[boost] Re: fix: intel.hpp

From: Gennaro Prota (gennaro_prota_at_[hidden])
Date: 2003-04-26 05:47:53

On Sat, 26 Apr 2003 01:33:38 +0200, "Pavel Vozenilek"
<pavel_vozenilek_at_[hidden]> wrote:

>It is even more complicated than that. I tried 12 different combinations
>(all Windows) and put results into table (hope formatting stays readable):
>
>[table omitted]

Thank you very much for that hard work Pavel! I guess such a table
would turn out very useful for all of us if you added some other
relevant information. Intel behavior depends very much on the switches
that are in icl6.cfg, icl7.cfg (when used from the VC++ IDE) and
icl.cfg (when used from the command line). That's, at least, for
Win32. I also suspect there are important differences between the
Win32 version and the Linux one (the Win32 version tends, by default,
to behave as Microsoft; the Linux one tends to be more sensible: for
instance IIRC it enables ADL by default). Even if you don't have
access to the Linux version, I think your table would be very handy if
it also included the test program and the switches, among e.g. -Qvc6,
-Za, -Qoption,c,--no_microsoft and -Qoption,c,--strict, that affect
the issue at hand (I have listed the Win32 options; BTW, this paper
contains a useful cross-reference:

http://www.intel.com/software/products/compilers/c60l/resources/c_ug_lnx.pdf

)

Genny.