stealth@engin.umich.edu (Mike Pelletier) (04/04/91)
Having set up the software to use the fast stdio routines, I get the following
error when it attempts to compile them:
/bin/cc '-DPTR_TYPE=unsigned char *' -O -s -c rdwr.c
ccom: Error: rdwr.c, line 20: redeclaration of fread
int
fread(ptr, size, count, fp)
char *ptr;
---^
ccom: Error: rdwr.c, line 70: redeclaration of fwrite
int
fwrite(ptr, size, count, fp)
char *ptr;
---^
*** Error code 1
Stop.
+ exit 1
Is there any way I can coax the C compiler to accept these new
versions of stdio routines? Considering the speedup they provide,
I would like to use them if at all possible.henry@zoo.toronto.edu (Henry Spencer) (04/04/91)
In article <1991Apr3.173459.3081@engin.umich.edu> stealth@engin.umich.edu (Mike Pelletier) writes: > char *ptr; > ---^ >ccom: Error: rdwr.c, line 70: redeclaration of fwrite You can try changing those "char *"s to "void *"s, although this will mean further work within the routines to make it all turn out right. Not a trivial fix, I'm afraid. -- "The stories one hears about putting up | Henry Spencer @ U of Toronto Zoology SunOS 4.1.1 are all true." -D. Harrison| henry@zoo.toronto.edu utzoo!henry
gamiddle@watmath.waterloo.edu (Guy Middleton) (04/04/91)
In article <1991Apr3.190011.1117@zoo.toronto.edu> henry@zoo.toronto.edu (Henry Spencer) writes: > In article <1991Apr3.173459.3081@engin.umich.edu> stealth@engin.umich.edu (Mike Pelletier) writes: > > char *ptr; > > ---^ > >ccom: Error: rdwr.c, line 70: redeclaration of fwrite > > You can try changing those "char *"s to "void *"s, although this will mean > further work within the routines to make it all turn out right. Not a > trivial fix, I'm afraid. Actually, the fix is to change int fread(ptr, size, count, fp) to size_t fread(ptr, size, count, fp) Changing "char *" to "void *" fixes the warnings, but not the errors.
stealth@caen.engin.umich.edu (Mike Pelletier) (04/04/91)
In article <1991Apr4.035820.8952@watmath.waterloo.edu> gamiddle@watmath.waterloo.edu (Guy Middleton) writes: >In article <1991Apr3.190011.1117@zoo.toronto.edu> henry@zoo.toronto.edu (Henry Spencer) writes: >> In article <1991Apr3.173459.3081@engin.umich.edu> I write: >> > char *ptr; >> > ---^ >> >ccom: Error: rdwr.c, line 70: redeclaration of fwrite >> >> You can try changing those "char *"s to "void *"s, although this will mean >> further work within the routines to make it all turn out right. Not a >> trivial fix, I'm afraid. > >Actually, the fix is to change > > int > fread(ptr, size, count, fp) > >to size_t > fread(ptr, size, count, fp) > >Changing "char *" to "void *" fixes the warnings, but not the errors. Thanks, Guy, that worked perfectly! I still get two warnings out of that file, namely: ccom: Warning: rdwr.c, line 44: illegal pointer combination fp->_ptr = (unsigned char *)bp; --------------------------------------------------------^ ccom: Warning: rdwr.c, line 93: illegal pointer combination fp->_ptr = (unsigned char *)bp; --------------------------------------------------------^ but it does indeed produce a .o file in spite of these complaints. Why does changing "int" to "size_t" cure this particular problem? Isn't size_t just typedef'ed to int in one of the include files? -- Mike Pelletier | The University of Michigan's | [this section intentionally] Computer Aided Engineering Network | [ left blank ] Usenet, UUCP, IRC and mail admin |
gamiddle@watmath.waterloo.edu (Guy Middleton) (04/05/91)
In article <1991Apr4.150426.25479@engin.umich.edu> stealth@caen.engin.umich.edu (Mike Pelletier) writes: > Why does changing "int" to "size_t" cure this particular problem? > Isn't size_t just typedef'ed to int in one of the include files? It is, but fread() and fwrite() were never declared in <stdio.h> before, and are now. They always returned an int before, and size_t is now an unsigned int. I was wrong about the void * fix for the warnings; I think you really want to change '-DPTR_TYPE=unsigned char *' in the makefile to '-DPTR_TYPE=char *'. The difference here is with the compiler; it never used to complain about conversions from unsigned char to char. -Guy Middleton, University of Waterloo gamiddleton@watmath.waterloo.edu (+1 519 885 1211 x3472) gamiddleton@watmath.uwaterloo.ca