Bob Rossi wrote:
Yes, that worked! I suggest putting a comment in the patch that says
that crypt must be done before postgresql, otherwise it will fail to
test link.
Actually that's half-assed. We have to back out the changes that the DBD
detection causes for LDFLAGS and LIBS, etc, in order to ensure that later
tests aren't bogus. That's true all over the place throughout apr and httpd
m4 macros, but apparently the dbd detection authors overlooked it.
Just glad you caught it and identified the root problem!
Bill
