From: Gianluca S. <gi...@gm...> - 2007-03-18 08:57:23
|
On 3/18/07, techtonik <tec...@us...> wrote: > > I don't know the gains either, my guess is none for mingw and w32api. > > The -O3 adds inline optimazations additionally over -O2. For debugging > > -O0 is what I use and recommend because any optimization may confuse > > where to look in the code including confusing the debugger but we are > > not talking about a debugging release. > > It is possible to extract debugging information and build stacktraces > with symbolic information from release binaries. It is the aim of > breakpad project I''ve mentioned here some time ago. I am not an > expert to say if -O3 switch will influence the process, but to me it > seems that it is very likely to happen. Yes. RPMs are built using both -O2 and -g, then the debug info are stripped and put into a separate -debuginfo package: when one need to debug a binary, installing the -debuginfo bits is all that is needed (i.e. no recompilation with -g) to get started Again, I am not sure if this is applicable to Windows executables, nor if there is any interstest in such a feature for MinGW binaries. |