Il 11/01/2014 08:46, Stefan Weil ha scritto: > Hi Steven, > > --disable-stack-protector would also be useful for platforms which make > debugging of executables with stack protection difficult. When I must > debug Windows executables, I always disable stack protection, because > otherwise the stack back traces are unreadable. > > So, for MinGW it might be reasonable to set the default to no stack > protection if --enable-debug is selected. This requires some > modifications in your patch. > > # Don't set it to "yes" initially: > stack_protector="" > > # Do the compile test if it is not "no": > if test "$stack_protector" != "no"; then
Apart from this little detail, the patch looks good. Steven, please resend the patch as a top-level message with the patch inline rather than attached. This is needed so that Anthony's script will pick it up. Including these changes would be nice too. Paolo > The usual logic is do nothing if the user says "no". Run the compile > tests otherwise. Show an error message if the compile tests fail and the > user said "yes". See the code which handles $pie for an example. > > Please send your next patch inline - this makes it easier to add comments. > > Best regards > > Stefan >