Re: Fwd: 5.8: LTO exposes some new issues

["Daniel Shahaf" <d.s@xxxxxxxxxxxxxxxxxx>]

Tomasz, if you devise a patch for these issues, please post it so it can
be reviewed and applied.  (If you prefer to wait for the Internet to
write the patch you want, feel free to do so, but that strategy doesn't
always work.)

Cheers,

Daniel


Bart Schaefer wrote on Sat, 25 Jul 2020 20:05 +00:00:
> I am not likely to pursue this further, I don't have a suitable build
> environment.
> 
> ---------- Forwarded message ---------
> From: Tomasz Kłoczko <kloczko.tomasz@xxxxxxxxx>
> Date: Sat, Jul 25, 2020 at 11:59 AM
> Subject: Re: 5.8: LTO exposes some new issues
> To: Bart Schaefer <schaefer@xxxxxxxxxxxxxxxx>
> 
> 
> On Sat, 25 Jul 2020 at 18:43, Bart Schaefer <schaefer@xxxxxxxxxxxxxxxx> wrote:
> >
> > On Tue, Jul 21, 2020 at 11:00 PM Daniel Shahaf <d.s@xxxxxxxxxxxxxxxxxx> wrote:
> > >
> > > Haven't looked at the other warnings you reported.
> >
> > They all appear to be clashes where the same pointer is declared
> > "extern" in a header file but not so in the original source.  It looks
> > as if the "mod_import_variable" preprocessor definition is supposed to
> > deal with this, perhaps it's not being properly defined in the
> > situation where Tomasz is compiling.
> 
> 
> Yes and they need to be resolved because even single such warnings and
> ld abandons applying LTO.
> If you will decide which version should be used in both locations and
> will have some patch for that please let me know about that because I
> can test that kind of patch instantly proving that there is no other
> coalition with LTO.
> 
> kloczek
> -- 
> Tomasz Kłoczko | LinkedIn: http://lnkd.in/FXPWxH
>