1 /* Copyright (c) 2000 TXT DataKonsult Ab & Monty Program Ab
2    Copyright (c) 2009-2011, Monty Program Ab
3 
4    Redistribution and use in source and binary forms, with or without
5    modification, are permitted provided that the following conditions are
6    met:
7 
8    1. Redistributions of source code must retain the above copyright
9    notice, this list of conditions and the following disclaimer.
10 
11    2. Redistributions in binary form must the following disclaimer in
12      the documentation and/or other materials provided with the
13      distribution.
14 
15    THIS SOFTWARE IS PROVIDED BY <COPYRIGHT HOLDER> ``AS IS'' AND ANY
16    EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
17    IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
18    PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL <COPYRIGHT HOLDER> OR
19    CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
20    SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
21    LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
22    USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
23    ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
24    OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
25    OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26    SUCH DAMAGE.
27 */
28 
29 /*
30   str2int(src, radix, lower, upper, &val)
31   converts the string pointed to by src to an integer and stores it in
32   val.	It skips leading spaces and tabs (but not newlines, formfeeds,
33   backspaces), then it accepts an optional sign and a sequence of digits
34   in the specified radix.  The result should satisfy lower <= *val <= upper.
35   The result is a pointer to the first character after the number;
36   trailing spaces will NOT be skipped.
37 
38   If an error is detected, the result will be NullS, the value put
39   in val will be 0, and errno will be set to
40 	EDOM	if there are no digits
41 	ERANGE	if the result would overflow or otherwise fail to lie
42 		within the specified bounds.
43   Check that the bounds are right for your machine.
44   This looks amazingly complicated for what you probably thought was an
45   easy task.  Coping with integer overflow and the asymmetric range of
46   twos complement machines is anything but easy.
47 
48   So that users of atoi and atol can check whether an error occurred,
49   I have taken a wholly unprecedented step: errno is CLEARED if this
50   call has no problems.
51 */
52 
53 #include "strings_def.h"
54 #include <m_ctype.h>
55 #include "my_sys.h"			/* defines errno */
56 #include <errno.h>
57 
58 #define char_val(X) (X >= '0' && X <= '9' ? X-'0' :\
59 		     X >= 'A' && X <= 'Z' ? X-'A'+10 :\
60 		     X >= 'a' && X <= 'z' ? X-'a'+10 :\
61 		     '\177')
62 
str2int(register const char * src,register int radix,long int lower,long int upper,long int * val)63 char *str2int(register const char *src, register int radix, long int lower,
64 	      long int upper, long int *val)
65 {
66   int sign;			/* is number negative (+1) or positive (-1) */
67   int n;			/* number of digits yet to be converted */
68   long limit;			/* "largest" possible valid input */
69   long scale;			/* the amount to multiply next digit by */
70   long sofar;			/* the running value */
71   register int d;		/* (negative of) next digit */
72   char *start;
73   int digits[32];		/* Room for numbers */
74 
75   /*  Make sure *val is sensible in case of error  */
76 
77   *val = 0;
78 
79   /*  Check that the radix is in the range 2..36  */
80 
81 #ifndef DBUG_OFF
82   if (radix < 2 || radix > 36) {
83     errno=EDOM;
84     return NullS;
85   }
86 #endif
87 
88   /*  The basic problem is: how do we handle the conversion of
89       a number without resorting to machine-specific code to
90       check for overflow?  Obviously, we have to ensure that
91       no calculation can overflow.  We are guaranteed that the
92       "lower" and "upper" arguments are valid machine integers.
93       On sign-and-magnitude, twos-complement, and ones-complement
94       machines all, if +|n| is representable, so is -|n|, but on
95       twos complement machines the converse is not true.  So the
96       "maximum" representable number has a negative representative.
97       Limit is set to MY_MIN(-|lower|,-|upper|); this is the "largest"
98       number we are concerned with.	*/
99 
100   /*  Calculate Limit using Scale as a scratch variable  */
101 
102   if ((limit = lower) > 0) limit = -limit;
103   if ((scale = upper) > 0) scale = -scale;
104   if (scale < limit) limit = scale;
105 
106   /*  Skip leading spaces and check for a sign.
107       Note: because on a 2s complement machine MinLong is a valid
108       integer but |MinLong| is not, we have to keep the current
109       converted value (and the scale!) as *negative* numbers,
110       so the sign is the opposite of what you might expect.
111       */
112   while (my_isspace(&my_charset_latin1,*src)) src++;
113   sign = -1;
114   if (*src == '+') src++; else
115     if (*src == '-') src++, sign = 1;
116 
117   /*  Skip leading zeros so that we never compute a power of radix
118       in scale that we won't have a need for.  Otherwise sticking
119       enough 0s in front of a number could cause the multiplication
120       to overflow when it neededn't.
121       */
122   start=(char*) src;
123   while (*src == '0') src++;
124 
125   /*  Move over the remaining digits.  We have to convert from left
126       to left in order to avoid overflow.  Answer is after last digit.
127       */
128 
129   for (n = 0; (digits[n]=char_val(*src)) < radix && n < 20; n++,src++) ;
130 
131   /*  Check that there is at least one digit  */
132 
133   if (start == src) {
134     errno=EDOM;
135     return NullS;
136   }
137 
138   /*  The invariant we want to maintain is that src is just
139       to the right of n digits, we've converted k digits to
140       sofar, scale = -radix**k, and scale < sofar < 0.	Now
141       if the final number is to be within the original
142       Limit, we must have (to the left)*scale+sofar >= Limit,
143       or (to the left)*scale >= Limit-sofar, i.e. the digits
144       to the left of src must form an integer <= (Limit-sofar)/(scale).
145       In particular, this is true of the next digit.  In our
146       incremental calculation of Limit,
147 
148       IT IS VITAL that (-|N|)/(-|D|) = |N|/|D|
149       */
150 
151   for (sofar = 0, scale = -1; --n >= 1;)
152   {
153     if ((long) -(d=digits[n]) < limit) {
154       errno=ERANGE;
155       return NullS;
156     }
157     limit = (limit+d)/radix, sofar += d*scale; scale *= radix;
158   }
159   if (n == 0)
160   {
161     if ((long) -(d=digits[n]) < limit)		/* get last digit */
162     {
163       errno=ERANGE;
164       return NullS;
165     }
166     sofar+=d*scale;
167   }
168 
169   /*  Now it might still happen that sofar = -32768 or its equivalent,
170       so we can't just multiply by the sign and check that the result
171       is in the range lower..upper.  All of this caution is a right
172       pain in the neck.  If only there were a standard routine which
173       says generate thus and such a signal on integer overflow...
174       But not enough machines can do it *SIGH*.
175       */
176   if (sign < 0)
177   {
178     if (sofar < -LONG_MAX || (sofar= -sofar) > upper)
179     {
180       errno=ERANGE;
181       return NullS;
182     }
183   }
184   else if (sofar < lower)
185   {
186     errno=ERANGE;
187     return NullS;
188   }
189   *val = sofar;
190   errno=0;			/* indicate that all went well */
191   return (char*) src;
192 }
193 
194 	/* These are so slow compared with ordinary, optimized atoi */
195 
196 #ifdef WANT_OUR_ATOI
197 
atoi(const char * src)198 int atoi(const char *src)
199 {
200   long val;
201   str2int(src, 10, (long) INT_MIN, (long) INT_MAX, &val);
202   return (int) val;
203 }
204 
205 
atol(const char * src)206 long atol(const char *src)
207 {
208   long val;
209   str2int(src, 10, LONG_MIN, LONG_MAX, &val);
210   return val;
211 }
212 
213 #endif /* WANT_OUR_ATOI */
214