Hi,
On 2023-03-16 13:54:29 -0400, Tom Lane wrote:
> Andres Freund <andres@anarazel.de> writes:
> > On 2023-03-16 12:10:27 -0400, Tom Lane wrote:
> >> It wouldn't be entirely surprising if meson is selecting some -W
> >> switches that the configure script doesn't ... but I don't know
> >> where to check or change that.
>
> > I think it's just that meson defaults to -O3 (fwiw, I see substantial gains of
> > that over -O2). I see such warnings with autoconf as well if I make it use
> > -O3.
>
> Oh, interesting. Should we try to standardize the two build systems
> on the same -O level, and if so which one?
I'm on the fence on this one (and posed it as a question before). O3 does
result in higher performance for me, but it also does take longer to build,
and increases the numbers of warnings.
So I just elected to leave it at the default for meson.
> To my mind, you should ideally get the identical built bits out of
> either system, so defaulting to a different -O level seems bad.
I doubt that is attainable, unfortunately. My experience is that even trivial
changes can lead to substantial changes in output. Even just being in a
different directory (the root build directory for meson vs the subdirectory in
make builds) apparently sometimes leads to different compiler output.
> I'm not sure if we're prepared to go to -O3 by default though,
> especially for some of the older buildfarm critters where that
> might be buggy. (I'd imagine you take a hit in gdb-ability too.)
My experience is that debuggability is already bad enough at O2 that the
difference to O3 is pretty marginal. But it certainly depends a bit on the
compiler version and what level of debug information one enables.
Greetings,
Andres Freund
