[OpenJDK 2D-Dev] New warnings in layout code

[Andrew Hughes]

----- Original Message -----
> Hi,

snip...

> 
> > Arguably the test could be removed but I would prefer to leave it in
> > and change the code it so it is actually tested.
> >
> > -fno-strict-overflow is just too compiler-specific and too far removed
> > from the code to be something I would want to rely upon long term.
> 
> -fwrapv is another option and seems to be supported by gcc and clang, at
> least.
> 

This is very compiler-specific. We had to add compiler options to work
around this as best as possible for the security update, but I don't think
that logic should go upstream if possible. -fwrapv was necessary on GCC < 4.2
(used by RHEL 5) and -fno-strict-overflow on later versions. I also didn't see
the warnings locally with GCC 5.1. They were flagged after our RPMs were created.

> That said, do you think some sort of add-and-check-for-overflow test
> would be suitable here? I can take a shot at implementing it, if you
> like.
> 
> > Interestingly I see the warning in the 8u71 logs using gcc 4.3.0,
> > but don't see the warning in RE's JDK 9 build log using gcc 4.9.2
> > I don't see it directly suppressed either but perhaps -Wno-type-limits
> > masks
> > it ?
> 
> On my box, I see the same warnings on jdk9. Please see the attached
> partial build log (from `make all LOG=trace`).

I can confirm these are the same warnings I saw in building the RHEL RPMs for u71.

> 
> Thanks,
> Omair
> 
> --
> PGP Key: 66484681 (http://pgp.mit.edu/)
> Fingerprint = F072 555B 0A17 3957 4E95  0056 F286 F14F 6648 4681
> 

-- 
Andrew :)

Senior Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

PGP Key: ed25519/35964222 (hkp://keys.gnupg.net)
Fingerprint = 5132 579D D154 0ED2 3E04  C5A0 CFDA 0F9B 3596 4222