stever@Octopus.COM (Steve Resnick ) (09/08/89)
I have a shell I am writing for MS DOS which will extend the batch language
and add some UNIX style enhancements (aliasing, history, and getting rid of
that damn backslash). As I write this I want to port it to UNIX/Xenix as well.
my problem is that the Xenix C compiler bitches at using pointers to
functions. For example, consider the follow fragment...
int echo(int, char **); /* The protos allow the forward refernce to */
int setenv(int, char **); /* the functions */
struct ci /* This is the command table */
{
char * name; /* The shell uses this name to compare
against the entered command */
int (* fptr)(int, char **); /* and here is the function pointer */
} Cmds[] =
"ECHO",echo,
"SET",setenv,
} ;
.
. /* Main and other stuff here */
.
.
/* When I get to the point where I execute an internal function, I call the
routine here..
*/
(Cmds[idx].fptr)(c,v); /* Here is where I call the function */
This is what the compiler doesn't like. The error message is something like
Term does not evaluate to a function.
This seems to be legitimate C in the MS DOS world...
Help! Please!karl@haddock.ima.isc.com (Karl Heuer) (09/09/89)
In article <1989Sep8.021802.29203@Octopus.COM> stever@octopus.UUCP (Steve Resnick ) writes: >When I get to the point where I execute an internal function... > (Cmds[idx].fptr)(c,v); /* Here is where I call the function */ >This is what the compiler doesn't like. The ability to invoke a function via a pointer without an explicit dereference is new to ANSI C, and is not supported by all pre-ANSI compilers. The more portable syntax is (*Cmds[idx].fptr)(c,v); Karl W. Z. Heuer (ima!haddock!karl or karl@haddock.isc.com), The Walking Lint
scjones@sdrc.UUCP (Larry Jones) (09/09/89)
In article <1989Sep8.021802.29203@Octopus.COM>, stever@Octopus.COM (Steve Resnick ) writes: > my problem is that the Xenix C compiler bitches at using pointers to > functions. For example, consider the follow fragment... > (Cmds[idx].fptr)(c,v); /* Here is where I call the function */ > > This is what the compiler doesn't like. The error message is something like > > Term does not evaluate to a function. The error message says it all -- (Cmds[idx].fptr) is not a function, it's a POINTER to a function. You need to use (*Cmds[idx].fptr) for many existing implementations, although ANSI C allows as few or as many asterisks as you like. Once again, let's all repeat in unison: In C, the use of a variable and it's declaration should look the same. ---- Larry Jones UUCP: uunet!sdrc!scjones SDRC scjones@SDRC.UU.NET 2000 Eastman Dr. BIX: ltl Milford, OH 45150-2789 AT&T: (513) 576-2070 "I have plenty of good sense. I just choose to ignore it." -Calvin
schaut@madnix.UUCP (Rick Schaut) (09/10/89)
In article <1989Sep8.021802.29203@Octopus.COM> stever@octopus.UUCP (Steve Resnick ) writes: [Defines an array of structures where one element of the structure is a pointer to a function returning an int.] | /* When I get to the point where I execute an internal function, I call the | routine here.. | */ | (Cmds[idx].fptr)(c,v); /* Here is where I call the function */ | | This is what the compiler doesn't like. The error message is something like | | Term does not evaluate to a function. Exactly, the term evaluates to a pointer. You need to go through a level of indirection to get to the function that the pointer points to. The correct expression is: (*Cmds[idx].fptr)(c,v); See K&R 5.11 (p. 118, 2nd Ed.). See, also, operator precedence on p. 52. -- Richard Schaut ArpaNet: madnix!schaut@cs.wisc.edu UseNet: ...uwvax!astroatc!nicmad!madnix!schaut {decvax!att}!
perry@ccssrv.UUCP (Perry Hutchison) (09/13/89)
In article <1989Sep8.021802.29203@Octopus.COM> stever@octopus.UUCP (Steve Resnick) writes: > int echo(int, char **); /* The protos allow the forward refernce to */ > int setenv(int, char **); /* the functions */ > > struct ci /* This is the command table */ > { > char * name; /* The shell uses this name to compare > against the entered command */ > int (* fptr)(int, char **); /* and here is the function pointer */ > } Cmds[] = > "ECHO",echo, ^^^^ > "SET",setenv, ^^^^^^ > } ; ... > (Cmds[idx].fptr)(c,v); /* Here is where I call the function */ > > The error message is something like > > Term does not evaluate to a function. The type of member fptr is "pointer to a function", but the initial values given are "functions". Try preceding them with "&" thus "ECHO", &echo, ^ "SET", &setenv, ^ I would call it a bug that the compiler complained only about the call. It should have at least given a warning on the initialization. Also, in the call, you need to dereference the "pointer to function" in order to obtain a "function" which you can call, thus (* Cmds[idx].fptr)(c,v); /* Here is where I call the function */ ^
stever@Octopus.COM (Steve Resnick ) (09/15/89)
I appreciate the help everyone has given me. It seems that I've grown quite used to my ANSI compiler on the DOS machine. The lack of the indirection was the problem. I could live with better error messages than Microsoft C but alas that is the Xenix C compiler. Ah well.... Thanx to all who've helped. ---------------------------------------------------------------------------- | Confusing and counfounding a hopelessley defenseless department who was | | unfortunate to ask for the informationin the first place. | |--------------------------------------------------------------------------| | Steve Resnick | ...!sun!vsi1!octopus!psisj!steve | |---------------| or ...!sun!vsi1!octopus!stever | |--------------------------------------------------------------------------|