = 4294967296; o[36] = 040000000000; x[36] = 0x100000000; d[37] = 68719476735; o[37] = 0777777777777; x[37] = 0xfffffffff; d[38] = 68719476736; o[38] = 01000000000000; x[38] = 0x1000000000; /* WHEW! */ for (j=0; j<39; j++){ if ( g[j] != d[j] || d[j] != o[j] || o[j] != x[j]) { if( pd0 -> flgm != 0 ) { /* printf(s241er,16); save in case opinions change... */ printf("Decimal and octal/hex constants sometimes give\n"); printf(" different results when assigned to longs.\n"); } /* lrc = 1; save... */ } } if (lrc != 0) rc =16; return rc; }

Here is the call graph for this function:

s243 (  struct defs *  pd0  )

Definition at line 428 of file cq.c. References printf(), defs::rfs, sumof(), and zerofill(). Referenced by main(). { static char s243er[] = "s243,er%d\n"; static char qs243[8] = "s243 "; char *ps, *pt; int rc; char chars[256]; rc = 0; ps = qs243; pt = pd0->rfs; while(*pt++ = *ps++); /* One of the problems that arises when testing character constants is that of definition: What, exactly, is the character set? In order to guarantee a certain amount of machine independence, the character set we will use here is the set of characters writ- able as escape sequences in C, plus those characters used in writ- ing C programs, i.e., letters: ABCDEFGHIJKLMNOPQRSTUVWXYZ 26 abcdefghijklmnopqrstuvwxyz 26 numbers: 0123456789 10 special characters: ~!"#%&()_=-^|{}[]+;*:<>,.?/ 27 extra special characters: newline \n horizontal tab \t backspace \b carriage return \r form feed \f backslash \\ single quote \' 7 blank & NUL 2 --- 98 Any specific implementation of C may of course support additional characters. */ /* Since the value of a character constant is the numerical value of the character in the machine's character set, there should be a one-to-one correspondence between characters and values. */ zerofill(chars); chars['a'] = 1; chars['A'] = 1; chars['~'] = 1; chars['0'] = 1; chars['b'] = 1; chars['B'] = 1; chars['!'] = 1; chars['1'] = 1; chars['c'] = 1; chars['C'] = 1; chars['"'] = 1; chars['2'] = 1; chars['d'] = 1; chars['D'] = 1; chars['#'] = 1; chars['3'] = 1; chars['e'] = 1; chars['E'] = 1; chars['%'] = 1; chars['4'] = 1; chars['f'] = 1; chars['F'] = 1; chars['&'] = 1; chars['5'] = 1; chars['g'] = 1; chars['G'] = 1; chars['('] = 1; chars['6'] = 1; chars['h'] = 1; chars['H'] = 1; chars[')'] = 1; chars['7'] = 1; chars['i'] = 1; chars['I'] = 1; chars['_'] = 1; chars['8'] = 1; chars['j'] = 1; chars['J'] = 1; chars['='] = 1; chars['9'] = 1; chars['k'] = 1; chars['K'] = 1; chars['-'] = 1; chars['l'] = 1; chars['L'] = 1; chars['^'] = 1; chars['m'] = 1; chars['M'] = 1; chars['|'] = 1; chars['\n'] = 1; chars['n'] = 1; chars['N'] = 1; chars['\t'] = 1; chars['o'] = 1; chars['O'] = 1; chars['{'] = 1; chars['\b'] = 1; chars['p'] = 1; chars['P'] = 1; chars['}'] = 1; chars['\r'] = 1; chars['q'] = 1; chars['Q'] = 1; chars['['] = 1; chars['\f'] = 1; chars['r'] = 1; chars['R'] = 1; chars[']'] = 1; chars['s'] = 1; chars['S'] = 1; chars['+'] = 1; chars['\\'] = 1; chars['t'] = 1; chars['T'] = 1; chars[';'] = 1; chars['\''] = 1; chars['u'] = 1; chars['U'] = 1; chars['*'] = 1; chars['v'] = 1; chars['V'] = 1; chars[':'] = 1; chars['\0'] = 1; chars['w'] = 1; chars['W'] = 1; chars['<'] = 1; chars[' '] = 1; chars['x'] = 1; chars['X'] = 1; chars['>'] = 1; chars['y'] = 1; chars['Y'] = 1; chars[','] = 1; chars['z'] = 1; chars['Z'] = 1; chars['.'] = 1; chars['?'] = 1; chars['/'] = 1; if(sumof(chars) != 98){ rc = rc+1; if(pd0->flgd != 0) printf(s243er,1); } /* Finally, the escape \ddd consists of the backslash followed by 1, 2, or 3 octal digits which are taken to specify the desired character. */ if( '\0' != 0 || '\01' != 1 || '\02' != 2 || '\03' != 3 || '\04' != 4 || '\05' != 5 || '\06' != 6 || '\07' != 7 || '\10' != 8 || '\17' != 15 || '\20' != 16 || '\77' != 63 || '\100' != 64 || '\177' != 127 ){ rc = rc+8; if(pd0->flgd != 0) printf(s243er,8); } return rc; }

Here is the call graph for this function:

s244 (  struct defs *  pd0  )

Definition at line 546 of file cq.c. References a, e, j, printf(), and defs::rfs. Referenced by main(). { double a[8]; int rc, lrc, j; static char s244er[] = "s244,er%d\n"; static char qs244[8] = "s244 "; char *ps, *pt; ps = qs244; pt = pd0->rfs; while(*pt++ = *ps++); rc = 0; lrc = 0; /* Unfortunately, there's not a lot we can do with floating constants. We can check to see that the various representations can be com- piled, that the conversion is such that they yield the same hard- ware representations in all cases, and that all representations thus checked are double precision. */ a[0] = .1250E+04; a[1] = 1.250E3; a[2] = 12.50E02; a[3] = 125.0e+1; a[4] = 1250e00; a[5] = 12500.e-01; a[6] = 125000e-2; a[7] = 1250.; lrc = 0; for (j=0; j<7; j++) if(a[j] != a[j+1]) lrc = 1; if(lrc != 0) { if(pd0->flgd != 0) printf(s244er,1); rc = rc+1; } if ( (sizeof .1250E+04 ) != sizeof(double) || (sizeof 1.250E3 ) != sizeof(double) || (sizeof 12.50E02 ) != sizeof(double) || (sizeof 1.250e+1 ) != sizeof(double) || (sizeof 1250e00 ) != sizeof(double) || (sizeof 12500.e-01) != sizeof(double) || (sizeof 125000e-2 ) != sizeof(double) || (sizeof 1250. ) != sizeof(double)){ if(pd0->flgd != 0) printf(s244er,2); rc = rc+2; } return rc; }

Here is the call graph for this function:

s25 (  struct defs *  pd0  )

Definition at line 599 of file cq.c. References j, printf(), defs::rfs, and s. Referenced by main(). { char *s, *s2; int rc, lrc, j; static char s25er[] = "s25,er%d\n"; static char qs25[8] = "s25 "; char *ps, *pt; ps = qs25; pt = pd0->rfs; while(*pt++ = *ps++); rc = 0; /* A string is a sequence of characters surrounded by double quotes, as in "...". */ s = "..."; /* A string has type "array of characters" and storage class static and is initialized with the given characters. */ if ( s[0] != s[1] || s[1] != s[2] || s[2] != '.' ) { rc = rc+1; if(pd0->flgd != 0) printf(s25er,1); } /* The compiler places a null byte \0 at the end of each string so the program which scans the string can find its end. */ if( s[3] != '\0' ){ rc = rc+4; if(pd0->flgd != 0) printf(s25er,4); } /* In a string, the double quote character " must be preceded by a \. */ if( ".\"."[1] != '"' ){ rc = rc+8; if(pd0->flgd != 0) printf(s25er,8); } /* In addition, the same escapes described for character constants may be used. */ s = "\n\t\b\r\f\\\'"; if( s[0] != '\n' || s[1] != '\t' || s[2] != '\b' || s[3] != '\r' || s[4] != '\f' || s[5] != '\\' || s[6] != '\'' ){ rc = rc+16; if( pd0->flgd != 0) printf(s25er,16); } /* Finally, a \ and an immediately following newline are ignored */ s2 = "queep!"; s = "queep!"; lrc = 0; for (j=0; j<sizeof "queep!"; j++) if(s[j] != s2[j]) lrc = 1; if (lrc != 0){ rc = rc+32; if(pd0->flgd != 0) printf(s25er,32); } return rc; }

Here is the call graph for this function:

s26 (  struct defs *  pd0  )

Definition at line 674 of file cq.c. References defs::cbits, defs::dbits, defs::dprec, defs::fbits, defs::fprec, defs::ibits, defs::lbits, one(), printf(), defs::rfs, s, defs::sbits, and defs::ubits. Referenced by main(). { static char qs26[8] = "s26 "; char *ps, *pt; char c0, c1; float temp, one, delta; double tempd, oned; static char s[] = "%3d bits in %ss.\n"; static char s2[] = "%e is the least number that can be added to 1. (%s).\n"; ps = qs26; pt = pd0->rfs; while(*pt++ = *ps++); /* Here, we shake the machinery a little to see what falls out. First, we find out how many bits are in a char. */ pd0->cbits = 0; c0 = 0; c1 = 1; while(c0 != c1) { c1 = c1<<1; pd0->cbits = pd0->cbits+1; } /* That information lets us determine the size of everything else. */ pd0->ibits = pd0->cbits * sizeof(int); pd0->sbits = pd0->cbits * sizeof(short); pd0->lbits = pd0->cbits * sizeof(long); pd0->ubits = pd0->cbits * sizeof(unsigned); pd0->fbits = pd0->cbits *