Hi Pavel!

On Mon, May 22, 2000 at 05:24:52PM -0400, Pavel Roskin wrote:
> I believe that it's a good default for those who compile packages (as
> opposed to those who download binaries and those who develop that
> software).

This is one of those things that I have to disagree with.  :-)

Rather than having autoconf make yet another decision on behalf of the
package maintainer/user/developer, wouldn't it be better to provide a
"--enable-debug" flag and a "--enable-optimize=level", instead?

> This way you get an executable that you can occasionally debug if you
> encounter some problems, and yet it runs on the full speed. You can
> strip the executable if you don't want debug information.

Right, but shouldn't the user have the ability to choose not to have
debugging information right from the start, without having to fake out
configure by setting CXXFLAGS beforehand?

> Package managers are welcome to provide different defaults (such as
> -mpentium -O6)

Sure, I do that too.  However, suppose I add flags to CXXFLAGS from
within `configure.in' after AC_PROG_CXX has run, and the user hasn't
supplied any CXXFLAGS to the configure script then the CXXFLAGS
variable will contain "-g -O2," despite the fact those flags may not
be desired.

> If you want to debug the software seriously and/or compile it fast
> (perhaps many times) you are welcome to write a script for running
> configure with your preferred arguments.

Isn't that a bit extreme for this simple case?  Why should I, as a
configure script maintainer, force the user to write a script that
runs configure simply to get rid of the "-g -O2?"  Wouldn't it be
better to provide a means to disable one or both of these flags
without having to fake out the the configure script, but have them
enabled by default?

-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

Reply via email to