On 04/26/2017 01:57 PM, Jakub Jelinek wrote:
> On Wed, Apr 26, 2017 at 01:48:03PM +0200, Tom de Vries wrote:
>> [ reposting with patch removed, was too big for gcc-patches@ ]
>
> I see in various tests you've ended up with ... { target ... } . }
> I think it would be better to just use ... { target ... } }
> in that case, . is the default, and if target is *-*-*, also remove the {
> target ... }.
>

I've written sed commands for both simplifications, untested patches 
attached.

[ I suppose we could drop an empty comment as the last argument as well. 
{ ... "" }. ]

> Are there any spots where the relative line number is with too big bias
> (say .+10 or .-10 and more)?

In the original patch, which rewrote all absolute line numbers, yes, it 
generated -62 ( and Mike commented on that here: 
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80221#c6 ).

In the committed patch, no. The greatest relative line number magnitude 
was 6. The script was limited to rewrite only dg directives with -1 .. 
+1 relative line number, and adjacent directives referring to the same 
line.

Thanks,
- Tom

> I guess those should be decided case by case
> whether we want relative, absolute or saved line numbers.  If the
> diagnostic is within the same function or code block as the stuff it is
> relative to, relative is fine, but if e.g. the messages are just at the end
> of file and there are unrelated functions in between, it is not a good idea
> to use relative.
>
> 	Jakub
>