This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: proposal: extend TARGET_OPTIONS structure



> > 	-mno-align	(provides default of 2)
> > 	-malign		(provides default of 16)
> > 	-malign=4
> > 	-malign=8
> >
> > 	-mhard-float	(provides default of "")
> > 	-msoft-float	(provides default of " ")
> > 	-msoft-float=-lfloat

They're *examples*.  I made them up, you don't have to tell me what
other options should be used because that's not my point.  Let me try
again with an option that gcc doesn't already have:

-mno-align-jcxz-loops
-malign-jcxz-loops
-malign-jcxz-loops=4

My point is that sometimes even boolean options can benefit from a
not-strictly-boolean state, which is what we get if we use an extended
TARGET_SWITCHES to replace TARGET_OPTIONS.  Of course, I'm not
advocating removing TARGET_OPTIONS yet either - just extending
TARGET_SWITCHES to be more flexible.

As for -msoft-float, that meant to use a floating point library *and*
specify the library.  DJGPP would use such an option (although not
that specific one), since we have three floating point emulator
options, with one being the default and the other two rarely used.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]