From: luke AT jet DOT msk DOT su (Oleg Yu. Polyanski)
Newsgroups: comp.os.msdos.djgpp
Subject: Re: DJGPP 2.01 to 2.72 conversion
Date: 05 Jun 1998 15:23:25 +0400
Organization: Jet Infosystems
Lines: 31
Message-ID: <huug1hk6qj6.fsf@goliath.jet.msk.su>
References: <35778687 DOT 1E93871E AT kiwi DOT gen DOT nz>
NNTP-Posting-Host: goliath.service.jet.msk.su
Mime-Version: 1.0 (generated by tm-edit 7.108)
Content-Type: text/plain; charset=US-ASCII
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Precedence: bulk

>>>>> "L" == L T de La Haye <lance AT kiwi DOT gen DOT nz> writes:

L> I am porting some C code, which was compiled under DJGPP 2.01
L> originally. I am compiling it in 2.72 now, and it appears to run
L> fine.  During the make I had some hassles with command line
L> arguments though:

	you misunderstand djgpp distribution version (which is currently
2.01) and gnu c compiler version (there are available 2.7.2.1, 2.8.0,
2.8.1 in distrbution). to know exactly version of your compiler, try:

	gcc -v

L> -mpentium seemed to exist on 2.01 but not 2.72.  The highest x86
L> supported on 2.72 seems to be -m486 -Wno-sign-compare also seems to
L> have disappeared. I can't find the new equivalent. This one
L> disables the "Comparison between signed and unsigned" (or some
L> such) warning.  Considering the program (MAME) compiles with
L> -Werrors (all warnings are errors - presently disabled) I would
L> like to resolve this one. I get about 100 of those warnings during
L> the make. I really don't want to have fix all those. Most are
L> pretty trivial.

	use command `info gcc invo' and select the chapter in manual
that is interesting for you.
	gcc 2.8.x has extra optimization abilities for p5 and p6 but
optimization for p6 is broken (conditional move instructions are
issued incorrectly) and should not be used. use `-mcpu=i586 -march=i586'
to produce p5 optimized programs.

/Oleg