Re: define bool in pgtypeslib_extern.h

Amit Kapila <amit(dot)kapila16(at)gmail(dot)com> writes:
> On Sat, Oct 26, 2019 at 10:49 PM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> I'm inclined to think that we need to make ecpglib.h's bool-related
>> definitions exactly match c.h, which will mean that it has to pull in
>> <stdbool.h> on most platforms, which will mean adding a control symbol
>> for that to ecpg_config.h. I do not think we should export
>> HAVE_STDBOOL_H and SIZEOF_BOOL there though; probably better to have
>> configure make the choice and export something named like PG_USE_STDBOOL.

> This sounds reasonable to me, but we also might want to do something
> for probes.d. To be clear, I am not immediately planning to write a
> patch for this.

As far as probes.d goes, it seems to work to do

@@ -20,7 +20,7 @@
#define BlockNumber unsigned int
#define Oid unsigned int
#define ForkNumber int
-#define bool char
+#define bool _Bool

provider postgresql {

although removing the macro altogether leads to compilation failures.
I surmise that dtrace is trying to compile the generated code without
any #include's, so that only compiler built-in types will do.

(I tried this on macOS, FreeBSD, and NetBSD, to the extent of seeing
whether a build with --enable-dtrace goes through. I don't know
enough about dtrace to test the results easily, but I suppose that
if it compiles then this is OK.)

This would, of course, not work on any platform where we're not
using <stdbool.h>, but I doubt that the set of platforms where
dtrace works includes any such.

A plausible alternative is to do

-#define bool char
+#define bool unsigned char

which is correct on platforms where we don't use <stdbool.h>,
and is at least no worse than now on those where we do. In
practice, since we know sizeof(_Bool) == 1 on platforms where
we use it, this is probably just fine for dtrace's purposes.

Anyone have a preference?

regards, tom lane