greim@sbsvax.UUCP (Michael Greim) (12/06/88)
Hello,
Just recently I tried the following program:
---- cut ----
# include <stdio.h>
struct misty {
unsigned int a1:1;
unsigned int a2:15;
};
struct misty mist;
char c1, c2;
int i;
float r;
char s [20];
main ()
{
printf ("sizeof(mist.a1) [1 bit] = %d\n", sizeof(mist.a1));
printf ("sizeof(c1<c2) = %d\n", sizeof(c1<c2));
printf ("sizeof(i+r) = %d\n", sizeof (i+r));
printf ("sizeof(i+s) = %d\n", sizeof (i+s));
}
---- cut ----
It printed
sizeof(mist.a1) [1 bit] = 4
sizeof(c1<c2) = 4
sizeof(i+r) = 8
sizeof(i+s) = 4
This was on a VAX 11/780 running 4.3BSD.
Now comes my question:
What should "sizeof expression" return? And specially : why does it
return 8 for the third expression?
Thanks for any answers,
-mg
--
email : greim@sbsvax.informatik.uni-saarland.dbp.de
(some mailers might not like this. Then use greim@sbsvax.uucp)
or : ...!uunet!unido!sbsvax!greim
# include <disclaimers/std.h>seanf@sco.COM (Sean Fagan) (12/08/88)
In article <654@sbsvax.UUCP> greim@sbsvax.UUCP (Michael Greim) writes: [struct mist { int a1:1, a2:16}; int i; float r; char s[20]; >sizeof(mist.a1) [1 bit] = 4 Because an int, on your machine, is 4 bytes long, and mist.a1 is expanded to an unsigned int. >sizeof(c1<c2) = 4 This is an integer expression, 4 bytes. >sizeof(i+r) = 8 An integer and a floating point form a floating point expression, which, on your machine, is 8 bytes. >sizeof(i+s) = 4 This is a pointer expression (equivalent to &s[i]), and a character pointer on your machine is 4 bytes long. >What should "sizeof expression" return? The size, in char's, of the expression. -- Sean Eric Fagan | "Engineering without management is *ART*" seanf@sco.UUCP | Jeff Johnson (jeffj@sco) (408) 458-1422 | Any opinions expressed are my own, not my employers'.
evil@arcturus.UUCP (Wade Guthrie) (12/09/88)
Well, I am a relative novice, but how do we learn without trying and experiencing the flame of disappointment :-) In article <654@sbsvax.UUCP>, greim@sbsvax.UUCP (Michael Greim) asks about the output of the sizeof operator in regards to the following program: [abridged for the line counter] > struct misty {unsigned int a1:1; unsigned int a2:15;}mist; > int i; > float r; > char s [20]; > > main () > { > printf ("sizeof(mist.a1) [1 bit] = %d\n", sizeof(mist.a1)); > printf ("sizeof(c1<c2) = %d\n", sizeof(c1<c2)); > printf ("sizeof(i+r) = %d\n", sizeof (i+r)); > printf ("sizeof(i+s) = %d\n", sizeof (i+s)); > } well, each of the sizeofs are in the form of sizeof expression, so let's dicect each expression to figure out what it should print. First, sizeof(mist.a1) -- well, my guess (and this is pretty shakey) is that mist.a1 is really just a single bit of an unsigned int and not an entity of itself, so its size is that of an unsigned int, which, on the VAX, is 4 bytes. sizeof(c1<c2) -- this is an expression, the type of which is independent of the types of c1 and c2. A "logical" expression in C is just an integer expression, and an integer requires 4 bytes on a VAX. sizeof (i+r) -- well, here we have to look at the promotion rules for expressions. On silly compilers (like that on a VAX running VMS), all floats get promoted to doubles (rather than having the option of having them not be promoted to allow single precision math like on SUNs, but I digress). So, the float r gets promoted to a double, the other argument of the binary operator '+' gets promoted to a double, and the result is a double which, on a VAX, requires 8 bytes (am I doing okay so far?) and, lastly, sizeof (i+s) -- s the name of an array of 20 chars. The name of an array is a pointer to the zeroth element of that array (pointers require 4 bytes on a VAX). Adding an int to a pointer (even in this order) should provide a pointer -- 4 bytes. does this help? Wade Guthrie Rockwell International Anaheim, CA (Rockwell doesn't necessarily believe / stand by what I'm saying; how could they when *I* don't even know what I'm talking about???)
henry@utzoo.uucp (Henry Spencer) (12/11/88)
In article <2976@arcturus> evil@arcturus.UUCP (Wade Guthrie) writes: >sizeof (i+r) -- well, here we have to look at the promotion rules for >expressions. On silly compilers (like that on a VAX running VMS), >all floats get promoted to doubles... Not too silly, actually, since this behavior was required by the old C standards (K&R and its immediate descendants). -- SunOSish, adj: requiring | Henry Spencer at U of Toronto Zoology 32-bit bug numbers. | uunet!attcan!utzoo!henry henry@zoo.toronto.edu