Tom Lane wrote:
> William Gray <grayw@mail.montclair.edu> writes:
>
>> I'm not sure how I'd go about determining that. I don't recall messing
>> with any of my system's .h files. Do you know where I might look?
>>
>
> grep through all the files under /usr/include looking for "bool" ...
>
> regards, tom lane
>
> ---------------------------(end of broadcast)---------------------------
> TIP 9: In versions below 8.0, the planner will ignore your desire to
> choose an index scan if your joining column's datatypes do not
> match
>
Looks like there are quite a few possible culprits, as it seems to be an
often-defined thing (but I must not shrink from the task...)
c.h defines a bool, and this looks correct and un-tampered with (grep
output)
> c.h: * file. Added bool enum from machine/types.h for regular users
> c.h: * that want a real boolean type.
> c.h:#ifndef bool
> c.h:typedef enum { false = 0, true = 1 } bool;
> c.h:#endif /* bool */
Both curses.h and ncurses.h appear to use whatever the compiler
considers bool, and if absent, defines their own.
There's a file stdbool.h that may be at fault:
> stdbool.h: * $FreeBSD: src/include/stdbool.h,v 1.6 2002/08/16 07:33:14
> alfred Exp $
> stdbool.h:#define __bool_true_false_are_defined 1
> stdbool.h:#define bool _Bool
That is explicitly redefining 'bool' to be of type _Bool.
I feel a little lost here. I'm pretty sure I've never messed with these
files. Is it possible that at some point if I configured some piece of
software with a prefix of /usr that it could have added some header file
to /usr/include that would adversely affect other builds, like postgres?
Thanks for your help,
Bill