[stage1][PATCH] Change semantics of -frecord-gcc-switches and add -frecord-gcc-switches-format.

[Martin Liška]

On 1/15/21 8:14 PM, Alexandre Oliva wrote:
> Hello, Martin,
> 
> Our testing detected unexpected -dumpbase-ext options making to the
> producer string.
> 
> I tracked it down to something weird that happened in this patch:
> 
> On Dec  4, 2020, Martin Liška <mliska@suse.cz> wrote:
> 
>> +++ b/gcc/dwarf2out.c
>> -      case OPT_dumpbase:
>> -      case OPT_dumpbase_ext:
>> -      case OPT_dumpdir:
> 
>> +++ b/gcc/opts.c
>> +      case OPT_dumpbase:
>> +      case OPT_dumpdir:
> 
> Assuming you didn't really mean to drop the option, the following patch
> restores it to the exclusion list in the refactored gen_producer_string.

Hello.

Thank you for the fix. Yes, it was not an intentional change from my side.

Martin

> 
> Regstrapped on x86_64-linux-gnu, installing as obvious.
> 
> 
> drop -dumpbase-ext from producer string
> 
> From: Alexandre Oliva <oliva@adacore.com>
> 
> The -dumpbase and -dumpdir options are excluded from the producer
> string output in debug information, but -dumpbase-ext was not.  This
> patch excludes it as well.
> 
> 
> for  gcc/ChangeLog
> 
> 	* opts.c (gen_command_line_string): Exclude -dumpbase-ext.
> ---
>   gcc/opts.c |    1 +
>   1 file changed, 1 insertion(+)
> 
> diff --git a/gcc/opts.c b/gcc/opts.c
> index 527f0dde70685..437389b3de8e7 100644
> --- a/gcc/opts.c
> +++ b/gcc/opts.c
> @@ -3284,6 +3284,7 @@ gen_command_line_string (cl_decoded_option *options,
>         case OPT_o:
>         case OPT_d:
>         case OPT_dumpbase:
> +      case OPT_dumpbase_ext:
>         case OPT_dumpdir:
>         case OPT_quiet:
>         case OPT_version:
> 
>