This is the mail archive of the gcc-patches@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]

Re: Error handling change


    >     Well the ARM compiler has some, eg:
    > 
    >         fatal ("Invalid floating point emulation option: -mfpe-%s",
    >                target_fp_name);
    > 
    > That seems quite wrong to me.  The other ports just use "error".

    I don't see why.  Error simply means print a message and continue.  Fatal 
    means print a message and stop.  It doesn't mean the compiler is buggy.

Why not continue when an invalid -m option is specified?  That's what the
other config files do?

At one point, calling fatal used to mean there was an internal
compiler error, which is why it had the hook to call back to the front end.

I only see a handful of calls to fatal that don't do that and the oens I've
looked at all look rather dubiuos to me.

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