Hi Paul,
On Mon, May 22, 2000 at 05:44:44PM -0400, Paul D. Smith wrote:
> %% Ossama Othman <[EMAIL PROTECTED]> writes:
>
> oo> I've had complaints from users about the fact that my configure script
> oo> adds "-g -O2" to CXXFLAGS by default since they want to compile an
> oo> optimized version of the library without debugging symbols. I know
> oo> that gcc/g++ can generate debugging symbols for optimized code, but
> oo> that's not the issue. The issue here is that the user doesn't appear
> oo> to have a choice about the "-g -O2" flags. Is this correct? If so,
> oo> why does autoconf add these? Shouldn't it be up to the package
> oo> maintainer to decide whether these flags get added to the compiler
> oo> flags?
>
> It's not correct. If they read the INSTALL file (you do ship the
> autoconf INSTALL file with your package, right?)
Yes, of course. "automake --add-missing" is my friend. :-)
> they'll see they can do
> either of these things to override the default value:
>
> $ CXXFLAGS=-g ./configure ...
> $ make
>
> (that's sh/ksh/bash/zsh syntax; if you're using csh/tcsh you need to add
> an "env " to the beginning of the configure line) or
>
> $ ./configure ...
> $ make CXXFLAGS=-g
I knew about the environment options but I didn't realize that it
would disable the "-g -O2," for example. That's what I get for not
reading INSTALL more than once every two years. :-)
> (the latter assumes you're makefiles are properly constructed so that
> they don't add important options to CXXFLAGS and leave them to the user
> for modification, as the GNU coding standards require).
Yes, indeed. I don't reset CXXFLAGS in the Makefile at all. What I
would like is a way to disable the automatically selected "-g -O2"
flags by some configure script options, rather than using the
CXXFLAGS=whatever hack. Note that I'm not saying that CXXFLAGS is
itself a hack. What I feel is a hack is the fact that it must be used
to disable the "-g -O2" that autoconf adds to the compiler flags.
Does this make any sense?
> HTH! :)
Thanks! ;-)
-Ossama
--
Ossama Othman <[EMAIL PROTECTED]>
Distributed Object Computing Laboratory, Univ. of California at Irvine
1024D/F7A394A8 - 84ED AA0B 1203 99E4 1068 70E6 5EB7 5E71 F7A3 94A8
