On Sun, Jan 23, 2022 at 10:43:11AM +0100, Kornel Benko wrote:
> Am Sat, 22 Jan 2022 19:06:59 +0100
> schrieb Kornel Benko <kor...@lyx.org>:
> 
> > Am Sat, 22 Jan 2022 18:35:12 +0100
> > schrieb Kornel Benko <kor...@lyx.org>:
> > 
> > > Am Sat, 22 Jan 2022 10:04:44 -0500
> > > schrieb Scott Kostyshak <skost...@lyx.org>:
> > >   
> > > > With Clang and fsanitize, I get the following error when I open a 
> > > > buffer:
> > > > 
> > > >   runtime error: load of value 4294967293, which is not a valid value 
> > > > for type
> > > > 'const lyx::Update::flags'
> > > > 
> > > > Attached is the backtrace (which is printed to terminal if I set 'export
> > > > UBSAN_OPTIONS=print_stacktrace=1').
> > > > 
> > > > Scott    
> > > 
> > > I see the same here. (No crash, but the messages on terminal)
> > > 
> > >   Kornel  
> > 
> > If I understand it correctly, the routine
> >     inline flags operator&(flags const f, flags const g)
> > gets the value f as:
> >     (update_flags & ~Update::Force)
> > Since 32-bit-value 4294967293 is the same as '~Update::Force', which is too 
> > big for
> > 32-bit integer, we possibly have the same 'problem' as with colors.
> > 
> > BTW Update::Force is defined as '2' and the hex value of 4294967293 is 
> > FFFFFFFD.
> > 
> >     Kornel
> 
> Here 2 possible patches which do not show the behaviour.
> (I prefer the number 2)

Why is this issue that fsanitize found not actually a "real issue"? Why
does it not lead to any potential bug? I never understood that with the
color code either. Does it depend on the compiler?

Scott

Attachment: signature.asc
Description: PGP signature

-- 
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel

Reply via email to