Re: b41 cygwin suckfest, gcc -O3 -> -O2 *please*
At 12:06 AM 5/26/98 -0700, Martin Buchholz wrote:
>>>>> "Andy" == Andy Piper
<andyp(a)parallax.co.uk> writes:
Andy> b41 is really turning into a suckfest under cygwin for me.
Andy> Martin, gcc -O3 loses big time under cygwin. Can we possibly undo this
Andy> change?
Sure, but note that it is working for all us Unix guys. Before
turning back the opt level to -O2, try real hard to check whether this
is really a source code bug. Most `optimizer' bugs are in fact bugs
in the program being compiled. Especially in this case, where the
same compiler at the same optimization level on the same hardware has
already proven its worth.
Its definitiely a compiler bug. In one instance gcc optimizes away 1 of two
almost identical functions used almost identically. There is then a link
time error for the undefined function.
I am happy to set it to -O2 for cygwin only if you'll tell me how. Note
that I have had problems in the past with optimization under cygwin
especially with compiling C++.
andy
---------------------------------------------------------------------------
" .sigs are like your face - rarely seen by you and uglier than you think"
Dr Andy Piper, Technical Architect, Parallax Solutions Ltd
mail: andyp(a)parallax.co.uk web: www.parallax.co.uk/~andyp