I've fixed this bug in 7.1RC1. It may have been causing core dumps by
dereferencing a null function pointer. As it was the Sun compiler
flagged it as an error. (SC5.0 on Solaris 2.6.)
*** src/bin/pg_dump/pg_backup_null.c.orig Sat Mar 24 00:54:54 2001
--- src/bin/pg_dump/pg_backup_null.c Tue Mar 27 09:50:24 2001
***************
*** 98,104 ****
static void
_PrintTocData(ArchiveHandle *AH, TocEntry *te, RestoreOptions *ropt)
{
! if (*te->dataDumper)
{
AH->currToc = te;
(*te->dataDumper) ((Archive *) AH, te->oid, te->dataDumperArg);
--- 98,104 ----
static void
_PrintTocData(ArchiveHandle *AH, TocEntry *te, RestoreOptions *ropt)
{
! if (te->dataDumper)
{
AH->currToc = te;
(*te->dataDumper) ((Archive *) AH, te->oid, te->dataDumperArg);
Interestingly other compilers on other platforms accept a function as
a controlling expression in an if statement without complaint. That
expression should have scalar type, i.e. either arithmetic or pointer.
This test program demonstrates Sun catching the error.
int (*f)(int);
int g(int x)
{
return x;
}
int main(void)
{
int y = 0;
f = g;
if (*f) {
y = f(1);
}
return y;
}
$ cc funcptr.c
"funcptr.c", line 13: controlling expressions must have scalar type
cc: acomp failed for funcptr.c
--
Pete Forman -./\.- Disclaimer: This post is originated
WesternGeco -./\.- by myself and does not represent
[EMAIL PROTECTED] -./\.- opinion of Schlumberger, Baker
http://www.crosswinds.net/~petef -./\.- Hughes or their divisions.
---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
