Re: pgsql: Fix blatantly broken record_image_cmp() logic for pass-by-value

Kevin Grittner <kgrittn(at)ymail(dot)com> writes:
> Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> Kevin Grittner <kgrittn(at)ymail(dot)com> writes:
>>> I don't get a warning on this with either of these compilers,
>>> either with or without asserts enabled:

>> Perhaps you built with -O0? At least in older versions of gcc, you need
>> at least -O1 to get uninitialized-variable warnings.

> This is with default configure and compile options, which results
> in this:

> gcc -O2 -Wall -Wmissing-prototypes -Wpointer-arith -Wdeclaration-after-statement -Wendif-labels -Wmissing-format-attribute -Wformat-security -fno-strict-aliasing -fwrapv -fexcess-precision=standard -g -I../../../../src/include -D_GNU_SOURCE -I/usr/include/libxml2 -c -o rowtypes.o rowtypes.c -MMD -MP -MF .deps/rowtypes.Po

> ... It is pretty disturbing that the compilers in my
> Linux distro don't recognize that as leaving the value
> uninitialized.

That is a serious compiler bug which you should file with your distro
forthwith. It definitely does show up with what I'm using:
gcc (GCC) 4.4.7 20120313 (Red Hat 4.4.7-3)
It might be worth trawling the buildfarm records to see which compilers
did or didn't warn before.

regards, tom lane