Re: [PATCH/fortran] Add option to make 1st error fatal

[Steve Kargl <sgk at troutmask dot apl dot washington dot edu>]

On Sun, Oct 30, 2005 at 12:58:32PM -0800, Steve Kargl wrote:
> The attached patch has been bootstrapped and regression
> tested on i386-*-freebsd.
> 
> What does it do?  Consider the source in PR fortran/23538.
> This code will cause gfortran to go into an infinite loop,
> which isn't too unexpected in that that code is F66.  However,
> to try to debug the codei and/or gfortran, one see 438 error
> messages fly by the screen.  Many of these errors are simply
> cascaded from an earlier error are somewhat bogus.  This 
> patch disables the output of warnings and the first encountered
> error is fatal.  For example, instead of 438 errors, I now get
> 
> kargl[213] gfc41 -c -ffatal pr23538.f 
>  In file pr23538.f:308
> 
>    37 format(*0absolute addresses of first,nbaxo + nelpaz in elpa*3o21) 
>              1
> Error: Unexpected element in format string at (1)
> 
> 
> OK for mainline?
> 
> 2005-10-30  Steven G. Kargl  <kargls@comcast.net>
> 
> 	*lang.opt: Define -ffatal
> 	*gfortran.h: Add gfc_option.flag_fatal.
> 	*options.c(gfc_init_options,gfc_post_options): Set and use it.
> 	*error.c (gfc_error_check): Use it.
> 	*invoke.texi: Document -ffatal.
> 

Here's a new patch based on feedback from Andrew Pinski on IRC.
Bubblestrap and regression tested on i386-*-freebsd.

2005-10-30  Steven G. Kargl  <kargls@comcast.net>

        *lang.opt: Define -Wfatal-errors and -Werror.
        *gfortran.h: Add gfc_option.warn_error, gfc_option.warn_fatal_error.
        *options.c(gfc_init_options,gfc_post_options): Set and use it.
        *error.c (gfc_error_check): Use it.
        *invoke.texi: Document.

-- 
Steve

Attachment: fatal1.diff
Description: Text document