Patch: "GLOBAL_OFFSET_TABLE" leading underscore -- take 2

[Richard Henderson]

On Sat, Apr 17, 1999 at 08:21:33AM -0600, Donn Terry wrote:
> 	* i386/i386.c(load_pic_register): Proper number of leading _
> 	in "__GLOBAL_OFFSET_TABLE_" literal.
> 
> 	* i386/i386.h(PRINT_OPERAND_PUNCT_VALID_P): Allow _.
> 	* i386/i386.c(print_operand): new %_ operator.
> 	* i386/i386.c(load_pic_register): Use it.
> 	* i386/i386.md(prologue_get_pc_and_set_got): Likewise.
> 	* i386/unix.h(ASM_OUTPUT_MI_THUNK): Likewise.

Applied.

Please be more careful with ChangeLog style rules.  You're
missing a space before (, and don't list all changes to the
same file under the same *.  It should have been more like

        * i386.h (PRINT_OPERAND_PUNCT_VALID_P): Allow _.
        * i386.c (print_operand): New %_ operator.
        (load_pic_register): Proper number of leading _ in GOT literal.
        * i386.md (prologue_get_pc_and_set_got): Likewise.
        * i386/unix.h (ASM_OUTPUT_MI_THUNK): Likewise.

> +#ifdef YES_UNDERSCORES
>  					   gen_rtx (SYMBOL_REF, Pmode,
> -						    "$_GLOBAL_OFFSET_TABLE_"),
> +  					            "$__GLOBAL_OFFSET_TABLE_"),
> +#else
> +					            "$_GLOBAL_OFFSET_TABLE_"),
> +#endif

This wouldn't compile with !YES_UNDERSCORES.

> +	case '_':
> +#ifdef YES_UNDERSCORES
> +	    putc ('_', file);
> +#endif

Indentation is wrong.



r~