[wide-int] Fix comment about GCC_BAD

Kenneth Zadeck zadeck@naturalbridge.com
Sat Nov 9 14:59:00 GMT 2013


On 11/09/2013 05:13 AM, Richard Sandiford wrote:
> There are two instances of:
>
>        if (TREE_CODE (x) != INTEGER_CST)
> 	GCC_BAD ("invalid constant in %<#pragma pack%> - ignored");
>        /* Cannot use tree_to_uhwi here or it will ice if above message printed.  */
>        align = tree_to_hwi (x);
>
> But GCC_BAD includes a return statement.
>
> As far as tree_to_hwi vs. tree_to_uhwi goes: tree_to_uhwi will ICE
> for big constants, so we can't use it in this context without first
> checking tree_fits_uhwi_p and reporting an error if it's false.
> Which might be the right thing to do, but is separate from the
> wide-int conversion.
>
> The old code used TREE_INT_CST_LOW here, so tree_to_hwi is the direct
> equivalent.
>
> Tested on powerpc64-linux-gnu and by rerunning the assembly comparison.
> OK to install?
>
> Thanks,
> Richard
>
>
> Index: gcc/c-family/c-pragma.c
> ===================================================================
> --- gcc/c-family/c-pragma.c	2013-11-09 09:39:17.797740475 +0000
> +++ gcc/c-family/c-pragma.c	2013-11-09 09:39:27.662807150 +0000
> @@ -151,7 +151,6 @@ handle_pragma_pack (cpp_reader * ARG_UNU
>       {
>         if (TREE_CODE (x) != INTEGER_CST)
>   	GCC_BAD ("invalid constant in %<#pragma pack%> - ignored");
> -      /* Cannot use tree_to_uhwi here or it will ice if above message printed.  */
>         align = tree_to_hwi (x);
>         action = set;
>         if (pragma_lex (&x) != CPP_CLOSE_PAREN)
> @@ -184,7 +183,6 @@ #define GCC_BAD_ACTION do { if (action !
>   	    {
>   	      if (TREE_CODE (x) != INTEGER_CST)
>   		GCC_BAD ("invalid constant in %<#pragma pack%> - ignored");
> -	      /* Cannot use tree_to_uhwi here or it will ice if above message printed.  */
>   	      align = tree_to_hwi (x);
>   	      if (align == -1)
>   		action = set;
this is fine with me, but you should wait for richi to see if he would 
prefer to use tree_to_uhwi.

kenny



More information about the Gcc-patches mailing list