Re: [Plib-users] problem with fnt and release mode

[Scott S. <sjs...@um...>]

Yeah, it can be.  I don't fault the optimizer, though, if you turn on "full
optimizations" you should be prepared to face the consequences.  There is
certainly an option called "assume aliasing across functions."  Although I
do question MS's decision to not make this the default optimization choice.

As for only affecting the G200/G400/Radeon, my guess is that NVidia drivers
are more robust when it comes to error-handling.

------------------
Scott Shumaker
sjs...@um...


----- Original Message -----
From: "Steve Baker" <sjb...@ai...>
To: <pli...@li...>
Sent: Thursday, December 21, 2000 1:07 AM
Subject: Re: [Plib-users] problem with fnt and release mode


> Scott Shumaker wrote:
>
> > Well, it only causes problems across functions, i.e.:
> <snip>
>
> Yes - that seems *incredibly* dangerous to me.  I can imagine a TON of
> places in PLIB where that would break.
>
> I wonder why it only affects G200/G400 and Radion?
>
> The OpenGL driver is a DLL - so the compiler doesn't know what driver
> you are using...it's hard to imagine a pointer alias that would trip
> up the optimiser with one driver and not the other.
>
> Just out of interest, what version of MSVC++ is everyone who sees this
> bug using?   I recollect someone telling me that the optimisations that
> V6 makes are *far* more dangerous than the ones in V7.
>
> --
> Steve Baker   HomeEmail: <sjb...@ai...>
>               WorkEmail: <sj...@li...>
>               HomePage : http://web2.airmail.net/sjbaker1
>               Projects : http://plib.sourceforge.net
>                          http://tuxaqfh.sourceforge.net
>                          http://tuxkart.sourceforge.net
>                          http://prettypoly.sourceforge.net
>
>
> _______________________________________________
> plib-users mailing list
> pli...@li...
> http://lists.sourceforge.net/mailman/listinfo/plib-users